1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
6948
6949
6950
6951
6952
6953
6954
6955
6956
6957
6958
6959
6960
6961
6962
6963
6964
6965
6966
6967
6968
6969
6970
6971
6972
6973
6974
6975
6976
6977
6978
6979
6980
6981
6982
6983
6984
6985
6986
6987
6988
6989
6990
6991
6992
6993
6994
6995
6996
6997
6998
6999
7000
7001
7002
7003
7004
7005
7006
7007
7008
7009
7010
7011
7012
7013
7014
7015
7016
7017
7018
7019
7020
7021
7022
7023
7024
7025
7026
7027
7028
7029
7030
7031
7032
7033
7034
7035
7036
7037
7038
7039
7040
7041
7042
7043
7044
7045
7046
7047
7048
7049
7050
7051
7052
7053
7054
7055
7056
7057
7058
7059
7060
7061
7062
7063
7064
7065
7066
7067
7068
7069
7070
7071
7072
7073
7074
7075
7076
7077
7078
7079
7080
7081
7082
7083
7084
7085
7086
7087
7088
7089
7090
7091
7092
7093
7094
7095
7096
7097
7098
7099
7100
7101
7102
7103
7104
7105
7106
7107
7108
7109
7110
7111
7112
7113
7114
7115
7116
7117
7118
7119
7120
7121
7122
7123
7124
7125
7126
7127
7128
7129
7130
7131
7132
7133
7134
7135
7136
7137
7138
7139
7140
7141
7142
7143
7144
7145
7146
7147
7148
7149
7150
7151
7152
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7167
7168
7169
7170
7171
7172
7173
7174
7175
7176
7177
7178
7179
7180
7181
7182
7183
7184
7185
7186
7187
7188
7189
7190
7191
7192
7193
7194
7195
7196
7197
7198
7199
7200
7201
7202
7203
7204
7205
7206
7207
7208
7209
7210
7211
7212
7213
7214
7215
7216
7217
7218
7219
7220
7221
7222
7223
7224
7225
7226
7227
7228
7229
7230
7231
7232
7233
7234
7235
7236
7237
7238
7239
7240
7241
7242
7243
7244
7245
7246
7247
7248
7249
7250
7251
7252
7253
7254
7255
7256
7257
7258
7259
7260
7261
7262
7263
7264
7265
7266
7267
7268
7269
7270
7271
7272
7273
7274
7275
7276
7277
7278
7279
7280
7281
7282
7283
7284
7285
7286
7287
7288
7289
7290
7291
7292
7293
7294
7295
7296
7297
7298
7299
7300
7301
7302
7303
7304
7305
7306
7307
7308
7309
7310
7311
7312
7313
7314
7315
7316
7317
7318
7319
7320
7321
7322
7323
7324
7325
7326
7327
7328
7329
7330
7331
7332
7333
7334
7335
7336
7337
7338
7339
7340
7341
7342
7343
7344
7345
7346
7347
7348
7349
7350
7351
7352
7353
7354
7355
7356
7357
7358
7359
7360
7361
7362
7363
7364
7365
7366
7367
7368
7369
7370
7371
7372
7373
7374
7375
7376
7377
7378
7379
7380
7381
7382
7383
7384
7385
7386
7387
7388
7389
7390
7391
7392
7393
7394
7395
7396
7397
7398
7399
7400
7401
7402
7403
7404
7405
7406
7407
7408
7409
7410
7411
7412
7413
7414
7415
7416
7417
7418
7419
7420
7421
7422
7423
7424
7425
7426
7427
7428
7429
7430
7431
7432
7433
7434
7435
7436
7437
7438
7439
7440
7441
7442
7443
7444
7445
7446
7447
7448
7449
7450
7451
7452
7453
7454
7455
7456
7457
7458
7459
7460
7461
7462
7463
7464
7465
7466
7467
7468
7469
7470
7471
7472
7473
7474
7475
7476
7477
7478
7479
7480
7481
7482
7483
7484
7485
7486
7487
7488
7489
7490
7491
7492
7493
7494
7495
7496
7497
7498
7499
7500
7501
7502
7503
7504
7505
7506
7507
7508
7509
7510
7511
7512
7513
7514
7515
7516
7517
7518
7519
7520
7521
7522
7523
7524
7525
7526
7527
7528
7529
7530
7531
7532
7533
7534
7535
7536
7537
7538
7539
7540
7541
7542
7543
7544
7545
7546
7547
7548
7549
7550
7551
7552
7553
7554
7555
7556
7557
7558
7559
7560
7561
7562
7563
7564
7565
7566
7567
7568
7569
7570
7571
7572
7573
7574
7575
7576
7577
7578
7579
7580
7581
7582
7583
7584
7585
7586
7587
7588
7589
7590
7591
7592
7593
7594
7595
7596
7597
7598
7599
7600
7601
7602
7603
7604
7605
7606
7607
7608
7609
7610
7611
7612
7613
7614
7615
7616
7617
7618
7619
7620
7621
7622
7623
7624
7625
7626
7627
7628
7629
7630
7631
7632
7633
7634
7635
7636
7637
7638
7639
7640
7641
7642
7643
7644
7645
7646
7647
7648
7649
7650
7651
7652
7653
7654
7655
7656
7657
7658
7659
7660
7661
7662
7663
7664
7665
7666
7667
7668
7669
7670
7671
7672
7673
7674
7675
7676
7677
7678
7679
7680
7681
7682
7683
7684
7685
7686
7687
7688
7689
7690
7691
7692
7693
7694
7695
7696
7697
7698
7699
7700
7701
7702
7703
7704
7705
7706
7707
7708
7709
7710
7711
7712
7713
7714
7715
7716
7717
7718
7719
7720
7721
7722
7723
7724
7725
7726
7727
7728
7729
7730
7731
7732
7733
7734
7735
7736
7737
7738
7739
7740
7741
7742
7743
7744
7745
7746
7747
7748
7749
7750
7751
7752
7753
7754
7755
7756
7757
7758
7759
7760
7761
7762
7763
7764
7765
7766
7767
7768
7769
7770
7771
7772
7773
7774
7775
7776
7777
7778
7779
7780
7781
7782
7783
7784
7785
7786
7787
7788
7789
7790
7791
7792
7793
7794
7795
7796
7797
7798
7799
7800
7801
7802
7803
7804
7805
7806
7807
7808
7809
7810
7811
7812
7813
7814
7815
7816
7817
7818
7819
7820
7821
7822
7823
7824
7825
7826
7827
7828
7829
7830
7831
7832
7833
7834
7835
7836
7837
7838
7839
7840
7841
7842
7843
7844
7845
7846
7847
7848
7849
7850
7851
7852
7853
7854
7855
7856
7857
7858
7859
7860
7861
7862
7863
7864
7865
7866
7867
7868
7869
7870
7871
7872
7873
7874
7875
7876
7877
7878
7879
7880
7881
7882
7883
7884
7885
7886
7887
7888
7889
7890
7891
7892
7893
7894
7895
7896
7897
7898
7899
7900
7901
7902
7903
7904
7905
7906
7907
7908
7909
7910
7911
7912
7913
7914
7915
7916
7917
7918
7919
7920
7921
7922
7923
7924
7925
7926
7927
7928
7929
7930
7931
7932
7933
7934
7935
7936
7937
7938
7939
7940
7941
7942
7943
7944
7945
7946
7947
7948
7949
7950
7951
7952
7953
7954
7955
7956
7957
7958
7959
7960
7961
7962
7963
7964
7965
7966
7967
7968
7969
7970
7971
7972
7973
7974
7975
7976
7977
7978
7979
7980
7981
7982
7983
7984
7985
7986
7987
7988
7989
7990
7991
7992
7993
7994
7995
7996
7997
7998
7999
8000
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
8012
8013
8014
8015
8016
8017
8018
8019
8020
8021
8022
8023
8024
8025
8026
8027
8028
8029
8030
8031
8032
8033
8034
8035
8036
8037
8038
8039
8040
8041
8042
8043
8044
8045
8046
8047
8048
8049
8050
8051
8052
8053
8054
8055
8056
8057
8058
8059
8060
8061
8062
8063
8064
8065
8066
8067
8068
8069
8070
8071
8072
8073
8074
8075
8076
8077
8078
8079
8080
8081
8082
8083
8084
8085
8086
8087
8088
8089
8090
8091
8092
8093
8094
8095
8096
8097
8098
8099
8100
8101
8102
8103
8104
8105
8106
8107
8108
8109
8110
8111
8112
8113
8114
8115
8116
8117
8118
8119
8120
8121
8122
8123
8124
8125
8126
8127
8128
8129
8130
8131
8132
8133
8134
8135
8136
8137
8138
8139
8140
8141
8142
8143
8144
8145
8146
8147
8148
8149
8150
8151
8152
8153
8154
8155
8156
8157
8158
8159
8160
8161
8162
8163
8164
8165
8166
8167
8168
8169
8170
8171
8172
8173
8174
8175
8176
8177
8178
8179
8180
8181
8182
8183
8184
8185
8186
8187
8188
8189
8190
8191
8192
8193
8194
8195
8196
8197
8198
8199
8200
8201
8202
8203
8204
8205
8206
8207
8208
8209
8210
8211
8212
8213
8214
8215
8216
8217
8218
8219
8220
8221
8222
8223
8224
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
8239
8240
8241
8242
8243
8244
8245
8246
8247
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
8261
8262
8263
8264
8265
8266
8267
8268
8269
8270
8271
8272
8273
8274
8275
8276
8277
8278
8279
8280
8281
8282
8283
8284
8285
8286
8287
8288
8289
8290
8291
8292
8293
8294
8295
8296
8297
8298
8299
8300
8301
8302
8303
8304
8305
8306
8307
8308
8309
8310
8311
8312
8313
8314
8315
8316
8317
8318
8319
8320
8321
8322
8323
8324
8325
8326
8327
8328
8329
8330
8331
8332
8333
8334
8335
8336
8337
8338
8339
8340
8341
8342
8343
8344
8345
8346
8347
8348
8349
8350
8351
8352
8353
8354
8355
8356
8357
8358
8359
8360
8361
8362
8363
8364
8365
8366
8367
8368
8369
8370
8371
8372
8373
8374
8375
8376
8377
8378
8379
8380
8381
8382
8383
8384
8385
8386
8387
8388
8389
8390
8391
8392
8393
8394
8395
8396
8397
8398
8399
8400
8401
8402
8403
8404
8405
8406
8407
8408
8409
8410
8411
8412
8413
8414
8415
8416
8417
8418
8419
8420
8421
8422
8423
8424
8425
8426
8427
8428
8429
8430
8431
8432
8433
8434
8435
8436
8437
8438
8439
8440
8441
8442
8443
8444
8445
8446
8447
8448
8449
8450
8451
8452
8453
8454
8455
8456
8457
8458
8459
8460
8461
8462
8463
8464
8465
8466
8467
8468
8469
8470
8471
8472
8473
8474
8475
8476
8477
8478
8479
8480
8481
8482
8483
8484
8485
8486
8487
8488
8489
8490
8491
8492
8493
8494
8495
8496
8497
8498
8499
8500
8501
8502
8503
8504
8505
8506
8507
8508
8509
8510
8511
8512
8513
8514
8515
8516
8517
8518
8519
8520
8521
8522
8523
8524
8525
8526
8527
8528
8529
8530
8531
8532
8533
8534
8535
8536
8537
8538
8539
8540
8541
8542
8543
8544
8545
8546
8547
8548
8549
8550
8551
8552
8553
8554
8555
8556
8557
8558
8559
8560
8561
8562
8563
8564
8565
8566
8567
8568
8569
8570
8571
8572
8573
8574
8575
8576
8577
8578
8579
8580
8581
8582
8583
8584
8585
8586
8587
8588
8589
8590
8591
8592
8593
8594
8595
8596
8597
8598
8599
8600
8601
8602
8603
8604
8605
8606
8607
8608
8609
8610
8611
8612
8613
8614
8615
8616
8617
8618
8619
8620
8621
8622
8623
8624
8625
8626
8627
8628
8629
8630
8631
8632
8633
8634
8635
8636
8637
8638
8639
8640
8641
8642
8643
8644
8645
8646
8647
8648
8649
8650
8651
8652
8653
8654
8655
8656
8657
8658
8659
8660
8661
8662
8663
8664
8665
8666
8667
8668
8669
8670
8671
8672
8673
8674
8675
8676
8677
8678
8679
8680
8681
8682
8683
8684
8685
8686
8687
8688
8689
8690
8691
8692
8693
8694
8695
8696
8697
8698
8699
8700
8701
8702
8703
8704
8705
8706
8707
8708
8709
8710
8711
8712
8713
8714
8715
8716
8717
8718
8719
8720
8721
8722
8723
8724
8725
8726
8727
8728
8729
8730
8731
8732
8733
8734
8735
8736
8737
8738
8739
8740
8741
8742
8743
8744
8745
8746
8747
8748
8749
8750
8751
8752
8753
8754
8755
8756
8757
8758
8759
8760
8761
8762
8763
8764
8765
8766
8767
8768
8769
8770
8771
8772
8773
8774
8775
8776
8777
8778
8779
8780
8781
8782
8783
8784
8785
8786
8787
8788
8789
8790
8791
8792
8793
8794
8795
8796
8797
8798
8799
8800
8801
8802
8803
8804
8805
8806
8807
8808
8809
8810
8811
8812
8813
8814
8815
8816
8817
8818
8819
8820
8821
8822
8823
8824
8825
8826
8827
8828
8829
8830
8831
8832
8833
8834
8835
8836
8837
8838
8839
8840
8841
8842
8843
8844
8845
8846
8847
8848
8849
8850
8851
8852
8853
8854
8855
8856
8857
8858
8859
8860
8861
8862
8863
8864
8865
8866
8867
8868
8869
8870
8871
8872
8873
8874
8875
8876
8877
8878
8879
8880
8881
8882
8883
8884
8885
8886
8887
8888
8889
8890
8891
8892
8893
8894
8895
8896
8897
8898
8899
8900
8901
8902
8903
8904
8905
8906
8907
8908
8909
8910
8911
8912
8913
8914
8915
8916
8917
8918
8919
8920
8921
8922
8923
8924
8925
8926
8927
8928
8929
8930
8931
8932
8933
8934
8935
8936
8937
8938
8939
8940
8941
8942
8943
8944
8945
8946
8947
8948
8949
8950
8951
8952
8953
8954
8955
8956
8957
8958
8959
8960
8961
8962
8963
8964
8965
8966
8967
8968
8969
8970
8971
8972
8973
8974
8975
8976
8977
8978
8979
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
8992
8993
8994
8995
8996
8997
8998
8999
9000
9001
9002
9003
9004
9005
9006
9007
9008
9009
9010
9011
9012
9013
9014
9015
9016
9017
9018
9019
9020
9021
9022
9023
9024
9025
9026
9027
9028
9029
9030
9031
9032
9033
9034
9035
9036
9037
9038
9039
9040
9041
9042
9043
9044
9045
9046
9047
9048
9049
9050
9051
9052
9053
9054
9055
9056
9057
9058
9059
9060
9061
9062
9063
9064
9065
9066
9067
9068
9069
9070
9071
9072
9073
9074
9075
9076
9077
9078
9079
9080
9081
9082
9083
9084
9085
9086
9087
9088
9089
9090
9091
9092
9093
9094
9095
9096
9097
9098
9099
9100
9101
9102
9103
9104
9105
9106
9107
9108
9109
9110
9111
9112
9113
9114
9115
9116
9117
9118
9119
9120
9121
9122
9123
9124
9125
9126
9127
9128
9129
9130
9131
9132
9133
9134
9135
9136
9137
9138
9139
9140
9141
9142
9143
9144
9145
9146
9147
9148
9149
9150
9151
9152
9153
9154
9155
9156
9157
9158
9159
9160
9161
9162
9163
9164
9165
9166
9167
9168
9169
9170
9171
9172
9173
9174
9175
9176
9177
9178
9179
9180
9181
9182
9183
9184
9185
9186
9187
9188
9189
9190
9191
9192
9193
9194
9195
9196
9197
9198
9199
9200
9201
9202
9203
9204
9205
9206
9207
9208
9209
9210
9211
9212
9213
9214
9215
9216
9217
9218
9219
9220
9221
9222
9223
9224
9225
9226
9227
9228
9229
9230
9231
9232
9233
9234
9235
9236
9237
9238
9239
9240
9241
9242
9243
9244
9245
9246
9247
9248
9249
9250
9251
9252
9253
9254
9255
9256
9257
9258
9259
9260
9261
9262
9263
9264
9265
9266
9267
9268
9269
9270
9271
9272
9273
9274
9275
9276
9277
9278
9279
9280
9281
9282
9283
9284
9285
9286
9287
9288
9289
9290
9291
9292
9293
9294
9295
9296
9297
9298
9299
9300
9301
9302
9303
9304
9305
9306
9307
9308
9309
9310
9311
9312
9313
9314
9315
9316
9317
9318
9319
9320
9321
9322
9323
9324
9325
9326
9327
9328
9329
9330
9331
9332
9333
9334
9335
9336
9337
9338
9339
9340
9341
9342
9343
9344
9345
9346
9347
9348
9349
9350
9351
9352
9353
9354
9355
9356
9357
9358
9359
9360
9361
9362
9363
9364
9365
9366
9367
9368
9369
9370
9371
9372
9373
9374
9375
9376
9377
9378
9379
9380
9381
9382
9383
9384
9385
9386
9387
9388
9389
9390
9391
9392
9393
9394
9395
9396
9397
9398
9399
9400
9401
9402
9403
9404
9405
9406
9407
9408
9409
9410
9411
9412
9413
9414
9415
9416
9417
9418
9419
9420
9421
9422
9423
9424
9425
9426
9427
9428
9429
9430
9431
9432
9433
9434
9435
9436
9437
9438
9439
9440
9441
9442
9443
9444
9445
9446
9447
9448
9449
9450
9451
9452
9453
9454
9455
9456
9457
9458
9459
9460
9461
9462
9463
9464
9465
9466
9467
9468
9469
9470
9471
9472
9473
9474
9475
9476
9477
9478
9479
9480
9481
9482
9483
9484
9485
9486
9487
9488
9489
9490
9491
9492
9493
9494
9495
9496
9497
9498
9499
9500
9501
9502
9503
9504
9505
9506
9507
9508
9509
9510
9511
9512
9513
9514
9515
9516
9517
9518
9519
9520
9521
9522
9523
9524
9525
9526
9527
9528
9529
9530
9531
9532
9533
9534
9535
9536
9537
9538
9539
9540
9541
9542
9543
9544
9545
9546
9547
9548
9549
9550
9551
9552
9553
9554
9555
9556
9557
9558
9559
9560
9561
9562
9563
9564
9565
9566
9567
9568
9569
9570
9571
9572
9573
9574
9575
9576
9577
9578
9579
9580
9581
9582
9583
9584
9585
9586
9587
9588
9589
9590
9591
9592
9593
9594
9595
9596
9597
9598
9599
9600
9601
9602
9603
9604
9605
9606
9607
9608
9609
9610
9611
9612
9613
9614
9615
9616
9617
9618
9619
9620
9621
9622
9623
9624
9625
9626
9627
9628
9629
9630
9631
9632
9633
9634
9635
9636
9637
9638
9639
9640
9641
9642
9643
9644
9645
9646
9647
9648
9649
9650
9651
9652
9653
9654
9655
9656
9657
9658
9659
9660
9661
9662
9663
9664
9665
9666
9667
9668
9669
9670
9671
9672
9673
9674
9675
9676
9677
9678
9679
9680
9681
9682
9683
9684
9685
9686
9687
9688
9689
9690
9691
9692
9693
9694
9695
9696
9697
9698
9699
9700
9701
9702
9703
9704
9705
9706
9707
9708
9709
9710
9711
9712
9713
9714
9715
9716
9717
9718
9719
9720
9721
9722
9723
9724
9725
9726
9727
9728
9729
9730
9731
9732
9733
9734
9735
9736
9737
9738
9739
9740
9741
9742
9743
9744
9745
9746
9747
9748
9749
9750
9751
9752
9753
9754
9755
9756
9757
9758
9759
9760
9761
9762
9763
9764
9765
9766
9767
9768
9769
9770
9771
9772
9773
9774
9775
9776
9777
9778
9779
9780
9781
9782
9783
9784
9785
9786
9787
9788
9789
9790
9791
9792
9793
9794
9795
9796
9797
9798
9799
9800
9801
9802
9803
9804
9805
9806
9807
9808
9809
9810
9811
9812
9813
9814
9815
9816
9817
9818
9819
9820
9821
9822
9823
9824
9825
9826
9827
9828
9829
9830
9831
9832
9833
9834
9835
9836
9837
9838
9839
9840
9841
9842
9843
9844
9845
9846
9847
9848
9849
9850
9851
9852
9853
9854
9855
9856
9857
9858
9859
9860
9861
9862
9863
9864
9865
9866
9867
9868
9869
9870
9871
9872
9873
9874
9875
9876
9877
9878
9879
9880
9881
9882
|
<!DOCTYPE html>
<html class="writer-html5" lang="en" >
<head>
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>User Guide — nsight-systems 2023.4 documentation</title>
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<link rel="stylesheet" href="../_static/design-style.b7bb847fb20b106c3d81b95245e65545.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/omni-style.css" type="text/css" />
<link rel="stylesheet" href="../_static/api-styles.css" type="text/css" />
<link rel="shortcut icon" href="../_static/favicon.ico"/>
<!--[if lt IE 9]>
<script src="../_static/js/html5shiv.min.js"></script>
<![endif]-->
<script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
<script src="../_static/jquery.js"></script>
<script src="../_static/underscore.js"></script>
<script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
<script src="../_static/doctools.js"></script>
<script src="../_static/sphinx_highlight.js"></script>
<script src="../_static/mermaid-init.js"></script>
<script src="../_static/design-tabs.js"></script>
<script src="../_static/version.js"></script>
<script src="../_static/social-media.js"></script>
<script src="../_static/js/theme.js"></script>
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
<link rel="next" title="Copyright and Licenses" href="../CopyrightAndLicenses/index.html" />
<link rel="prev" title="Installation Guide" href="../InstallationGuide/index.html" />
</head>
<body class="wy-body-for-nav">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side">
<div class="wy-side-scroll">
<div class="wy-side-nav-search" >
<a href="../index.html">
<img src="../_static/devzone.png" class="logo" alt="Logo"/>
</a>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../ReleaseNotes/index.html">Release Notes</a></li>
<li class="toctree-l1"><a class="reference internal" href="../InstallationGuide/index.html">Installation Guide</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">User Guide</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#profiling-from-the-cli">Profiling from the CLI</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#installing-the-cli-on-your-target">Installing the CLI on Your Target</a></li>
<li class="toctree-l3"><a class="reference internal" href="#command-line-options">Command Line Options</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#cli-global-options">CLI Global Options</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#cli-command-switches">CLI Command Switches</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#cli-analyze-command-switch-options">CLI Analyze Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-cancel-command-switch-options">CLI Cancel Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-export-command-switch-options">CLI Export Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-launch-command-switch-options">CLI Launch Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-profile-command-switch-options">CLI Profile Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-sessions-command-switch-subcommands">CLI Sessions Command Switch Subcommands</a><ul>
<li class="toctree-l5"><a class="reference internal" href="#cli-sessions-list-command-switch-options">CLI Sessions List Command Switch Options</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="#cli-shutdown-command-switch-options">CLI Shutdown Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-start-command-switch-options">CLI Start Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-stats-command-switch-options">CLI Stats Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-status-command-switch-options">CLI Status Command Switch Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#cli-stop-command-switch-options">CLI Stop Command Switch Options</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#example-single-command-lines">Example Single Command Lines</a></li>
<li class="toctree-l3"><a class="reference internal" href="#example-interactive-cli-command-sequences">Example Interactive CLI Command Sequences</a></li>
<li class="toctree-l3"><a class="reference internal" href="#example-stats-command-sequences">Example Stats Command Sequences</a></li>
<li class="toctree-l3"><a class="reference internal" href="#example-output-from-stats-option">Example Output from <code class="docutils literal notranslate"><span class="pre">--stats</span></code> Option</a></li>
<li class="toctree-l3"><a class="reference internal" href="#importing-and-viewing-command-line-results-files">Importing and Viewing Command Line Results Files</a></li>
<li class="toctree-l3"><a class="reference internal" href="#using-the-cli-to-analyze-mpi-codes">Using the CLI to Analyze MPI Codes</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#tracing-mpi-api-calls">Tracing MPI API calls</a></li>
<li class="toctree-l4"><a class="reference internal" href="#using-the-cli-to-profile-applications-launched-with-mpirun">Using the CLI to Profile Applications Launched with mpirun</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#profiling-from-the-gui">Profiling from the GUI</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#profiling-linux-targets-from-the-gui">Profiling Linux Targets from the GUI</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#connecting-to-the-target-device">Connecting to the Target Device</a></li>
<li class="toctree-l4"><a class="reference internal" href="#system-wide-profiling-options">System-Wide Profiling Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#target-sampling-options">Target Sampling Options</a></li>
<li class="toctree-l4"><a class="reference internal" href="#hotkey-trace-start-stop">Hotkey Trace Start/Stop</a></li>
<li class="toctree-l4"><a class="reference internal" href="#launching-processes">Launching Processes</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#profiling-windows-targets-from-the-gui">Profiling Windows Targets from the GUI</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#remoting-to-a-windows-based-machine">Remoting to a Windows Based Machine</a></li>
<li class="toctree-l4"><a class="reference internal" href="#id1">Hotkey Trace Start/Stop</a></li>
<li class="toctree-l4"><a class="reference internal" href="#target-sampling-options-on-windows">Target Sampling Options on Windows</a></li>
<li class="toctree-l4"><a class="reference internal" href="#symbol-locations">Symbol Locations</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#profiling-qnx-targets-from-the-gui">Profiling QNX Targets from the GUI</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#container-support-on-linux-servers">Container Support on Linux Servers</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#gui-vnc-container">GUI VNC container</a></li>
<li class="toctree-l3"><a class="reference internal" href="#gui-webrtc-container">GUI WebRTC container</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#migrating-from-nvidia-nvprof">Migrating from NVIDIA nvprof</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#using-the-product-name-cli-nvprof-command">Using the Nsight Systems CLI nvprof Command</a></li>
<li class="toctree-l3"><a class="reference internal" href="#cli-nvprof-command-switch-options">CLI nvprof Command Switch Options</a></li>
<li class="toctree-l3"><a class="reference internal" href="#next-steps">Next Steps</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#direct3d-trace">Direct3D Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#d3d11-api-trace">D3D11 API trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#d3d12-api-trace">D3D12 API Trace</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#wddm-queues">WDDM Queues</a></li>
<li class="toctree-l2"><a class="reference internal" href="#wddm-hw-scheduler">WDDM HW Scheduler</a></li>
<li class="toctree-l2"><a class="reference internal" href="#vulkan-api-trace">Vulkan API Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#vulkan-overview">Vulkan Overview</a></li>
<li class="toctree-l3"><a class="reference internal" href="#pipeline-creation-feedback">Pipeline Creation Feedback</a></li>
<li class="toctree-l3"><a class="reference internal" href="#vulkan-gpu-trace-notes">Vulkan GPU Trace Notes</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#stutter-analysis">Stutter Analysis</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#fps-overview">FPS Overview</a></li>
<li class="toctree-l3"><a class="reference internal" href="#frame-health">Frame Health</a></li>
<li class="toctree-l3"><a class="reference internal" href="#gpu-memory">GPU Memory</a></li>
<li class="toctree-l3"><a class="reference internal" href="#vertical-synchronization">Vertical Synchronization</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#openmp-trace">OpenMP Trace</a></li>
<li class="toctree-l2"><a class="reference internal" href="#os-runtime-libraries-trace">OS Runtime Libraries Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#locking-a-resource">Locking a Resource</a></li>
<li class="toctree-l3"><a class="reference internal" href="#limitations">Limitations</a></li>
<li class="toctree-l3"><a class="reference internal" href="#os-runtime-libraries-trace-filters">OS Runtime Libraries Trace Filters</a></li>
<li class="toctree-l3"><a class="reference internal" href="#os-runtime-default-function-list">OS Runtime Default Function List</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#nvtx-trace">NVTX Trace</a></li>
<li class="toctree-l2"><a class="reference internal" href="#cuda-trace">CUDA Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#cuda-gpu-memory-allocation-graph">CUDA GPU Memory Allocation Graph</a></li>
<li class="toctree-l3"><a class="reference internal" href="#unified-memory-transfer-trace">Unified Memory Transfer Trace</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#unified-memory-cpu-page-faults">Unified Memory CPU Page Faults</a></li>
<li class="toctree-l4"><a class="reference internal" href="#unified-memory-gpu-page-faults">Unified Memory GPU Page Faults</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#cuda-graph-trace">CUDA Graph Trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#cuda-python-backtrace">CUDA Python Backtrace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#cuda-default-function-list-for-cli">CUDA Default Function List for CLI</a></li>
<li class="toctree-l3"><a class="reference internal" href="#cudnn-function-list-for-x86-cli">cuDNN Function List for X86 CLI</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#openacc-trace">OpenACC Trace</a></li>
<li class="toctree-l2"><a class="reference internal" href="#opengl-trace">OpenGL Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#opengl-trace-using-command-line">OpenGL Trace Using Command Line</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#custom-etw-trace">Custom ETW Trace</a></li>
<li class="toctree-l2"><a class="reference internal" href="#gpu-metrics">GPU Metrics</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#overview">Overview</a></li>
<li class="toctree-l3"><a class="reference internal" href="#launching-gpu-metrics-from-the-cli">Launching GPU Metrics from the CLI</a></li>
<li class="toctree-l3"><a class="reference internal" href="#launching-gpu-metrics-from-the-gui">Launching GPU Metrics from the GUI</a></li>
<li class="toctree-l3"><a class="reference internal" href="#sampling-frequency">Sampling frequency</a></li>
<li class="toctree-l3"><a class="reference internal" href="#available-metrics">Available metrics</a></li>
<li class="toctree-l3"><a class="reference internal" href="#exporting-and-querying-data">Exporting and Querying Data</a></li>
<li class="toctree-l3"><a class="reference internal" href="#id2">Limitations</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#soc-metrics">SOC Metrics</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#id3">Overview</a></li>
<li class="toctree-l3"><a class="reference internal" href="#id4">Available metrics</a></li>
<li class="toctree-l3"><a class="reference internal" href="#launching-soc-metrics-from-the-cli">Launching SOC Metrics from the CLI</a></li>
<li class="toctree-l3"><a class="reference internal" href="#launching-soc-metrics-from-the-gui">Launching SOC Metrics from the GUI</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#cpu-profiling-using-linux-os-perf-subsystem">CPU Profiling Using Linux OS Perf Subsystem</a></li>
<li class="toctree-l2"><a class="reference internal" href="#nvidia-video-codec-sdk-trace">NVIDIA Video Codec SDK Trace</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#nv-encoder-api-functions-traced-by-default">NV Encoder API Functions Traced by Default</a></li>
<li class="toctree-l3"><a class="reference internal" href="#nv-decoder-api-functions-traced-by-default">NV Decoder API Functions Traced by Default</a></li>
<li class="toctree-l3"><a class="reference internal" href="#nv-jpeg-api-functions-traced-by-default">NV JPEG API Functions Traced by Default</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#network-communication-profiling">Network Communication Profiling</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#mpi-api-trace">MPI API Trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#openshmem-library-trace">OpenSHMEM Library Trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#ucx-library-trace">UCX Library Trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#nvidia-nvshmem-and-nccl-trace">NVIDIA NVSHMEM and NCCL Trace</a></li>
<li class="toctree-l3"><a class="reference internal" href="#nic-metric-sampling">NIC Metric Sampling</a></li>
<li class="toctree-l3"><a class="reference internal" href="#infiniband-switch-metric-sampling">InfiniBand Switch Metric Sampling</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#python-profiling">Python Profiling</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#python-backtrace-sampling">Python Backtrace Sampling</a></li>
<li class="toctree-l3"><a class="reference internal" href="#python-nvtx-annotations">Python NVTX Annotations</a></li>
<li class="toctree-l3"><a class="reference internal" href="#python-gil-tracing">Python GIL Tracing</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#reading-your-report-in-gui">Reading Your Report in GUI</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#generating-a-new-report">Generating a New Report</a></li>
<li class="toctree-l3"><a class="reference internal" href="#opening-an-existing-report">Opening an Existing Report</a></li>
<li class="toctree-l3"><a class="reference internal" href="#sharing-a-report-file">Sharing a Report File</a></li>
<li class="toctree-l3"><a class="reference internal" href="#report-tab">Report Tab</a></li>
<li class="toctree-l3"><a class="reference internal" href="#analysis-summary-view">Analysis Summary View</a></li>
<li class="toctree-l3"><a class="reference internal" href="#timeline-view">Timeline View</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#timeline">Timeline</a></li>
<li class="toctree-l4"><a class="reference internal" href="#events-view">Events View</a></li>
<li class="toctree-l4"><a class="reference internal" href="#function-table-modes">Function Table Modes</a></li>
<li class="toctree-l4"><a class="reference internal" href="#function-table-notes">Function Table Notes</a></li>
<li class="toctree-l4"><a class="reference internal" href="#filter-dialog">Filter Dialog</a></li>
<li class="toctree-l4"><a class="reference internal" href="#example-of-using-timeline-with-function-table">Example of Using Timeline with Function Table</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#diagnostics-summary-view">Diagnostics Summary View</a></li>
<li class="toctree-l3"><a class="reference internal" href="#symbol-resolution-logs-view">Symbol Resolution Logs View</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#adding-report-to-the-timeline">Adding Report to the Timeline</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#time-synchronization">Time Synchronization</a></li>
<li class="toctree-l3"><a class="reference internal" href="#timeline-hierarchy">Timeline Hierarchy</a></li>
<li class="toctree-l3"><a class="reference internal" href="#example-mpi">Example: MPI</a></li>
<li class="toctree-l3"><a class="reference internal" href="#id5">Limitations</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#post-collection-analysis">Post-Collection Analysis</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#available-export-formats">Available Export Formats</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#sqlite-schema-reference">SQLite Schema Reference</a></li>
<li class="toctree-l4"><a class="reference internal" href="#sqlite-schema-event-values">SQLite Schema Event Values</a></li>
<li class="toctree-l4"><a class="reference internal" href="#common-sqlite-examples">Common SQLite Examples</a></li>
<li class="toctree-l4"><a class="reference internal" href="#arrow-format-description">Arrow Format Description</a></li>
<li class="toctree-l4"><a class="reference internal" href="#json-and-text-format-description">JSON and Text Format Description</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#statistical-analysis">Statistical Analysis</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#statistical-reports-shipped-with-product-name">Statistical Reports Shipped With Nsight Systems</a></li>
<li class="toctree-l4"><a class="reference internal" href="#report-formatters-shipped-with-product-name">Report Formatters Shipped With Nsight Systems</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#expert-systems-analysis">Expert Systems Analysis</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#using-expert-system-from-the-cli">Using Expert System from the CLI</a></li>
<li class="toctree-l4"><a class="reference internal" href="#using-expert-system-from-the-gui">Using Expert System from the GUI</a></li>
<li class="toctree-l4"><a class="reference internal" href="#expert-system-rules">Expert System Rules</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#multi-report-analysis">Multi-Report Analysis</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#available-recipes">Available Recipes</a></li>
<li class="toctree-l4"><a class="reference internal" href="#opening-in-jupyter-notebook">Opening in Jupyter Notebook</a></li>
<li class="toctree-l4"><a class="reference internal" href="#configuring-dask">Configuring Dask</a></li>
<li class="toctree-l4"><a class="reference internal" href="#tutorial-create-a-user-defined-recipe">Tutorial: Create a User-Defined Recipe</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#import-nvtxt">Import NVTXT</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#commands">Commands</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#visual-studio-integration">Visual Studio Integration</a></li>
<li class="toctree-l2"><a class="reference internal" href="#troubleshooting">Troubleshooting</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#general-troubleshooting">General Troubleshooting</a></li>
<li class="toctree-l3"><a class="reference internal" href="#cli-troubleshooting">CLI Troubleshooting</a></li>
<li class="toctree-l3"><a class="reference internal" href="#launch-processes-in-stopped-state">Launch Processes in Stopped State</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#ld-preload">LD_PRELOAD</a></li>
<li class="toctree-l4"><a class="reference internal" href="#launcher">Launcher</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#gui-troubleshooting">GUI Troubleshooting</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#ubuntu-18-04-20-04-22-04-and-centos-7-8-9-with-root-privileges">Ubuntu 18.04/20.04/22.04 and CentOS 7/8/9 with root privileges</a></li>
<li class="toctree-l4"><a class="reference internal" href="#ubuntu-18-04-20-04-22-04-and-centos-7-8-9-without-root-privileges">Ubuntu 18.04/20.04/22.04 and CentOS 7/8/9 without root privileges</a></li>
<li class="toctree-l4"><a class="reference internal" href="#other-platforms-or-if-the-previous-steps-did-not-help">Other platforms, or if the previous steps did not help</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#symbol-resolution">Symbol Resolution</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#broken-backtraces-on-tegra">Broken Backtraces on Tegra</a></li>
<li class="toctree-l4"><a class="reference internal" href="#debug-versions-of-elf-files">Debug Versions of ELF Files</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="#logging">Logging</a><ul>
<li class="toctree-l4"><a class="reference internal" href="#verbose-remote-logging-on-linux-targets">Verbose Remote Logging on Linux Targets</a></li>
<li class="toctree-l4"><a class="reference internal" href="#verbose-cli-logging-on-linux-targets">Verbose CLI Logging on Linux Targets</a></li>
<li class="toctree-l4"><a class="reference internal" href="#verbose-logging-on-windows-targets">Verbose Logging on Windows Targets</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#other-resources">Other Resources</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#training-seminars">Training Seminars</a></li>
<li class="toctree-l3"><a class="reference internal" href="#blog-posts">Blog Posts</a></li>
<li class="toctree-l3"><a class="reference internal" href="#feature-videos">Feature Videos</a></li>
<li class="toctree-l3"><a class="reference internal" href="#conference-presentations">Conference Presentations</a></li>
<li class="toctree-l3"><a class="reference internal" href="#for-more-support">For More Support</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../CopyrightAndLicenses/index.html">Copyright and Licenses</a></li>
</ul>
<p class="caption" role="heading"><span class="caption-text">Archives</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../Archives/index.html">Archives</a></li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../index.html">nsight-systems</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="Page navigation">
<ul class="wy-breadcrumbs">
<li><a href="../index.html" class="icon icon-home"></a> »</li>
<li>User Guide</li>
<li class="wy-breadcrumbs-aside">
</li>
<li class="wy-breadcrumbs-aside">
<span>v2023.4 |</span>
<a href="https://developer.nvidia.com/nsight-systems/" class="reference external">Archive</a>
<span> </span>
</li>
</ul>
<hr/>
</div>
<div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
<div itemprop="articleBody">
<section id="user-guide">
<h1>User Guide<a class="headerlink" href="#user-guide" title="Permalink to this heading"></a></h1>
<p>NVIDIA Nsight Systems user guide</p>
<section id="profiling-from-the-cli">
<span id="cli-profiling"></span><h2>Profiling from the CLI<a class="headerlink" href="#profiling-from-the-cli" title="Permalink to this heading"></a></h2>
<section id="installing-the-cli-on-your-target">
<h3>Installing the CLI on Your Target<a class="headerlink" href="#installing-the-cli-on-your-target" title="Permalink to this heading"></a></h3>
<p>The Nsight Systems CLI provides a simple interface to collect on a target without using the GUI. The collected data can then be copied to any system and analyzed later.</p>
<p>The CLI is distributed in the Target directory of the standard Nsight Systems download package. Users who want to install the CLI as a standalone tool can do so by copying the files within the Target directory. If you want the CLI output file (.qdstrm) to be auto-converted (to .nsys-rep) after the analysis is complete, you will need to copy the host directory as well.</p>
<p>If you wish to run the CLI without root (recommended mode), you will want to install in a directory where you have full access.</p>
<p>Note that you must run the CLI on Windows as administrator.</p>
</section>
<section id="command-line-options">
<h3>Command Line Options<a class="headerlink" href="#command-line-options" title="Permalink to this heading"></a></h3>
<p>The Nsight Systems command lines can have one of two forms:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global_option</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>or</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">command_switch</span><span class="p">][</span><span class="n">optional</span><span class="w"> </span><span class="n">command_switch_options</span><span class="p">][</span><span class="n">application</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">optional</span><span class="w"> </span><span class="n">application_options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>All command line options are case sensitive. For command switch options, when short options are used, the parameters should follow the switch after a space; e.g. <code class="docutils literal notranslate"><span class="pre">-s</span> <span class="pre">process-tree</span></code>. When long options are used, the switch should be followed by an equal sign and then the parameter(s); e.g. <code class="docutils literal notranslate"><span class="pre">--sample=process-tree</span></code>.</p>
<p>For this version of Nsight Systems, if you launch a process from the command line to begin analysis, the launched process will be terminated when collection is complete, including runs with <code class="docutils literal notranslate"><span class="pre">--duration</span></code> set, unless the user specifies the <code class="docutils literal notranslate"><span class="pre">--kill</span> <span class="pre">none</span></code> option (details below). The exception is that if the user uses NVTX, cudaProfilerStart/Stop, or hotkeys to control the duration, the application will continue unless <code class="docutils literal notranslate"><span class="pre">--kill</span></code> is set.</p>
<p>The Nsight Systems CLI supports concurrent analysis by using sessions. Each Nsight Systems session is defined by a sequence of CLI commands that define one or more collections (e.g. when and what data is collected). A session begins with either a start, launch, or profile command. A session ends with a shutdown command, when a profile command terminates, or, if requested, when all the process tree(s) launched in the session exit. Multiple sessions can run concurrently on the same system.</p>
<section id="cli-global-options">
<h4>CLI Global Options<a class="headerlink" href="#cli-global-options" title="Permalink to this heading"></a></h4>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>-h</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p>Help message providing information about available command switches and their options.</p></td>
</tr>
<tr class="row-odd"><td><p>-v</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--version</span></code></p></td>
<td><p>Output Nsight Systems CLI version information.</p></td>
</tr>
</tbody>
</table>
</section>
</section>
<section id="cli-command-switches">
<h3>CLI Command Switches<a class="headerlink" href="#cli-command-switches" title="Permalink to this heading"></a></h3>
<p>The Nsight Systems command line interface can be used in two modes. You may launch your application and begin analysis with options specified to the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> command. Alternatively, you can control the launch of an application and data collection using interactive CLI commands.</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Command</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>analyze</p></td>
<td><p>Post process existing Nsight Systems result, either in .nsys-rep or SQLite format, to generate expert systems report.</p></td>
</tr>
<tr class="row-odd"><td><p>cancel</p></td>
<td><p>Cancels an existing collection started in interactive mode. All data already collected in the current collection is discarded.</p></td>
</tr>
<tr class="row-even"><td><p>export</p></td>
<td><p>Generates an export file from an existing .nsys-rep file. For more information about the exported formats see the /documentation/nsys-exporter directory in your Nsight Systems installation directory.</p></td>
</tr>
<tr class="row-odd"><td><p>launch</p></td>
<td><p>In interactive mode, launches an application in an environment that supports the requested options. The launch command can be executed before or after a start command.</p></td>
</tr>
<tr class="row-even"><td><p>nvprof</p></td>
<td><p>Special option to help with transition from legacy NVIDIA nvprof tool. Calling <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">nvprof</span> <span class="pre">[options]</span></code> will provide the best available translation of <code class="docutils literal notranslate"><span class="pre">nvprof</span> <span class="pre">[options]</span></code> See <strong>Migrating from NVIDIA nvprof</strong> topic for details. No additional functionality of nsys will be available when using this option. Note: Not available on IBM Power targets.</p></td>
</tr>
<tr class="row-odd"><td><p>profile</p></td>
<td><p>A fully formed profiling description requiring and accepting no further input. The command switch options used (see below table) determine when the collection starts, stops, what collectors are used (e.g. API trace, IP sampling, etc.), what processes are monitored, etc.</p></td>
</tr>
<tr class="row-even"><td><p>recipe</p></td>
<td><p><strong>PREVIEW FEATURE</strong> Post process multiple existing Nsight Systems results, in .nsys-rep or SQLite to generate statistical information and create various plots. See <strong>Multi-Node Analysis</strong> topic for details.</p></td>
</tr>
<tr class="row-odd"><td><p>sessions</p></td>
<td><p>Gives information about all sessions running on the system.</p></td>
</tr>
<tr class="row-even"><td><p>shutdown</p></td>
<td><p>Disconnects the CLI process from the launched application and forces the CLI process to exit. If a collection is pending or active, it is cancelled</p></td>
</tr>
<tr class="row-odd"><td><p>start</p></td>
<td><p>Start a collection in interactive mode. The start command can be executed before or after a launch command.</p></td>
</tr>
<tr class="row-even"><td><p>stats</p></td>
<td><p>Post process existing Nsight Systems result, either in .nsys-rep or SQLite format, to generate statistical information.</p></td>
</tr>
<tr class="row-odd"><td><p>status</p></td>
<td><p>Reports on the status of a CLI-based collection or the suitability of the profiling environment.</p></td>
</tr>
<tr class="row-even"><td><p>stop</p></td>
<td><p>Stop a collection that was started in interactive mode. When executed, all active collections stop, the CLI process terminates but the application continues running.</p></td>
</tr>
</tbody>
</table>
<section id="cli-analyze-command-switch-options">
<h4>CLI Analyze Command Switch Options<a class="headerlink" href="#cli-analyze-command-switch-options" title="Permalink to this heading"></a></h4>
<p>The <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">analyze</span></code> command generates and outputs to the terminal a report using expert system rules on existing results. Reports are generated from an SQLite export of a .nsys-rep file. If a .nsys-rep file is specified, Nsight Systems will look for an accompanying SQLite file and use it. If no SQLite export file exists, one will be created.</p>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">analyze</span></code> command switch, the following options are available. Usage:</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">[global-options]</span> <span class="pre">analyze</span> <span class="pre">[options]</span> <span class="pre">[input-file]</span></code></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options
relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--format</span></code></p></td>
<td><p>column, table, csv, tsv, json, hdoc, htable, .</p></td>
<td></td>
<td><p>Specify the output format. The special name “.” indicates the default format for the given output. The default format for console is column, while files and process outputs default to csv. This option may be used multiple times. Multiple formats may also be specified using a comma-separated list (<name[:args…][,name[:args…]…]>). See <strong>Report Scripts</strong> for options available with each format.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-export</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Force a re-export of the SQLite file from the specified .nsys-rep file, even if an SQLite file already exists.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Overwrite any existing output files.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help-formats</span></code></p></td>
<td><p><format_name>, ALL, [none]</p></td>
<td><p>none</p></td>
<td><p>With no argument, list a summary of the available output formats. If a format name is given, a more detailed explanation of the the format is displayed. If <code class="docutils literal notranslate"><span class="pre">ALL</span></code> is given, a more detailed explanation of all available formats is displayed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help-rules</span></code></p></td>
<td><p><rule_name>, ALL, [none]</p></td>
<td><p>none</p></td>
<td><p>With no argument, list available rules with a short description. If a rule name is given, a more detailed explanation of the rule is displayed. If <code class="docutils literal notranslate"><span class="pre">ALL</span></code> is given, a more detailed explanation of all available rules is displayed.</p></td>
</tr>
<tr class="row-even"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p>-, @<command>, <basename>, .</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">-</span></code></p></td>
<td><p>Specify the output mechanism. There are three output mechanisms: print to console, output to file, or output to command. This option may be used multiple times. Multiple outputs may also be specified using a comma-separated list. If the given output name is “-”, the output will be displayed on the console. If the output name starts with “@”, the output designates a command to run. The nsys command will be executed and the analysis output will be piped into the command. Any other output is assumed to be the base path and name for a file. If a file basename is given, the filename used will be: <basename>_<analysis&args>.<output_format>. The default base (including path) is the name of the SQLite file (as derived from the input file or <code class="docutils literal notranslate"><span class="pre">--sqlite</span></code> option), minus the extension. The output “.” can be used to indicate the analysis should be output to a file, and the default basename should be used. To write one or more analysis outputs to files using the default basename, use <code class="docutils literal notranslate"><span class="pre">--output</span></code>. If the output starts with “@”, the nsys command output is piped to the given command. The command is run, and the output is piped to the command’s stdin (standard-input). The command’s stdout and stderr remain attached to the console, so any output will be displayed directly to the console. Be aware there are some limitations in how the command string is parsed. No shell expansions (including *, ?, [], and ~) are supported. The command cannot be piped to another command, nor redirected to a file using shell syntax. The command and command arguments are split on whitespace, and no quotes (within the command syntax) are supported. For commands that require complex command line syntax, it is suggested that the command be put into a shell script file, and the script designated as the output command.</p></td>
</tr>
<tr class="row-odd"><td><p>-q</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--quiet</span></code></p></td>
<td></td>
<td></td>
<td><p>Do not display verbose messages, only display errors.</p></td>
</tr>
<tr class="row-even"><td><p>-r</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--rule</span></code></p></td>
<td><p>cuda_memcpy_async, cuda_memcpy_sync, cuda_memset_sync, cuda_api_sync, gpu_gaps, gpu_time_util, dx12_mem_ops</p></td>
<td><p>all</p></td>
<td><p>Specify the rules(s) to execute, including any arguments. This option may be used multiple times. Multiple rules may also be specified using a comma-separated list. See <strong>Expert Systems</strong> section and <code class="docutils literal notranslate"><span class="pre">--help-rules</span></code> switch for details on all rules.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sqlite</span></code></p></td>
<td><p><file.sqlite></p></td>
<td></td>
<td><p>Specify the SQLite export filename. If this file exists, it will be used. If this file doesn’t exist (or if <code class="docutils literal notranslate"><span class="pre">--force-export</span></code> was given) this file will be created from the specified .nsys-rep file before processing. This option cannot be used if the specified input file is also an SQLite file.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--timeunit</span></code></p></td>
<td><p>nsec, nanoseconds, usec, microseconds, msec, milliseconds, seconds</p></td>
<td><p>nanoseconds</p></td>
<td><p>Set basic unit of time. The argument of the switch is matched by using the longest prefix matching. Meaning that it is not necessary to write a whole word as the switch argument. It is similar to passing a “:time=<unit>” argument to every formatter, although the formatter uses more strict naming conventions. See <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">analyze</span> <span class="pre">--help-formats</span> <span class="pre">column</span></code> for more detailed information on unit conversion.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-cancel-command-switch-options">
<h4>CLI Cancel Command Switch Options<a class="headerlink" href="#cli-cancel-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">cancel</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">cancel</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p><session identifier></p></td>
<td><p>none</p></td>
<td><p>Cancel the collection in the given session. The option argument must represent a valid session name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern in the option argument will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern in the option argument will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern in the option argument will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-export-command-switch-options">
<h4>CLI Export Command Switch Options<a class="headerlink" href="#cli-export-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">export</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="k">export</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="o">-</span><span class="n">file</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, overwrite all existing result files with same output filename (QDSTRM, nsys-rep, SQLITE, HDF, TEXT, ARROW, JSON).</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-even"><td><p>-l</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--lazy</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>Controls if table creation is lazy or not. When true, a table will only be created when it contains data. This option will be deprecated in the future, and all exports will be non-lazy. This affects SQLite, HDF5, and Arrow exports only.</p></td>
</tr>
<tr class="row-odd"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p><filename></p></td>
<td><p><inputfile.ext></p></td>
<td><p>Set the .output filename. The default is the input filename with the extension for the chosen format.</p></td>
</tr>
<tr class="row-even"><td><p>-q</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--quiet</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, do not display progress bar</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--separate-strings</span></code></p></td>
<td><p>true,false</p></td>
<td><p>false</p></td>
<td><p>Output stored strings and thread names separately, with one value per line. This affects JSON and text output only.</p></td>
</tr>
<tr class="row-even"><td><p>-t</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--type</span></code></p></td>
<td><p>arrow, hdf, info, json, sqlite, text</p></td>
<td><p>sqlite</p></td>
<td><p>Export format type. HDF format is supported only on x86_64 Linux and Windows</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ts-normalize</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, all timestamp values in the report will be shifted to UTC wall-clock time, as defined by the UNIX epoch. This option can be used in conjunction with the <code class="docutils literal notranslate"><span class="pre">--ts-shift</span></code> option, in which case both adjustments will be applied. If this option is used to align a series of reports from a cluster or distributed system, the accuracy of the alignment is limited by the synchronization precision of the system clocks. For detailed analysis, the use of PTP or another high-precision synchronization methodology is recommended. NTP is unlikely to produce desirable results. This option only applies to Arrow, HDF5, and SQLite exports.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ts-shift</span></code></p></td>
<td><p>signed integer, in nanoseconds</p></td>
<td><p>0</p></td>
<td><p>If given, all timestamp values in the report will be shifted by the given amount. This option can be used in conjunction with the <code class="docutils literal notranslate"><span class="pre">--ts-normalize</span></code> option, in which case both adjustments will be applied. This option can be used to “hand-align” report files captured at different times, or reports captured on distributed systems with poorly synchronized system clocks. This option only applies to Arrow, HDF5, and SQLite exports.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-launch-command-switch-options">
<h4>CLI Launch Command Switch Options<a class="headerlink" href="#cli-launch-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">launch</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">launch</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="o"><</span><span class="n">application</span><span class="o">></span><span class="w"> </span><span class="p">[</span><span class="n">application</span><span class="o">-</span><span class="n">arguments</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>-b</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--backtrace</span></code></p></td>
<td><p>auto,fp,lbr,dwarf,none</p></td>
<td></td>
<td><p>Select the backtrace method to use while sampling. The option ‘lbr’ uses Intel(c) Corporation’s Last
Branch Record registers, available only with Intel(c) CPUs codenamed Haswell and later. The option ‘fp’
is frame pointer and assumes that frame pointers were enabled during compilation. The option ‘dwarf’ uses
DWARF’s CFI (Call Frame Information). Setting the value to ‘none’ can reduce collection overhead.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--clock-frequency-changes</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect clock frequency changes. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-cluster-events</span></code></p></td>
<td><p>0x16, 0x17, …, none</p></td>
<td><p>none</p></td>
<td><p>Collect per-cluster Uncore PMU counters. Multiple values can be selected, separated by commas only (no
spaces). Use the <code class="docutils literal notranslate"><span class="pre">--cpu-cluster-events=help</span></code> switch to see the full list of values. Available in
Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--command-file</span></code></p></td>
<td><p>< filename ></p></td>
<td><p>none</p></td>
<td><p>Open a file that contains profile switches and parse the switches. Note additional switches on the command
line will override switches in the file. This flag can be specified more than once.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-core-events</span></code> (Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>0x11,0x13,…,none</p></td>
<td><p>none</p></td>
<td><p>Collect per-core PMU counters. Multiple values can be selected, separated by commas only (no spaces). Use
the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> switch to see the full list of values.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-socket-events</span></code></p></td>
<td><p>0x2a, 0x2c, …, none</p></td>
<td><p>none</p></td>
<td><p>Collect per-socket Uncore PMU counters. Multiple values can be selected, separated by commas only (no spaces).
Use the <code class="docutils literal notranslate"><span class="pre">--cpu-socket-events=help</span></code> switch to see the full list of values. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code></p></td>
<td><p>process-tree, system-wide,
none</p></td>
<td><p>process-tree</p></td>
<td><p>Trace OS thread scheduling activity. Select ‘none’ to disable tracing CPU context switches. Depending on
the platform, some values may require admin or root privileges. Note: if the <code class="docutils literal notranslate"><span class="pre">--sample</span></code> switch is set
to a value other than ‘none’, the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> setting is hardcoded to the same value as the <code class="docutils literal notranslate"><span class="pre">--sample</span></code>
switch. If <code class="docutils literal notranslate"><span class="pre">--sample=none</span></code> and a target application is launched, the default is ‘process-tree’, otherwise
the default is ‘none’. Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch in Nsight Systems Embedded Platforms Edition</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-flush-interval</span></code></p></td>
<td><p>milliseconds</p></td>
<td><p>See Description</p></td>
<td><p>Set the interval, in milliseconds, when buffered CUDA data is automatically saved to storage. CUDA data
buffer saves may cause profiler overhead. Buffer save behavior can be controlled with this switch. If the
CUDA flush interval is set to 0 on systems running CUDA 11.0 or newer, buffers are saved when they fill.
If a flush interval is set to a non-zero value on such systems, buffers are saved only when the flush
interval expires. If a flush interval is set and the profiler runs out of available buffers before the
flush interval expires, additional buffers will be allocated as needed. In this case, setting a flush
interval can reduce buffer save overhead but increase memory use by the profiler. If the flush interval
is set to 0 on systems running older versions of CUDA, buffers are saved at the end of the collection. If
the profiler runs out of available buffers, additional buffers are allocated as needed. If a flush interval
is set to a non-zero value on such systems, buffers are saved when the flush interval expires. A
cuCtxSynchronize call may be inserted into the workflow before the buffers are saved which will cause
application overhead. In this case, setting a flush interval can reduce memory use by the profiler but may
increase save overhead. For collections over 30 seconds an interval of 10 seconds is recommended. Default is
10000 for Nsight Systems Embedded Platforms Edition and 0 otherwise.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-memory-usage</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Track the GPU memory usage by CUDA kernels. Applicable only when CUDA tracing is enabled. Note: This feature
may cause significant runtime overhead.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-um-cpu-page-faults</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>This switch tracks the page faults that occur when CPU code tries to access a memory page that resides on the
device. Note that this feature may cause significant runtime overhead. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-um-gpu-page-faults</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>This switch tracks the page faults that occur when GPU code tries to access a memory page that resides on the
host. Note that this feature may cause significant runtime overhead. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cudabacktrace</span></code></p></td>
<td><p>all, none, kernel, memory,
sync, other</p></td>
<td><p>none</p></td>
<td><p>When tracing CUDA APIs, enable the collection of a backtrace when a CUDA API is invoked. Significant runtime
overhead may occur. Values may be combined using ‘,’. Each value except ‘none’ may be appended with a
threshold after ‘:’. Threshold is duration, in nanoseconds, that CUDA APIs must execute before backtraces
are collected, e.g. ‘kernel:500’. Default value for each threshold is 1000ns (1us). Note: CPU sampling must
be enabled. Note: Not available on IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-graph-trace</span></code></p></td>
<td><p>graph, node</p></td>
<td><p>graph</p></td>
<td><p>If ‘graph’ is selected, CUDA graphs will be traced as a whole and node activities will not be collected. This
will reduce overhead to a minimum, but requires CUDA driver version 515.43 or higher. If ‘node’ is selected,
node activities will be collected, but CUDA graphs will not be traced as a whole. This may cause significant
runtime overhead. Default is ‘graph’ if available, otherwise default is ‘node’.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx-force-declare-adapter-removal-support</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>The Nsight Systems trace initialization involves creating a D3D device and discarding it. Enabling this flag
makes a call to DXGIDeclareAdapterRemovalSupport() before device creation. Requires DX11 or DX12 trace to be
enabled.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx12-gpu-workload</span></code></p></td>
<td><p>true, false, individual,
batch, none</p></td>
<td><p>individual</p></td>
<td><p>If individual or true, trace each DX12 workload’s GPU activity individually. If batch, trace DX12 workloads’
GPU activity in ExecuteCommandLists call batches. If none or false, do not trace DX12 workloads’ GPU
activity. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=dx12</span></code> is specified. This option is only
supported on Windows targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx12-wait-calls</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, trace wait calls that block on fences for DX12. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--trace=dx12</span></code> is specified. This option is only supported on Windows targets.</p></td>
</tr>
<tr class="row-even"><td><p>-e</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--env-var</span></code></p></td>
<td><p>A=B</p></td>
<td><p>NA</p></td>
<td><p>Set environment variable(s) for the application process to be launched. Environment variables should be
defined as A=B. Multiple environment variables can be specified as A=B,C=D.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag
is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--hotkey-capture</span></code></p></td>
<td><p>‘F1’ to ‘F12’</p></td>
<td><p>‘F12’</p></td>
<td><p>Hotkey to trigger the profiling session. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--capture-range=hotkey</span></code> is specified.</p></td>
</tr>
<tr class="row-odd"><td><p>-n</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--inherit-environment</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>When true, the current environment variables and the tool’s environment variables will be specified
for the launched process. When false, only the tool’s environment variables will be specified for the
launched process.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--injection-use-detours</span></code></p></td>
<td><p>true,false</p></td>
<td><p>true</p></td>
<td><p>Use detours for injection. If false, process injection will be performed by windows hooks which
allows it to bypass anti-cheat software.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--isr</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Trace Interrupt Service Routines (ISRs) and Deferred Procedure Calls (DPCs). Requires administrative
privileges. Available only on Windows devices.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code></p></td>
<td><p>openmpi,mpich</p></td>
<td><p>openmpi</p></td>
<td><p>When using <code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code> to trace MPI APIs use <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code> to specify which MPI implementation
the application is using. If no MPI implementation is specified, nsys tries to automatically detect
it based on the dynamic linker’s search path. If this fails, ‘openmpi’ is used. Calling <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code>
without <code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code> is not supported.</p></td>
</tr>
<tr class="row-odd"><td><p>-p</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-capture</span></code></p></td>
<td><p>range@domain, range, range@*</p></td>
<td><p>none</p></td>
<td><p>Specify NVTX range and domain to trigger the profiling session. This option is applicable only when
used along with <code class="docutils literal notranslate"><span class="pre">--capture-range=nvtx</span></code>.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code></p></td>
<td><p>default, <domain_names></p></td>
<td></td>
<td><p>Choose to exclude NVTX events from a comma separated list of domains. ‘default’ excludes NVTX events
without a domain. A domain with this name or commas in a domain name must be escaped with ‘\’.
Note: Only one of <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code> and <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code> can be used. This option
is only applicable when <code class="docutils literal notranslate"><span class="pre">--trace=nvtx</span></code> is specified.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code></p></td>
<td><p>default, <domain_names></p></td>
<td></td>
<td><p>Choose to only include NVTX events from a comma separated list of domains. ‘default’ filters the
NVTX default domain. A domain with this name or commas in a domain name must be escaped with ‘\’.
Note: Only one of <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code> and <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code> can be used. This option
is only applicable when <code class="docutils literal notranslate"><span class="pre">--trace=nvtx</span></code> is specified.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-nvtx-annotations</span></code></p></td>
<td><p><json_file></p></td>
<td></td>
<td><p>Specify the path to the JSON file containing the requested NVTX annotations.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--opengl-gpu-workload</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, trace the OpenGL workloads’ GPU activity. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--trace=opengl</span></code> is specified. This option is not supported on IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--os-events</span></code></p></td>
<td><p>‘help’ or the end users
selected events in the
format ‘x,y’</p></td>
<td></td>
<td><p>Select the OS events to sample. Use the <code class="docutils literal notranslate"><span class="pre">--os-events=help</span></code> switch to see the full list of events.
Multiple values can be selected, separated by commas only (no spaces). Use the <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code>
switch to enable. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-depth</span></code></p></td>
<td><p>integer</p></td>
<td><p>24</p></td>
<td><p>Set the depth for the backtraces collected for OS runtime libraries calls.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-stack-size</span></code></p></td>
<td><p>integer</p></td>
<td><p>6144</p></td>
<td><p>Set the stack dump size, in bytes, to generate backtraces for OS runtime libraries calls.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-threshold</span></code></p></td>
<td><p>nanoseconds</p></td>
<td><p>80000</p></td>
<td><p>Set the duration, in nanoseconds, that all OS runtime libraries calls must execute before backtraces
are collected.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-threshold</span></code></p></td>
<td><p>< nanoseconds ></p></td>
<td><p>1000 ns</p></td>
<td><p>Set the duration, in nanoseconds, that Operating System Runtime (osrt) APIs must execute before
they are traced. Values significantly less than 1000 may cause significant overhead and result in
extremely large result files. Note: Not available for IBM Power targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-backtrace</span></code></p></td>
<td><p>cuda, none, false</p></td>
<td><p>none</p></td>
<td><p>Collect Python backtrace event when tracing the selected API’s trigger. This option is supported
on Arm server (SBSA) platforms and x86 Linux targets. Note: the selected API tracing must be
enabled. For example, <code class="docutils literal notranslate"><span class="pre">--cudabacktrace</span></code> must be set when using <code class="docutils literal notranslate"><span class="pre">--python-backtrace=cuda</span></code>.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-sampling</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect Python backtrace sampling events. This option is supported on Arm server (SBSA) platforms,
x86 Linux and Windows targets. Note: When profiling Python-only workflows, consider disabling the
CPU sampling option to reduce overhead.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-sampling-frequency</span></code></p></td>
<td><p>1 < integers < 2000</p></td>
<td><p>1000</p></td>
<td><p>Specify the Python sampling frequency. The minimum supported frequency is 1Hz. The maximum
supported frequency is 2KHz. This option is ignored if the <code class="docutils literal notranslate"><span class="pre">--python-sampling</span></code> option is set
to false.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events</span></code></p></td>
<td><p>class/event,event,
class/event:mode,
class:mode,help,none</p></td>
<td><p>none</p></td>
<td><p>Multiple values can be selected, separated by commas only (no spaces). See the
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events-mode</span></code> switch description for ‘:mode’ format. Use the
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events=help</span></code> switch to see the full list of values. Example:
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events=8/1:system:wide,_NTO_TRACE_THREAD:process:fast,</span>
<span class="pre">\_NTO_TRACE_KERCALLENTER/\__KER_BAD,_NTO_TRACE_COMM,13</span></code>.
Collect QNX kernel events.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events-mode</span></code></p></td>
<td><p>system,process,fast,wide</p></td>
<td><p>system:fast</p></td>
<td><p>Values are separated by a colon (‘:’) only (no spaces). ‘system’ and ‘process’ cannot be specified
at the same time. ‘fast’ and ‘wide’ cannot be specified at the same time. Please check the QNX
documentation to determine when to select the ‘fast’ or ‘wide’ mode. Specify the default mode
for QNX kernel events collection.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--resolve-symbols</span></code></p></td>
<td><p>true,false</p></td>
<td><p>true</p></td>
<td><p>Resolve symbols of captured samples and backtraces.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--retain-etw-files</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Retain ETW files generated by the trace, merge and move the files to the output directory.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--run-as</span></code></p></td>
<td><p>< username ></p></td>
<td><p>none</p></td>
<td><p>Run the target application as the specified username. If not specified, the target application will
be run by the same user as Nsight Systems. Requires root privileges. Available for Linux targets only.</p></td>
</tr>
<tr class="row-odd"><td><p>-s</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sample</span></code></p></td>
<td><p>process-tree, system-wide,
none</p></td>
<td><p>process-tree</p></td>
<td><p>Select how to collect CPU IP/backtrace samples. If ‘none’ is selected, CPU sampling is disabled.
Depending on the platform, some values may require admin or root privileges. If a target application
is launched, the default is ‘process-tree’, otherwise, the default is ‘none’. Note: ‘system-wide’
is not available on all platforms. Note: If set to ‘none’, CPU context switch data will still be
collected unless the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> switch is set to ‘none’.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--samples-per-backtrace</span></code></p></td>
<td><p>integer <= 32</p></td>
<td><p>1</p></td>
<td><p>The number of CPU IP samples collected for every CPU IP/backtrace sample collected. For example,
if set to 4, on the fourth CPU IP sample collected, a backtrace will also be collected. Lower
values increase the amount of data collected. Higher values can reduce collection overhead and
reduce the number of CPU IP samples dropped. If DWARF backtraces are collected, the default is 4,
otherwise the default is 1. This option is not available on Nsight Systems Embedded Platforms Edition or on non-Linux targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-frequency</span></code></p></td>
<td><p>100 < integers < 8000</p></td>
<td><p>1000</p></td>
<td><p>Specify the sampling/backtracing frequency. The minimum supported frequency is 100 Hz. The maximum
supported frequency is 8000 Hz. This option is supported only on QNX, Linux for Tegra, and Windows
targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of CPU Cycle events counted before a CPU instruction pointer (IP) sample is collected.
If configured, backtraces may also be collected. The smaller the sampling period, the higher the
sampling rate. Note that smaller sampling periods will increase overhead and significantly increase
the size of the result file(s). Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (not Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of events counted before a CPU instruction pointer (IP) sample is collected. The event
used to trigger the collection of a sample is determined dynamically. For example, on Intel based
platforms, it will probably be “Reference Cycles” and on AMD platforms, “CPU Cycles”. If configured,
backtraces may also be collected. The smaller the sampling period, the higher the sampling rate.
Note that smaller sampling periods will increase overhead and significantly increase the size of
the result file(s). This option is available only on Linux targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-trigger</span></code></p></td>
<td><p>timer, sched, perf, cuda</p></td>
<td><p>timer,sched</p></td>
<td><p>Specify backtrace collection trigger. Multiple APIs can be selected, separated by commas only
(no spaces). Available on Nsight Systems Embedded Platforms Edition targets only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p>session identifier</p></td>
<td><p>none</p></td>
<td><p>Launch the application in the indicated session. The option argument must represent a valid session
name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with
the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the
system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session-new</span></code></p></td>
<td><p>[a-Z][0-9,a-Z,spaces]</p></td>
<td><p>profile-<id>-<application></p></td>
<td><p>Name the session created by the command. Name must start with an alphabetical character followed by
printable or space characters. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the
environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any
<code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p>-w</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--show-output</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, send target process’s stdout and stderr streams to both the console and stdout/stderr files
which are added to the report file. If false, only send target process stdout and stderr streams to
the stdout/stderr files which are added to the report file.</p></td>
</tr>
<tr class="row-even"><td><p>-t</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace</span></code></p></td>
<td><p>cuda, nvtx, cublas,
cublas-verbose, cusparse,
cusparse-verbose, cudnn,
cudla, cudla-verbose,
cusolver, cusolver-verbose,
opengl, opengl-annotations,
openacc, openmp, osrt, mpi,
nvvideo, vulkan,
vulkan-annotations, dx11,
dx11-annotations, dx12,
dx12-annotations, oshmem, ucx,
wddm, tegra-accelerators,
python-gil, none</p></td>
<td><p>cuda, opengl, nvtx, osrt</p></td>
<td><p>Select the API(s) to be traced. The osrt switch controls the OS runtime libraries tracing. Multiple
APIs can be selected, separated by commas only (no spaces). Since OpenACC and cuXXX APIs
are tightly linked with CUDA, selecting one of those APIs will automatically enable CUDA tracing.
cublas, cudla, cusparse and cusolver all have XXX-verbose options available.
Reflex SDK latency markers will be automatically collected when DX or vulkan API trace is enabled.
See information on <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code> option below if mpi is selected. If ‘<api>-annotations’ is
selected, the corresponding API will also be traced. If the none option is selected, no APIs are
traced and no other API can be selected. Note: cublas, cudnn, nvvideo, opengl, and vulkan are not
available on IBM Power target. cuddn is not available on Windows target.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace-fork-before-exec</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, trace any child process after fork and before they call one of the exec functions. Beware,
tracing in this interval relies on undefined behavior and might cause your application to crash or
deadlock. Note: This option is only available on Linux target platforms.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--vulkan-gpu-workload</span></code></p></td>
<td><p>true, false, individual,
batch, none</p></td>
<td><p>individual</p></td>
<td><p>If individual or true, trace each Vulkan workload’s GPU activity individually. If batch, trace
Vulkan workloads’ GPU activity in vkQueueSubmit call batches. If none or false, do not trace Vulkan
workloads’ GPU activity. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=vulkan</span></code> is
specified. This option is not supported on QNX.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wait</span></code></p></td>
<td><p>primary,all</p></td>
<td><p>all</p></td>
<td><p>If primary, the CLI will wait on the application process termination. If all, the CLI will
additionally wait on re-parented processes created by the application.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wddm-additional-events</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, collect additional range of ETW events, including context status, allocations, sync wait
and signal events, etc. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=wddm</span></code> is specified.
This option is only supported on Windows targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wddm-backtraces</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, collect backtraces of WDDM events. Disabling this data collection can reduce overhead for
certain target applications. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=wddm</span></code> is
specified. This option is only supported on Windows targets.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-profile-command-switch-options">
<h4>CLI Profile Command Switch Options<a class="headerlink" href="#cli-profile-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">profile</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="o"><</span><span class="n">application</span><span class="o">></span><span class="w"> </span><span class="p">[</span><span class="n">application</span><span class="o">-</span><span class="n">arguments</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--accelerator-trace</span></code></p></td>
<td><p>none,tegra-accelerators</p></td>
<td><p>none</p></td>
<td><p>Collect other accelerators workload trace from the hardware engine units. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--auto-report-name</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Derive report file name from collected data uses details of profiled graphics application. Format:
[Process Name][GPU Name][Window Resolution][Graphics API] Timestamp .nsys-rep If true, automatically
generate report file names.</p></td>
</tr>
<tr class="row-even"><td><p>-b</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--backtrace</span></code></p></td>
<td><p>auto,fp,lbr,dwarf,none</p></td>
<td></td>
<td><p>Select the backtrace method to use while sampling. The option ‘lbr’ uses Intel(c) Corporation’s Last
Branch Record registers, available only with Intel(c) CPUs codenamed Haswell and later. The option ‘fp’
is frame pointer and assumes that frame pointers were enabled during compilation. The option ‘dwarf’ uses
DWARF’s CFI (Call Frame Information). Setting the value to ‘none’ can reduce collection overhead.</p></td>
</tr>
<tr class="row-odd"><td><p>-c</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--capture-range</span></code></p></td>
<td><p>none, cudaProfilerApi,
hotkey, nvtx</p></td>
<td><p>none</p></td>
<td><p>When <code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> is used, profiling will start only when appropriate start API or hotkey is
invoked. If <code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> is set to none, start/stop API calls and hotkeys will be ignored. Note:
Hotkey works for graphic applications only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--capture-range-end</span></code></p></td>
<td><p>none, stop, stop-shutdown,
repeat[:N], repeat-shutdown:N</p></td>
<td><p>stop-shutdown</p></td>
<td><p>Specify the desired behavior when a capture range ends. Applicable only when used along with
<code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> option. If <code class="docutils literal notranslate"><span class="pre">none</span></code>, capture range end will be ignored. If <code class="docutils literal notranslate"><span class="pre">stop</span></code>, collection
will stop at capture range end. Any subsequent capture ranges will be ignored. Target app will continue
running. If <code class="docutils literal notranslate"><span class="pre">stop-shutdown</span></code>, collection will stop at capture range end and session will be shutdown. If
<code class="docutils literal notranslate"><span class="pre">repeat[:N]</span></code>, collection will stop at capture range end and subsequent capture ranges will trigger more
collections. Use the optional <code class="docutils literal notranslate"><span class="pre">:N</span></code> to specify max number of capture ranges to be honored. Any subsequent
capture ranges will be ignored once N capture ranges are collected. If <code class="docutils literal notranslate"><span class="pre">repeat-shutdown:N</span></code>, same behavior
as <code class="docutils literal notranslate"><span class="pre">repeat:N</span></code> but session will be shutdown after N ranges. For <code class="docutils literal notranslate"><span class="pre">stop-shutdown</span></code> and <code class="docutils literal notranslate"><span class="pre">repeat-shutdown:N</span></code>,
as always, use <code class="docutils literal notranslate"><span class="pre">--kill</span></code> option to specify whether target app should be terminated when shutting down session.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--clock-frequency-changes</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect clock frequency changes. Available only in Nsight Systems Embedded Platforms Edition and Arm server (SBSA) platforms</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--command-file</span></code></p></td>
<td><p>< filename ></p></td>
<td><p>none</p></td>
<td><p>Open a file that contains profile switches and parse the switches. Note additional switches on the command
line will override switches in the file. This flag can be specified more than once.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-cluster-events</span></code></p></td>
<td><p>0x16, 0x17, …, none</p></td>
<td><p>none</p></td>
<td><p>Collect per-cluster Uncore PMU counters. Multiple values can be selected, separated by commas only (no
spaces). Use the <code class="docutils literal notranslate"><span class="pre">--cpu-cluster-events=help</span></code> switch to see the full list of values. Available in
Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-core-events</span></code> (Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>0x11,0x13,…,none</p></td>
<td><p>none</p></td>
<td><p>Collect per-core PMU counters. Multiple values can be selected, separated by commas only (no spaces). Use
the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> switch to see the full list of values.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-core-events</span></code> (not Nsight Systems Embedded Platforms Edition )</p></td>
<td><p>‘help’ or the end users
selected events in the format
‘x,y’</p></td>
<td><p>‘2’ i.e. Instructions Retired</p></td>
<td><p>Select the CPU Core events to sample. Use the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> switch to see the full list of
events and the number of events that can be collected simultaneously. Multiple values can be selected,
separated by commas only (no spaces). Use the <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code> switch to enable.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-socket-events</span></code></p></td>
<td><p>0x2a, 0x2c, …, none</p></td>
<td><p>none</p></td>
<td><p>Collect per-socket Uncore PMU counters. Multiple values can be selected, separated by commas only (no spaces).
Use the <code class="docutils literal notranslate"><span class="pre">--cpu-socket-events=help</span></code> switch to see the full list of values. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code></p></td>
<td><p>process-tree, system-wide,
none</p></td>
<td><p>process-tree</p></td>
<td><p>Trace OS thread scheduling activity. Select ‘none’ to disable tracing CPU context switches. Depending on
the platform, some values may require admin or root privileges. Note: if the <code class="docutils literal notranslate"><span class="pre">--sample</span></code> switch is set
to a value other than ‘none’, the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> setting is hardcoded to the same value as the <code class="docutils literal notranslate"><span class="pre">--sample</span></code>
switch. If <code class="docutils literal notranslate"><span class="pre">--sample=none</span></code> and a target application is launched, the default is ‘process-tree’, otherwise
the default is ‘none’. Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch in Nsight Systems Embedded Platforms Edition</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-flush-interval</span></code></p></td>
<td><p>milliseconds</p></td>
<td><p>See Description</p></td>
<td><p>Set the interval, in milliseconds, when buffered CUDA data is automatically saved to storage. CUDA data
buffer saves may cause profiler overhead. Buffer save behavior can be controlled with this switch. If the
CUDA flush interval is set to 0 on systems running CUDA 11.0 or newer, buffers are saved when they fill.
If a flush interval is set to a non-zero value on such systems, buffers are saved only when the flush
interval expires. If a flush interval is set and the profiler runs out of available buffers before the
flush interval expires, additional buffers will be allocated as needed. In this case, setting a flush
interval can reduce buffer save overhead but increase memory use by the profiler. If the flush interval
is set to 0 on systems running older versions of CUDA, buffers are saved at the end of the collection. If
the profiler runs out of available buffers, additional buffers are allocated as needed. If a flush interval
is set to a non-zero value on such systems, buffers are saved when the flush interval expires. A
cuCtxSynchronize call may be inserted into the workflow before the buffers are saved which will cause
application overhead. In this case, setting a flush interval can reduce memory use by the profiler but may
increase save overhead. For collections over 30 seconds an interval of 10 seconds is recommended. Default is
10000 for Nsight Systems Embedded Platforms Edition and 0 otherwise.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-graph-trace</span></code></p></td>
<td><p>graph, node</p></td>
<td><p>graph</p></td>
<td><p>If ‘graph’ is selected, CUDA graphs will be traced as a whole and node activities will not be collected. This
will reduce overhead to a minimum, but requires CUDA driver version 515.43 or higher. If ‘node’ is selected,
node activities will be collected, but CUDA graphs will not be traced as a whole. This may cause significant
runtime overhead. Default is ‘graph’ if available, otherwise default is ‘node’.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-memory-usage</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Track the GPU memory usage by CUDA kernels. Applicable only when CUDA tracing is enabled. Note: This feature
may cause significant runtime overhead.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-um-cpu-page-faults</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>This switch tracks the page faults that occur when CPU code tries to access a memory page that resides on the
device. Note that this feature may cause significant runtime overhead. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cuda-um-gpu-page-faults</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>This switch tracks the page faults that occur when GPU code tries to access a memory page that resides on the
host. Note that this feature may cause significant runtime overhead. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cudabacktrace</span></code></p></td>
<td><p>all, none, kernel, memory,
sync, other</p></td>
<td><p>none</p></td>
<td><p>When tracing CUDA APIs, enable the collection of a backtrace when a CUDA API is invoked. Significant runtime
overhead may occur. Values may be combined using ‘,’. Each value except ‘none’ may be appended with a
threshold after ‘:’. Threshold is duration, in nanoseconds, that CUDA APIs must execute before backtraces
are collected, e.g. ‘kernel:500’. Default value for each threshold is 1000ns (1us). Note: CPU sampling must
be enabled. Note: Not available on IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td><p>-y</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--delay</span></code></p></td>
<td><p>< seconds ></p></td>
<td><p>0</p></td>
<td><p>Collection start delay in seconds.</p></td>
</tr>
<tr class="row-odd"><td><p>-d</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--duration</span></code></p></td>
<td><p>< seconds ></p></td>
<td><p>NA</p></td>
<td><p>Collection duration in seconds, duration must be greater than zero. The launched process will be terminated
when the specified profiling duration expires unless the user specifies the <code class="docutils literal notranslate"><span class="pre">--kill</span></code> none option (details
below).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--duration-frames</span></code></p></td>
<td><p>60 <= integer</p></td>
<td></td>
<td><p>Stop the recording session after this many frames have been captured. Note when it is selected cannot include
any other stop options. If not specified, the default is disabled.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx-force-declare-adapter-removal-support</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>The Nsight Systems trace initialization involves creating a D3D device and discarding it. Enabling this flag
makes a call to DXGIDeclareAdapterRemovalSupport() before device creation. Requires DX11 or DX12 trace to be
enabled.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx12-gpu-workload</span></code></p></td>
<td><p>true, false, individual,
batch, none</p></td>
<td><p>individual</p></td>
<td><p>If individual or true, trace each DX12 workload’s GPU activity individually. If batch, trace DX12 workloads’
GPU activity in ExecuteCommandLists call batches. If none or false, do not trace DX12 workloads’ GPU
activity. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=dx12</span></code> is specified. This option is only
supported on Windows targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--dx12-wait-calls</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, trace wait calls that block on fences for DX12. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--trace=dx12</span></code> is specified. This option is only supported on Windows targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--el1-sampling</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Enable EL1 sampling. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--el1-sampling-config</span></code></p></td>
<td><p>< filepath config.json ></p></td>
<td><p>none</p></td>
<td><p>EL1 sampling config. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td><p>-e</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--env-var</span></code></p></td>
<td><p>A=B</p></td>
<td><p>NA</p></td>
<td><p>Set environment variable(s) for the application process to be launched. Environment variables should be
defined as A=B. Multiple environment variables can be specified as A=B,C=D.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--etw-provider</span></code></p></td>
<td><p>“<name>,<guid>”, or path
to JSON file</p></td>
<td><p>none</p></td>
<td><p>Add custom ETW trace provider(s). If you want to specify more attributes than Name and GUID, provide a JSON
configuration file as as outlined below. This switch can be used multiple times to add multiple providers.
Note: Only available for Windows targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--event-sample</span></code></p></td>
<td><p>system-wide, none</p></td>
<td><p>none</p></td>
<td><p>Use the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> and the <code class="docutils literal notranslate"><span class="pre">--os-events=help</span></code> switches to see the full list of events. If
event sampling is enabled and no events are selected, the CPU Core event ‘Instructions Retired’ is selected
by default. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--event-sampling-frequency</span></code></p></td>
<td><p>Integers from 1 to 20 Hz</p></td>
<td><p>3</p></td>
<td><p>The sampling frequency used to collect event counts. Minimum event sampling frequency is 1 Hz. Maximum event
sampling frequency is 20 Hz. Not available in Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--export</span></code></p></td>
<td><p>arrow, hdf, json, sqlite,
text, none</p></td>
<td><p>none</p></td>
<td><p>Create additional output file(s) based on the data collected. This option can be given more than once.
WARNING: If the collection captures a large amount of data, creating the export file may take several
minutes to complete.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--flush-on-cudaprofilerstop</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If set to true, any call to cudaProfilerStop() will cause the CUDA trace buffers to be flushed. Note that
the CUDA trace buffers will be flushed when the collection ends, irrespective of the value of this switch.</p></td>
</tr>
<tr class="row-even"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, overwrite all existing result files with same output filename (.qdstrm, .nsys-rep, .arrows, .h5,
.json, .sqlite, .txt).</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ftrace</span></code></p></td>
<td></td>
<td></td>
<td><p>Collect ftrace events. Argument should list events to collect as: subsystem1/event1,subsystem2/event2.
Requires root. No ftrace events are collected by default. Note: Not available on IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ftrace-keep-user-config</span></code></p></td>
<td></td>
<td></td>
<td><p>Skip initial ftrace setup and collect already configured events. Default resets the ftrace configuration.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-device</span></code></p></td>
<td><p>GPU ID, help, all, none</p></td>
<td><p>none</p></td>
<td><p>Collect GPU Metrics from specified devices. Determine GPU IDs by using <code class="docutils literal notranslate"><span class="pre">--gpu-metrics-device=help</span></code>
switch.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-frequency</span></code></p></td>
<td><p>integer</p></td>
<td><p>10000</p></td>
<td><p>Specify GPU Metrics sampling frequency. Minimum supported frequency is 10 (Hz). Maximum supported
frequency is 200000 (Hz).</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-set</span></code></p></td>
<td><p>index, alias</p></td>
<td></td>
<td><p>Specify metric set for GPU Metrics. The argument must be one of indices or aliases reported by
<code class="docutils literal notranslate"><span class="pre">--gpu-metrics-set=help</span></code> switch. If not specified, the default is the first metric set that supports
all selected GPUs.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpuctxsw</span></code></p></td>
<td><p>true,false</p></td>
<td><p>false</p></td>
<td><p>Trace GPU context switches. Note that this requires driver r435.17 or later and root permission. Not
supported on IBM Power targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag
is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--hotkey-capture</span></code></p></td>
<td><p>‘F1’ to ‘F12’</p></td>
<td><p>‘F12’</p></td>
<td><p>Hotkey to trigger the profiling session. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--capture-range=hotkey</span></code> is specified.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-device</span></code></p></td>
<td><p><IB switch GUIDs></p></td>
<td><p>none</p></td>
<td><p>A comma-separated list of InfiniBand switch GUIDs. Collect InfiniBand switch congestion events from
switches identified by the specified GUIDs. This switch can be used multiple times. System scope. Use
the –ib-switch-congestion-nic-device, –ib-switch-congestion-percent, and
–ib-switch-congestion-threshold-high switches to further control how congestion events are collected.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-nic-device</span></code></p></td>
<td><p><NIC name></p></td>
<td><p>none</p></td>
<td><p>The name of the NIC (HCA) through which InfiniBand switches will be accessed. By default, the first
active NIC will be used. One way to find a NIC’s name is via the <cite>ibnetdiscover –Hca_list | grep
“$(hostname)”</cite> command. Example usage: –ib-switch-congestion-nic-device=mlx5_3</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-percent</span></code></p></td>
<td><p>1 <= integer <= 100</p></td>
<td><p>50</p></td>
<td><p>Percent of InfiniBand switch congestion events to be collected. This option enables reducing the
network bandwidth consumed by reporting congestion events.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-threshold-high</span></code></p></td>
<td><p>1 <= integer <= 1023</p></td>
<td><p>75</p></td>
<td><p>High threshold percentage for InfiniBand switch egress port buffer size. Before a packet leaves an
InfiniBand switch, it is stored at an egress port buffer. The buffer’s size is checked and if it
exceeds the given threshold percentage, a congestion event is reported. The percentage can be greater
than 100.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-metrics-device</span></code></p></td>
<td><p><IB switch GUIDs></p></td>
<td><p>none</p></td>
<td><p>A comma-separated list of InfiniBand switch GUIDs. Collect metrics from the specified InfiniBand
switches. This switch can be used multiple times. System scope.</p></td>
</tr>
<tr class="row-even"><td><p>-n</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--inherit-environment</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>When true, the current environment variables and the tool’s environment variables will be specified
for the launched process. When false, only the tool’s environment variables will be specified for the
launched process.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--injection-use-detours</span></code></p></td>
<td><p>true,false</p></td>
<td><p>true</p></td>
<td><p>Use detours for injection. If false, process injection will be performed by windows hooks which
allows it to bypass anti-cheat software.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--isr</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Trace Interrupt Service Routines (ISRs) and Deferred Procedure Calls (DPCs). Requires administrative
privileges. Available only on Windows devices.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--kill</span></code></p></td>
<td><p>none, sigkill, sigterm,
signal number</p></td>
<td><p>sigterm</p></td>
<td><p>Send signal to the target application’s process group. Can be used with <code class="docutils literal notranslate"><span class="pre">--duration</span></code> or range
markers.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code></p></td>
<td><p>openmpi,mpich</p></td>
<td><p>openmpi</p></td>
<td><p>When using <code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code> to trace MPI APIs use <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code> to specify which MPI implementation
the application is using. If no MPI implementation is specified, nsys tries to automatically detect
it based on the dynamic linker’s search path. If this fails, ‘openmpi’ is used. Calling <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code>
without <code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code> is not supported.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nic-metrics</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect metrics from supported NIC/HCA devices. System scope. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td><p>-p</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-capture</span></code></p></td>
<td><p>range@domain, range, range@*</p></td>
<td><p>none</p></td>
<td><p>Specify NVTX range and domain to trigger the profiling session. This option is applicable only when
used along with <code class="docutils literal notranslate"><span class="pre">--capture-range=nvtx</span></code>.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code></p></td>
<td><p>default, <domain_names></p></td>
<td></td>
<td><p>Choose to exclude NVTX events from a comma separated list of domains. ‘default’ excludes NVTX events
without a domain. A domain with this name or commas in a domain name must be escaped with ‘\’.
Note: Only one of <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code> and <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code> can be used. This option
is only applicable when <code class="docutils literal notranslate"><span class="pre">--trace=nvtx</span></code> is specified.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code></p></td>
<td><p>default, <domain_names></p></td>
<td></td>
<td><p>Choose to only include NVTX events from a comma separated list of domains. ‘default’ filters the
NVTX default domain. A domain with this name or commas in a domain name must be escaped with ‘\’.
Note: Only one of <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code> and <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code> can be used. This option
is only applicable when <code class="docutils literal notranslate"><span class="pre">--trace=nvtx</span></code> is specified.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-nvtx-annotations</span></code></p></td>
<td><p><json_file></p></td>
<td></td>
<td><p>Specify the path to the JSON file containing the requested NVTX annotations.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--opengl-gpu-workload</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, trace the OpenGL workloads’ GPU activity. Note that this switch is applicable only when
<code class="docutils literal notranslate"><span class="pre">--trace=opengl</span></code> is specified. This option is not supported on IBM Power targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--os-events</span></code></p></td>
<td><p>‘help’ or the end users
selected events in the
format ‘x,y’</p></td>
<td></td>
<td><p>Select the OS events to sample. Use the <code class="docutils literal notranslate"><span class="pre">--os-events=help</span></code> switch to see the full list of events.
Multiple values can be selected, separated by commas only (no spaces). Use the <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code>
switch to enable. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-depth</span></code></p></td>
<td><p>integer</p></td>
<td><p>24</p></td>
<td><p>Set the depth for the backtraces collected for OS runtime libraries calls.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-stack-size</span></code></p></td>
<td><p>integer</p></td>
<td><p>6144</p></td>
<td><p>Set the stack dump size, in bytes, to generate backtraces for OS runtime libraries calls.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-backtrace-threshold</span></code></p></td>
<td><p>nanoseconds</p></td>
<td><p>80000</p></td>
<td><p>Set the duration, in nanoseconds, that all OS runtime libraries calls must execute before backtraces
are collected.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--osrt-threshold</span></code></p></td>
<td><p>< nanoseconds ></p></td>
<td><p>1000 ns</p></td>
<td><p>Set the duration, in nanoseconds, that Operating System Runtime (osrt) APIs must execute before
they are traced. Values significantly less than 1000 may cause significant overhead and result in
extremely large result files. Note: Not available for IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p>< filename ></p></td>
<td><p>report#</p></td>
<td><p>Set report file name. Any %q{ENV_VAR} pattern in the filename will be substituted with the value
of the environment variable. Any %h pattern in the filename will be substituted with the hostname
of the system. Any %p pattern in the filename will be substituted with the PID of the target
process or the PID of the root process if there is a process tree. Any %% pattern in the filename
will be substituted with %. Default is report#.{qdstrm,nsys-rep,sqlite,h5,txt,arrows,json} in the
working directory.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--process-scope</span></code></p></td>
<td><p>main, process-tree,
system-wide</p></td>
<td><p>main</p></td>
<td><p>Select which process(es) to trace. Available in Nsight Systems Embedded Platforms Edition only. Nsight Systems Workstation Edition will always
trace system-wide in this version of the tool.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-backtrace</span></code></p></td>
<td><p>cuda, none, false</p></td>
<td><p>none</p></td>
<td><p>Collect Python backtrace event when tracing the selected API’s trigger. This option is supported
on Arm server (SBSA) platforms and x86 Linux targets. Note: the selected API tracing must be
enabled. For example, <code class="docutils literal notranslate"><span class="pre">--cudabacktrace</span></code> must be set when using <code class="docutils literal notranslate"><span class="pre">--python-backtrace=cuda</span></code>.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-sampling</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect Python backtrace sampling events. This option is supported on Arm server (SBSA) platforms,
x86 Linux and Windows targets. Note: When profiling Python-only workflows, consider disabling the
CPU sampling option to reduce overhead.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--python-sampling-frequency</span></code></p></td>
<td><p>1 < integers < 2000</p></td>
<td><p>1000</p></td>
<td><p>Specify the Python sampling frequency. The minimum supported frequency is 1Hz. The maximum
supported frequency is 2KHz. This option is ignored if the <code class="docutils literal notranslate"><span class="pre">--python-sampling</span></code> option is set
to false.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events</span></code></p></td>
<td><p>class/event,event,
class/event:mode,
class:mode,help,none</p></td>
<td><p>none</p></td>
<td><p>Multiple values can be selected, separated by commas only (no spaces). See the
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events-mode</span></code> switch description for ‘:mode’ format. Use the
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events=help</span></code> switch to see the full list of values. Example:
<code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events=8/1:system:wide,_NTO_TRACE_THREAD:process:fast,</span>
<span class="pre">\_NTO_TRACE_KERCALLENTER/\__KER_BAD,_NTO_TRACE_COMM,13</span></code>.
Collect QNX kernel events.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--qnx-kernel-events-mode</span></code></p></td>
<td><p>system,process,fast,wide</p></td>
<td><p>system:fast</p></td>
<td><p>Values are separated by a colon (‘:’) only (no spaces). ‘system’ and ‘process’ cannot be specified
at the same time. ‘fast’ and ‘wide’ cannot be specified at the same time. Please check the QNX
documentation to determine when to select the ‘fast’ or ‘wide’ mode. Specify the default mode
for QNX kernel events collection.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--resolve-symbols</span></code></p></td>
<td><p>true,false</p></td>
<td><p>true</p></td>
<td><p>Resolve symbols of captured samples and backtraces.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--retain-etw-files</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Retain ETW files generated by the trace, merge and move the files to the output directory.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--run-as</span></code></p></td>
<td><p>< username ></p></td>
<td><p>none</p></td>
<td><p>Run the target application as the specified username. If not specified, the target application will
be run by the same user as Nsight Systems. Requires root privileges. Available for Linux targets only.</p></td>
</tr>
<tr class="row-even"><td><p>-s</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sample</span></code></p></td>
<td><p>process-tree, system-wide,
none</p></td>
<td><p>process-tree</p></td>
<td><p>Select how to collect CPU IP/backtrace samples. If ‘none’ is selected, CPU sampling is disabled.
Depending on the platform, some values may require admin or root privileges. If a target application
is launched, the default is ‘process-tree’, otherwise, the default is ‘none’. Note: ‘system-wide’
is not available on all platforms. Note: If set to ‘none’, CPU context switch data will still be
collected unless the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> switch is set to ‘none’.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--samples-per-backtrace</span></code></p></td>
<td><p>integer <= 32</p></td>
<td><p>1</p></td>
<td><p>The number of CPU IP samples collected for every CPU IP/backtrace sample collected. For example,
if set to 4, on the fourth CPU IP sample collected, a backtrace will also be collected. Lower
values increase the amount of data collected. Higher values can reduce collection overhead and
reduce the number of CPU IP samples dropped. If DWARF backtraces are collected, the default is 4,
otherwise the default is 1. This option is not available on Nsight Systems Embedded Platforms Edition or on non-Linux targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-frequency</span></code></p></td>
<td><p>100 < integers < 8000</p></td>
<td><p>1000</p></td>
<td><p>Specify the sampling/backtracing frequency. The minimum supported frequency is 100 Hz. The maximum
supported frequency is 8000 Hz. This option is supported only on QNX, Linux for Tegra, and Windows
targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of CPU Cycle events counted before a CPU instruction pointer (IP) sample is collected.
If configured, backtraces may also be collected. The smaller the sampling period, the higher the
sampling rate. Note that smaller sampling periods will increase overhead and significantly increase
the size of the result file(s). Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (not Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of events counted before a CPU instruction pointer (IP) sample is collected. The event
used to trigger the collection of a sample is determined dynamically. For example, on Intel based
platforms, it will probably be “Reference Cycles” and on AMD platforms, “CPU Cycles”. If configured,
backtraces may also be collected. The smaller the sampling period, the higher the sampling rate.
Note that smaller sampling periods will increase overhead and significantly increase the size of
the result file(s). This option is available only on Linux targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-trigger</span></code></p></td>
<td><p>timer, sched, perf, cuda</p></td>
<td><p>timer,sched</p></td>
<td><p>Specify backtrace collection trigger. Multiple APIs can be selected, separated by commas only
(no spaces). Available on Nsight Systems Embedded Platforms Edition targets only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session-new</span></code></p></td>
<td><p>[a-Z][0-9,a-Z,spaces]</p></td>
<td><p>profile-<id>-<application></p></td>
<td><p>Name the session created by the command. Name must start with an alphabetical character followed by
printable or space characters. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the
environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any
<code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p>-w</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--show-output</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, send target process’s stdout and stderr streams to both the console and stdout/stderr files
which are added to the report file. If false, only send target process stdout and stderr streams to
the stdout/stderr files which are added to the report file.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics</span></code></p></td>
<td><p>true,false</p></td>
<td><p>false</p></td>
<td><p>Collect SOC Metrics. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-frequency</span></code></p></td>
<td><p>integer</p></td>
<td><p>10000</p></td>
<td><p>Specify SOC Metrics sampling frequency. Minimum supported frequency is ‘100’ (Hz). Maximum supported
frequency is ‘1000000’ (Hz). Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-set</span></code></p></td>
<td><p>see description</p></td>
<td><p>see description</p></td>
<td><p>Specify metric set for SOC Metrics sampling. The option argument must be one of indices or aliases
reported by <code class="docutils literal notranslate"><span class="pre">--soc-metrics-set=help</span></code> switch. Default is the first supported set. Available in
Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--start-frame-index</span></code></p></td>
<td><p>1 <= integer</p></td>
<td></td>
<td><p>Start the recording session when the frame index reaches the frame number preceding the start frame
index. Note when it is selected cannot include any other start options. If not specified, the default
is disabled.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--stats</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Generate summary statistics after the collection. WARNING: When set to true, an SQLite database will
be created after the collection. If the collection captures a large amount of data, creating the
database file may take several minutes to complete.</p></td>
</tr>
<tr class="row-odd"><td><p>-x</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--stop-on-exit</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, stop collecting automatically when the launched process has exited or when the duration
expires - whichever occurs first. If false, duration must be set and the collection stops only when
the duration expires. Nsight Systems does not officially support runs longer than 5 minutes.</p></td>
</tr>
<tr class="row-even"><td><p>-t</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace</span></code></p></td>
<td><p>cuda, nvtx, cublas,
cublas-verbose, cusparse,
cusparse-verbose, cudnn,
cudla, cudla-verbose,
cusolver, cusolver-verbose,
opengl, opengl-annotations,
openacc, openmp, osrt, mpi,
nvvideo, vulkan,
vulkan-annotations, dx11,
dx11-annotations, dx12,
dx12-annotations, oshmem, ucx,
wddm, tegra-accelerators,
python-gil, none</p></td>
<td><p>cuda, opengl, nvtx, osrt</p></td>
<td><p>Select the API(s) to be traced. The osrt switch controls the OS runtime libraries tracing. Multiple
APIs can be selected, separated by commas only (no spaces). Since OpenACC and cuXXX APIs
are tightly linked with CUDA, selecting one of those APIs will automatically enable CUDA tracing.
cublas, cudla, cusparse and cusolver all have XXX-verbose options available.
Reflex SDK latency markers will be automatically collected when DX or vulkan API trace is enabled.
See information on <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code> option below if mpi is selected. If ‘<api>-annotations’ is
selected, the corresponding API will also be traced. If the none option is selected, no APIs are
traced and no other API can be selected. Note: cublas, cudnn, nvvideo, opengl, and vulkan are not
available on IBM Power target. cuddn is not available on Windows target.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace-fork-before-exec</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, trace any child process after fork and before they call one of the exec functions. Beware,
tracing in this interval relies on undefined behavior and might cause your application to crash or
deadlock. Note: This option is only available on Linux target platforms.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--vsync</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect vsync events. If collection of vsync events is enabled, display/display_scanline ftrace
events will also be captured. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--vulkan-gpu-workload</span></code></p></td>
<td><p>true, false, individual,
batch, none</p></td>
<td><p>individual</p></td>
<td><p>If individual or true, trace each Vulkan workload’s GPU activity individually. If batch, trace
Vulkan workloads’ GPU activity in vkQueueSubmit call batches. If none or false, do not trace Vulkan
workloads’ GPU activity. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=vulkan</span></code> is
specified. This option is not supported on QNX.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wait</span></code></p></td>
<td><p>primary,all</p></td>
<td><p>all</p></td>
<td><p>If primary, the CLI will wait on the application process termination. If all, the CLI will
additionally wait on re-parented processes created by the application.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wddm-additional-events</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, collect additional range of ETW events, including context status, allocations, sync wait
and signal events, etc. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=wddm</span></code> is specified.
This option is only supported on Windows targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--wddm-backtraces</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, collect backtraces of WDDM events. Disabling this data collection can reduce overhead for
certain target applications. Note that this switch is applicable only when <code class="docutils literal notranslate"><span class="pre">--trace=wddm</span></code> is
specified. This option is only supported on Windows targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--xhv-trace</span></code></p></td>
<td><p>< filepath pct.json ></p></td>
<td><p>none</p></td>
<td><p>Collect hypervisor trace. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--xhv-trace-events</span></code></p></td>
<td><p>all, none, core, sched,
irq, trap</p></td>
<td><p>all</p></td>
<td><p>Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-sessions-command-switch-subcommands">
<h4>CLI Sessions Command Switch Subcommands<a class="headerlink" href="#cli-sessions-command-switch-subcommands" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">sessions</span></code> command switch, the following subcommands are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">sessions</span><span class="w"> </span><span class="p">[</span><span class="n">subcommand</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Subcommand</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>list</p></td>
<td><p>List all active sessions including ID, name, and state information</p></td>
</tr>
</tbody>
</table>
<section id="cli-sessions-list-command-switch-options">
<h5>CLI Sessions List Command Switch Options<a class="headerlink" href="#cli-sessions-list-command-switch-options" title="Permalink to this heading"></a></h5>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">sessions</span> <span class="pre">list</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys [global-options] sessions list [options]
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td><p>-p</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--show-header</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>Controls whether a header should appear in the output.</p></td>
</tr>
</tbody>
</table>
</section>
</section>
<section id="cli-shutdown-command-switch-options">
<h4>CLI Shutdown Command Switch Options<a class="headerlink" href="#cli-shutdown-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">shutdown</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">shutdown</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--kill</span></code></p></td>
<td><p>On Linux: one, sigkill, sigterm, signal number</p>
<p>On Windows: true, false</p>
</td>
<td><p>On Linux: sigterm</p>
<p>On Windows: true</p>
</td>
<td><p>Send signal to the target application’s process group when shutting down session.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p>session identifier</p></td>
<td><p>none</p></td>
<td><p>Shutdown the indicated session. The option argument must represent a valid session name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-start-command-switch-options">
<h4>CLI Start Command Switch Options<a class="headerlink" href="#cli-start-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">start</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">start</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--accelerator-trace</span></code></p></td>
<td><p>none,tegra-accelerators</p></td>
<td><p>none</p></td>
<td><p>Collect other accelerators workload trace from the hardware engine units. Only available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td><p>-b</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--backtrace</span></code></p></td>
<td><p>auto,fp,lbr,dwarf,none</p></td>
<td></td>
<td><p>Select the backtrace method to use while sampling. The option ‘lbr’ uses Intel(c) Corporation’s Last Branch Record registers, available only with Intel(c) CPUs codenamed Haswell and later. The option ‘fp’ is frame pointer and assumes that frame pointers were enabled during compilation. The option ‘dwarf’ uses DWARF’s CFI (Call Frame Information). Setting the value to ‘none’ can reduce collection overhead.</p></td>
</tr>
<tr class="row-even"><td><p>-c</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--capture-range</span></code></p></td>
<td><p>none, cudaProfilerApi, hotkey, nvtx</p></td>
<td><p>none</p></td>
<td><p>When <code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> is used, profiling will start only when appropriate start API or hotkey is invoked. If <code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> is set to none, start/stop API calls and hotkeys will be ignored. Note: hotkey works for graphic applications only. CUDA or NVTX tracing must be enabled on the target application for ‘-c cudaProfilerApi’ or ‘-c nvtx’ to work.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--capture-range-end</span></code></p></td>
<td><p>none, stop, stop-shutdown, repeat[:N], repeat-shutdown:N</p></td>
<td><p>stop-shutdown</p></td>
<td><p>Specify the desired behavior when a capture range ends. Applicable only when used along with <code class="docutils literal notranslate"><span class="pre">--capture-range</span></code> option. If <code class="docutils literal notranslate"><span class="pre">none</span></code>, capture range end will be ignored. If <code class="docutils literal notranslate"><span class="pre">stop</span></code>, collection will stop at capture range end. Any subsequent capture ranges will be ignored. Target app will continue running. If <code class="docutils literal notranslate"><span class="pre">stop-shutdown</span></code>, collection will stop at capture range end and session will be shutdown. If <code class="docutils literal notranslate"><span class="pre">repeat[:N]</span></code>, collection will stop at capture range end and subsequent capture ranges will trigger more collections.
Use the optional <code class="docutils literal notranslate"><span class="pre">:N</span></code> to specify max number of capture ranges to be honored. Any subsequent capture ranges will be ignored once N capture ranges are collected. If <code class="docutils literal notranslate"><span class="pre">repeat-shutdown:N</span></code>, same behavior as <code class="docutils literal notranslate"><span class="pre">repeat:N</span></code> but session will be shutdown after N ranges. For <code class="docutils literal notranslate"><span class="pre">stop-shutdown</span></code> and <code class="docutils literal notranslate"><span class="pre">repeat-shutdown:N</span></code>, as always use <code class="docutils literal notranslate"><span class="pre">--kill</span></code> option to specify whether target app should be terminated when shutting down session.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpu-core-events</span></code> (not Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>‘help’ or the end users selected events in the format ‘x,y’</p></td>
<td><p>‘2’ i.e. Instructions Retired</p></td>
<td><p>Select the CPU Core events to sample. Use the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> switch to see the full list of events and the number of events that can be collected simultaneously. Multiple values can be selected, separated by commas only (no spaces). Use the <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code> switch to enable.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code></p></td>
<td><p>process-tree, system-wide, none</p></td>
<td><p>process-tree</p></td>
<td><p>Trace OS thread scheduling activity. Select ‘none’ to disable tracing CPU context switches. Depending on the platform, some values may require admin or root privileges. Note: if the <code class="docutils literal notranslate"><span class="pre">--sample</span></code> switch is set to a value other than ‘none’, the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> setting is hardcoded to the same value as the <code class="docutils literal notranslate"><span class="pre">--sample</span></code> switch. If <code class="docutils literal notranslate"><span class="pre">--sample=none</span></code> and a target application is launched, the default is ‘process-tree’, otherwise the default is ‘none’. Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch in Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--el1-sampling</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Enable EL1 sampling. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--el1-sampling-config</span></code></p></td>
<td><p>< filepath config.json ></p></td>
<td><p>none</p></td>
<td><p>EL1 sampling config. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--etw-provider</span></code></p></td>
<td><p>“<name>,<guid>”, or path to JSON file</p></td>
<td><p>none</p></td>
<td><p>Add custom ETW trace provider(s). If you want to specify more attributes than Name and GUID, provide a JSON configuration file as as outlined below. This switch can be used multiple times to add multiple providers. Note: Only available for Windows targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--event-sample</span></code></p></td>
<td><p>system-wide, none</p></td>
<td><p>none</p></td>
<td><p>Use the <code class="docutils literal notranslate"><span class="pre">--cpu-core-events=help</span></code> and the <code class="docutils literal notranslate"><span class="pre">--os-events=help</span></code> switches to see the full list of events. If event sampling is enabled and no events are selected, the CPU Core event ‘Instructions Retired’ is selected by default. Not available in Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--event-sampling-frequency</span></code></p></td>
<td><p>Integers from 1 to 20 Hz</p></td>
<td><p>3</p></td>
<td><p>The sampling frequency used to collect event counts. Minimum event sampling frequency is 1 Hz. Maximum event sampling frequency is 20 Hz. Not available in Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--export</span></code></p></td>
<td><p>arrow, hdf, json, sqlite, text, none</p></td>
<td><p>none</p></td>
<td><p>Create additional output file(s) based on the data collected. This option can be given more than once. WARNING: If the collection captures a large amount of data, creating the export file may take several minutes to complete.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--flush-on-cudaprofilerstop</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If set to true, any call to cudaProfilerStop() will cause the CUDA trace buffers to be flushed. Note that the CUDA trace buffers will be flushed when the collection ends, irrespective of the value of this switch.</p></td>
</tr>
<tr class="row-odd"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>If true, overwrite all existing result files with same output filename (.qdstrm, .nsys-rep, .arrows, .hdf, .json, .sqlite, .txt).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ftrace</span></code></p></td>
<td></td>
<td></td>
<td><p>Collect ftrace events. Argument should list events to collect as: subsystem1/event1,subsystem2/event2. Requires root. No ftrace events are collected by default. Note: Not supported on IBM Power targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ftrace-keep-user-config</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Skip initial ftrace setup and collect already configured events. Default resets the ftrace configuration.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-device</span></code></p></td>
<td><p>GPU ID, help, all, none</p></td>
<td><p>none</p></td>
<td><p>Collect GPU Metrics from specified devices. Determine GPU IDs by using <code class="docutils literal notranslate"><span class="pre">--gpu-metrics-device=help</span></code> switch.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-frequency</span></code></p></td>
<td><p>integer</p></td>
<td><p>10000</p></td>
<td><p>Specify GPU Metrics sampling frequency. Minimum supported frequency is 10 (Hz). Maximum supported frequency is 200000(Hz).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-set</span></code></p></td>
<td><p>index</p></td>
<td><p>first</p></td>
<td><p>Specify metric set for GPU Metrics sampling. The argument must be one of indices reported by <code class="docutils literal notranslate"><span class="pre">--gpu-metrics-set=help</span></code> switch. Default is the first metric set that supports selected GPU.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--gpuctxsw</span></code></p></td>
<td><p>true,false</p></td>
<td><p>false</p></td>
<td><p>Trace GPU context switches. Note that this requires driver r435.17 or later and root permission. Not supported on IBM Power targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-device</span></code></p></td>
<td><p><IB switch GUIDs></p></td>
<td><p>none</p></td>
<td><p>A comma-separated list of InfiniBand switch GUIDs. Collect InfiniBand switch congestion events from switches identified by the specified GUIDs. This switch can be used multiple times. System scope. Use the –ib-switch-congestion-nic-device, –ib-switch-congestion-percent, and –ib-switch-congestion-threshold-high switches to further control how congestion events are collected.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-nic-device</span></code></p></td>
<td><p><NIC name></p></td>
<td><p>none</p></td>
<td><p>The name of the NIC (HCA) through which InfiniBand switches will be accessed. By default, the first active NIC will be used. One way to find a NIC’s name is via the <cite>ibnetdiscover –Hca_list | grep “$(hostname)”</cite> command. Example usage: –ib-switch-congestion-nic-device=mlx5_3</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-percent</span></code></p></td>
<td><p>1 <= integer <= 100</p></td>
<td><p>50</p></td>
<td><p>Percent of InfiniBand switch congestion events to be collected. This option enables reducing the network bandwidth consumed by reporting congestion events.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ib-switch-congestion-threshold-high</span></code></p></td>
<td><p>1 <= integer <= 1023</p></td>
<td><p>75</p></td>
<td><p>High threshold percentage for InfiniBand switch egress port buffer size. Before a packet leaves an InfiniBand switch, it is stored at an egress port buffer. The buffer’s size is checked and if it exceeds the given threshold percentage, a congestion event is reported. The percentage can be greater than 100.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--isr</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Trace Interrupt Service Routines (ISRs) and Deferred Procedure Calls (DPCs). Requires administrative privileges. Available only on Windows devices.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nic-metrics</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect metrics from supported NIC/HCA devices. System scope. Not available on Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--os-events</span></code></p></td>
<td><p>‘help’ or the end users selected events in the format ‘x,y’</p></td>
<td><p>none</p></td>
<td><p>Select the OS events to sample. Use the <code class="docutils literal notranslate"><span class="pre">--os-events=help</span></code> switch to see the full list of events. Multiple values can be selected, separated by commas only (no spaces). Use the <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code> switch to enable. Not available in Nsight Systems Embedded Platforms Edition.</p></td>
</tr>
<tr class="row-even"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p>< filename ></p></td>
<td><p>report#</p></td>
<td><p>Set report file name. Any %q{ENV_VAR} pattern in the filename will be substituted with the value of the environment variable. Any %h pattern in the filename will be substituted with the hostname of the system. Any %p pattern in the filename will be substituted with the PID of the target process or the PID of the root process if there is a process tree. Any %% pattern in the filename will be substituted with %. Default is report#.{nsys-rep,sqlite,h5,txt,arrows,json} in the working directory.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--process-scope</span></code></p></td>
<td><p>main, process-tree, system-wide</p></td>
<td><p>main</p></td>
<td><p>Select which process(es) to trace. Available in Nsight Systems Embedded Platforms Edition only. Nsight Systems Workstation Edition will always trace system-wide in this version of the tool.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--retain-etw-files</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Retain ETW files generated by the trace, merge and move the files to the output directory.</p></td>
</tr>
<tr class="row-odd"><td><p>-s</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sample</span></code></p></td>
<td><p>process-tree, system-wide, none</p></td>
<td><p>process-tree</p></td>
<td><p>Select how to collect CPU IP/backtrace samples. If ‘none’ is selected, CPU sampling is disabled. Depending on the platform, some values may require admin or root privileges. If a target application is launched, the default is ‘process-tree’, otherwise, the default is ‘none’. Note: ‘system-wide’ is not available on all platforms. Note: If set to ‘none’, CPU context switch data will still be collected unless the <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code> switch is set to ‘none’.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--samples-per-backtrace</span></code></p></td>
<td><p>integer <= 32</p></td>
<td><p>1</p></td>
<td><p>The number of CPU IP samples collected for every CPU IP/backtrace sample collected. For example, if set to 4, on the fourth CPU IP sample collected, a backtrace will also be collected. Lower values increase the amount of data collected. Higher values can reduce collection overhead and reduce the number of CPU IP samples dropped. If DWARF backtraces are collected, the default is 4, otherwise the default is 1. This option is not available on Nsight Systems Embedded Platforms Edition or on non-Linux targets.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-frequency</span></code></p></td>
<td><p>integers between 100 and 8000</p></td>
<td><p>1000</p></td>
<td><p>Specify the sampling/backtracing frequency. The minimum supported frequency is 100 Hz. The maximum supported frequency is 8000 Hz. This option is supported only on QNX, Linux for Tegra, and Windows targets. Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch in Nsight Systems Embedded Platforms Edition</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of CPU Cycle events counted before a CPU instruction pointer (IP) sample is collected. If configured, backtraces may also be collected. The smaller the sampling period, the higher the sampling rate. Note that smaller sampling periods will increase overhead and significantly increase the size of the result file(s). Requires <code class="docutils literal notranslate"><span class="pre">--sampling-trigger=perf</span></code> switch.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> (not Nsight Systems Embedded Platforms Edition)</p></td>
<td><p>integer</p></td>
<td><p>determined dynamically</p></td>
<td><p>The number of events counted before a CPU instruction pointer (IP) sample is collected. The event used to trigger the collection of a sample is determined dynamically. For example, on Intel based platforms, it will probably be “Reference Cycles” and on AMD platforms, “CPU Cycles”. If configured, backtraces may also be collected. The smaller the sampling period, the higher the sampling rate. Note that smaller sampling periods will increase overhead and significantly increase the size of the result file(s). This option is available only on Linux targets.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling-trigger</span></code></p></td>
<td><p>timer, sched, perf, cuda</p></td>
<td><p>timer,sched</p></td>
<td><p>Specify backtrace collection trigger. Multiple APIs can be selected, separated by commas only (no spaces). Available on Nsight Systems Embedded Platforms Edition targets only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p>session identifier</p></td>
<td><p>none</p></td>
<td><p>Start the application in the indicated session. The option argument must represent a valid session name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session-new</span></code></p></td>
<td><p>[a-Z][0-9,a-Z,spaces]</p></td>
<td><p>[default]</p></td>
<td><p>Start the application in a new session. Name must start with an alphabetical character followed by printable or space characters. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics</span></code></p></td>
<td><p>true,false</p></td>
<td><p>false</p></td>
<td><p>Collect SOC Metrics. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-frequency</span></code></p></td>
<td><p>integer</p></td>
<td><p>10000</p></td>
<td><p>Specify SOC Metrics sampling frequency. Minimum supported frequency is ‘100’ (Hz). Maximum supported frequency is ‘1000000’ (Hz). Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-set</span></code></p></td>
<td><p>see description</p></td>
<td><p>see description</p></td>
<td><p>Specify metric set for SOC Metrics sampling. The option argument must be one of indices or aliases reported by <code class="docutils literal notranslate"><span class="pre">--soc-metrics-set=help</span></code> switch. Default is the first supported set. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--stats</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Generate summary statistics after the collection. WARNING: When set to true, an SQLite database will be created after the collection. If the collection captures a large amount of data, creating the database file may take several minutes to complete.</p></td>
</tr>
<tr class="row-odd"><td><p>-x</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--stop-on-exit</span></code></p></td>
<td><p>true, false</p></td>
<td><p>true</p></td>
<td><p>If true, stop collecting automatically when all tracked processes have exited or when <code class="docutils literal notranslate"><span class="pre">stop</span></code> command is issued - whichever occurs first. If false, stop only on <code class="docutils literal notranslate"><span class="pre">stop</span></code> command. Note: When this is true, <code class="docutils literal notranslate"><span class="pre">stop</span></code> command is optional. Nsight Systems does not officially support runs longer than 5 minutes.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--vsync</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Collect vsync events. If collection of vsync events is enabled, display/display_scanline ftrace events will also be captured. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--xhv-trace</span></code></p></td>
<td><p>< filepath pct.json ></p></td>
<td><p>none</p></td>
<td><p>Collect hypervisor trace. Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--xhv-trace-events</span></code></p></td>
<td><p>all, none, core, sched, irq, trap</p></td>
<td><p>all</p></td>
<td><p>Available in Nsight Systems Embedded Platforms Edition only.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-stats-command-switch-options">
<h4>CLI Stats Command Switch Options<a class="headerlink" href="#cli-stats-command-switch-options" title="Permalink to this heading"></a></h4>
<p>The <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span></code> command generates a series of summary or trace reports.
These reports can be output to the console, or to individual files, or piped to
external processes. Reports can be rendered in a variety of different output
formats, from human readable columns of text, to formats more appropriate for
data exchange, such as CSV.</p>
<p>Reports are generated from an SQLite export of a .nsys-rep file. If a .nsys-rep
file is specified, Nsight Systems will look for an accompanying SQLite file and
use it. If no SQLite file exists, one will be exported and created.</p>
<p>Individual reports are generated by calling out to scripts that read data from
the SQLite file and return their report data in CSV format. Nsight Systems
ingests this data and formats it as requested, then displays the data to the
console, writes it to a file, or pipes it to an external process. Adding new
reports is as simple as writing a script that can read the SQLite file and
generate the required CSV output. See the shipped scripts as an example. Both
reports and formatters may take arguments to tweak their processing. For details
on shipped scripts and formatters, see <strong>Report Scripts</strong> topic.</p>
<p>Reports are processed using a three-tuple that consists of 1) the requested
report (and any arguments), 2) the presentation format (and any arguments), and
3) the output (filename, console, or external process). The first report
specified uses the first format specified, and is presented via the first output
specified. The second report uses the second format for the second output, and
so forth. If more reports are specified than formats or outputs, the format
and/or output list is expanded to match the number of provided reports by
repeating the last specified element of the list (or the default, if nothing was
specified).</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span></code> is a very powerful command and can handle complex argument
structures, please see the topic below on Example Stats Command Sequences.</p>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">stats</span></code> command switch, the following options are
available. Usage:</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">[global-options]</span> <span class="pre">stats</span> <span class="pre">[options]</span> <span class="pre">[input-file]</span></code></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options
relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--format</span></code></p></td>
<td><p>column, table, csv, tsv,
json, hdoc, htable, .</p></td>
<td></td>
<td><p>Specify the output format. The special name “.” indicates the default format for the given output. The default format for console
is column, while files and process outputs default to csv. This option may be used multiple times. Multiple formats may also be
specified using a comma-separated list (<name[:args…][,name[:args…]…]>). See <strong>Report Scripts</strong> for options available with
each format.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-export</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Force a re-export of the SQLite file from the specified .nsys-rep file, even if an SQLite file already exists.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td><p>true, false</p></td>
<td><p>false</p></td>
<td><p>Overwrite any existing report file(s).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help-formats</span></code></p></td>
<td><p><format_name>, ALL, [none]</p></td>
<td><p>none</p></td>
<td><p>With no argument, give a summary of the available output formats. If a format name is given, a more detailed explanation of that
format is displayed. If <code class="docutils literal notranslate"><span class="pre">ALL</span></code> is given, a more detailed explanation of all available formats is displayed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help-reports</span></code></p></td>
<td><p><report_name>, ALL, [none]</p></td>
<td><p>none</p></td>
<td><p>With no argument, list a summary of the available summary and trace reports. If a report name is given, a more detailed explanation
of the report is displayed. If <code class="docutils literal notranslate"><span class="pre">ALL</span></code> is given, a more detailed explanation of all available reports is displayed.</p></td>
</tr>
<tr class="row-even"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p>-, @<command>, <basename>, .</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">-</span></code></p></td>
<td><p>Specify the output mechanism. There are three output mechanisms: print to console, output to file, or output to command. This
option may be used multiple times. Multiple outputs may also be specified using a comma-separated list. If the given output name
is “-”, the output will be displayed on the console. If the output name starts with “@”, the output designates a command to run.
The nsys command will be executed and the analysis output will be piped into the command. Any other output is assumed to be the
base path and name for a file. If a file basename is given, the filename used will be: <basename>_<analysis&args>.<output_format>.
The default base (including path) is the name of the SQLite file (as derived from the input file or <code class="docutils literal notranslate"><span class="pre">--sqlite</span></code> option), minus
the extension. The output “.” can be used to indicate the analysis should be output to a file, and the default basename should be
used. To write one or more analysis outputs to files using the default basename, use the option: <code class="docutils literal notranslate"><span class="pre">--output</span></code>. If the output starts
with “@”, the nsys command output is piped to the given command. The command is run, and the output is piped to the command’s stdin
(standard-input). The command’s stdout and stderr remain attached to the console, so any output will be displayed directly to the
console. Be aware there are some limitations in how the command string is parsed. No shell expansions (including *, ?, [], and ~)
are supported. The command cannot be piped to another command, nor redirected to a file using shell syntax. The command and command
arguments are split on whitespace, and no quotes (within the command syntax) are supported. For commands that require complex
command line syntax, it is suggested that the command be put in a shell script file, and that be designated as the output command.</p></td>
</tr>
<tr class="row-odd"><td><p>-q</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--quiet</span></code></p></td>
<td></td>
<td></td>
<td><p>Do not display verbose messages, only display errors.</p></td>
</tr>
<tr class="row-even"><td><p>-r</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--report</span></code></p></td>
<td><p>See <strong>Report Scripts</strong></p></td>
<td></td>
<td><p>Specify the report(s) to generate, including any arguments. This option may be used multiple times. Multiple reports may also be
specified using a comma-separated list (<name[:args…][,name[:args…]…]>). If no reports are specified, the following will be
used as the default report set: nvtx_sum, osrt_sum, cuda_api_sum, cuda_gpu_kern_sum, cuda_gpu_mem_time_sum, cuda_gpu_mem_size_sum,
openmp_sum, opengl_khr_range_sum, opengl_khr_gpu_range_sum, vulkan_marker_sum, vulkan_gpu_marker_sum, dx11_pix_sum,
dx12_gpu_marker_sum, dx12_pix_sum, wddm_queue_sum, um_sum, um_total_sum, um_cpu_page_faults_sum, openacc_sum. See
<strong>Report Scripts</strong> section for details about existing built-in scripts and how to make your own.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--report-dir</span></code></p></td>
<td><p><path></p></td>
<td></td>
<td><p>Add a directory to the path used to find report scripts. This is usually only needed if you have one or more directories with
personal scripts. This option may be used multiple times. Each use adds a new directory to the end of the path. A search path can
also be defined using the environment variable “NSYS_STATS_REPORT_PATH”. Directories added this way will be added after the
application flags. The last two entries in the path will always be the current working directory, followed by the directory
containing the shipped <code class="docutils literal notranslate"><span class="pre">nsys</span></code> reports.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sqlite</span></code></p></td>
<td><p><file.sqlite></p></td>
<td></td>
<td><p>Specify the SQLite export filename. If this file exists, it will be used. If this file doesn’t exist (or if <code class="docutils literal notranslate"><span class="pre">--force-export</span></code>
was given) this file will be created from the specified .nsys-rep file before processing. This option cannot be used if the
specified input file is also an SQLite file.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--timeunit</span></code></p></td>
<td><p>nsec, nanoseconds, usec,
microseconds, msec,
milliseconds, seconds</p></td>
<td><p>nanoseconds</p></td>
<td><p>Set basic unit of time. The argument of the switch is matched by using the longest prefix matching. Meaning that it is not
necessary to write a whole word as the switch argument. It is similar to passing a “:time=<unit>” argument to every formatter,
although the formatter uses more strict naming conventions. See <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">--help-formats</span> <span class="pre">column</span></code> for more detailed information
on unit conversion.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-status-command-switch-options">
<h4>CLI Status Command Switch Options<a class="headerlink" href="#cli-status-command-switch-options" title="Permalink to this heading"></a></h4>
<p>The <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span></code> command returns the current state of the CLI. After choosing the <code class="docutils literal notranslate"><span class="pre">status</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">status</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--all</span></code></p></td>
<td></td>
<td></td>
<td><p>Prints information for all the available profiling environments.</p></td>
</tr>
<tr class="row-odd"><td><p>-e</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--environment</span></code></p></td>
<td></td>
<td></td>
<td><p>Returns information about the system regarding suitability of the profiling environment.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td><p>-n</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--network</span></code></p></td>
<td></td>
<td></td>
<td><p>Returns information about the system regarding suitability of the network profiling environment.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p>session identifier</p></td>
<td><p>none</p></td>
<td><p>Print the status of the indicated session. The option argument must represent a valid session name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
</tbody>
</table>
</section>
<section id="cli-stop-command-switch-options">
<h4>CLI Stop Command Switch Options<a class="headerlink" href="#cli-stop-command-switch-options" title="Permalink to this heading"></a></h4>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">stop</span></code> command switch, the following options are available. Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">stop</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Possible Parameters</p></th>
<th class="head"><p>Default</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p><tag></p></td>
<td><p>none</p></td>
<td><p>Print the help message. The option can take one optional argument that will be used as a tag. If a tag is provided, only options relevant to the tag will be printed.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--session</span></code></p></td>
<td><p>session identifier</p></td>
<td><p>none</p></td>
<td><p>Stop the indicated session. The option argument must represent a valid session name or ID as reported by <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">sessions</span> <span class="pre">list</span></code>. Any <code class="docutils literal notranslate"><span class="pre">%q{ENV_VAR}</span></code> pattern will be substituted with the value of the environment variable. Any <code class="docutils literal notranslate"><span class="pre">%h</span></code> pattern will be substituted with the hostname of the system. Any <code class="docutils literal notranslate"><span class="pre">%%</span></code> pattern will be substituted with <code class="docutils literal notranslate"><span class="pre">%</span></code>.</p></td>
</tr>
</tbody>
</table>
</section>
</section>
<section id="example-single-command-lines">
<h3>Example Single Command Lines<a class="headerlink" href="#example-single-command-lines" title="Permalink to this heading"></a></h3>
<p><strong>Version Information</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys -v
</pre></div>
</div>
<p>Effect: Prints tool version information to the screen.</p>
<p><strong>Run with elevated privilege</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>sudo nsys profile <app>
</pre></div>
</div>
<p>Effect: Nsight Systems CLI (and target application) will run with elevated privilege. This is necessary for some features, such as FTrace or system-wide CPU sampling. If you don’t want the target application to be elevated, use <code class="docutils literal notranslate"><span class="pre">--run-as</span></code> option.</p>
<p><strong>Default analysis run</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile <application>
[application-arguments]
</pre></div>
</div>
<p>Effect: Launch the application using the given arguments. Start collecting immediately and end collection when the application stops. Trace CUDA, OpenGL, NVTX, and OS runtime libraries APIs. Collect CPU sampling information and thread scheduling information. With Nsight Systems Embedded Platforms Edition this will only analysis the single process. With Nsight Systems Workstation Edition this will trace the process tree. Generate the report#.nsys-rep file in the default location, incrementing the report number if needed to avoid overwriting any existing output files.</p>
<p><strong>Limited trace only run</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --trace=cuda,nvtx -d 20
--sample=none --cpuctxsw=none -o my_test <application>
[application-arguments]
</pre></div>
</div>
<p>Effect: Launch the application using the given arguments. Start collecting immediately and end collection after 20 seconds or when the application ends. Trace CUDA and NVTX APIs. Do not collect CPU sampling information or thread scheduling information. Profile any child processes. Generate the output file as my_test.nsys-rep in the current working directory.</p>
<p><strong>Delayed start run</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile -e TEST_ONLY=0 -y 20
<application> [application-arguments]
</pre></div>
</div>
<p>Effect: Set environment variable TEST_ONLY=0. Launch the application using the given arguments. Start collecting after 20 seconds and end collection at application exit. Trace CUDA, OpenGL, NVTX, and OS runtime libraries APIs. Collect CPU sampling and thread schedule information. Profile any child processes. Generate the report#.nsys-rep file in the default location, incrementing if needed to avoid overwriting any existing output files.</p>
<p><strong>Collect ftrace events</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --ftrace=drm/drm_vblank_event
-d 20
</pre></div>
</div>
<p>Effect: Collect ftrace <code class="docutils literal notranslate"><span class="pre">drm_vblank_event</span></code> events for 20 seconds. Generate the report#.nsys-rep file in the current working directory. Note that ftrace event collection requires running as root. To get a list of ftrace events available from the kernel, run the following:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>sudo cat /sys/kernel/debug/tracing/available_events
</pre></div>
</div>
<p><strong>Run GPU metric sampling on one TU10x</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --gpu-metrics-device=0
--gpu-metrics-set=tu10x-gfxt <application>
</pre></div>
</div>
<p>Effect: Launch application. Collect default options and GPU metrics for the first GPU (a TU10x), using the tu10x-gfxt metric set at the default frequency (10 kHz). Profile any child processes. Generate the report#.nsys-rep file in the default location, incrementing if needed to avoid overwriting any existing output files.</p>
<p><strong>Run GPU metric sampling on all GPUs at a set frequency</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --gpu-metrics-device=all
--gpu-metrics-frequency=20000 <application>
</pre></div>
</div>
<p>Effect: Launch application. Collect default options and GPU metrics for all available GPUs using the first suitable metric set for each and sampling at 20 kHz. Profile any child processes. Generate the report#.nsys-rep file in the default location, incrementing if needed to avoid overwriting any existing output files.</p>
<p><strong>Collect CPU IP/backtrace and CPU context switch</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --sample=system-wide --duration=5
</pre></div>
</div>
<p>Effect: Collects both CPU IP/backtrace samples using the default backtrace mechanism and traces CPU context switch activity for the whole system for 5 seconds. Note that it requires root permission to run. No hardware or OS events are sampled. Post processing of this collection will take longer due to the large number of symbols to be resolved caused by system-wide sampling.</p>
<p><strong>Get list of available CPU core events</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --cpu-core-events=help
</pre></div>
</div>
<p>Effect: Lists the CPU events that can be sampled and the maximum number of CPU events that can be sampled concurrently.</p>
<p><strong>Collect system-wide CPU events and trace application</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --event-sample=system-wide
--cpu-core-events='1,2' --event-sampling-frequency=5 <app> [app args]
</pre></div>
</div>
<p>Effect:Collects CPU IP/backtrace samples using the default backtrace mechanism, traces CPU context switch activity, and samples each CPU’s “CPU Cycles” and “Instructions Retired” event every 200 ms for the whole system. Note that it requires root permission to run. Note that CUDA, NVTX, OpenGL, and OSRT within the app launched by Nsight Systems are traced by default while using this command. Post processing of this collection will take longer due to the large number of symbols to be resolved caused by system-wide sampling.</p>
<p><strong>Collect custom ETW trace using configuration file</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --etw-provider=file.JSON
</pre></div>
</div>
<p>Effect: Configure custom ETW collectors using the contents of file.JSON. Collect data for 20 seconds. Generate the report#.nsys-rep file in the current working directory.</p>
<p>A template JSON configuration file is located at in the Nsight Systems installation directory as \target-windows-x64\etw_providers_template.json. This path will show up automatically if you call</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --help
</pre></div>
</div>
<p>The <strong>level</strong> attribute can only be set to one of the following:</p>
<ul class="simple">
<li><p>TRACE_LEVEL_CRITICAL</p></li>
<li><p>TRACE_LEVEL_ERROR</p></li>
<li><p>TRACE_LEVEL_WARNING</p></li>
<li><p>TRACE_LEVEL_INFORMATION</p></li>
<li><p>TRACE_LEVEL_VERBOSE</p></li>
</ul>
<p>The <strong>flags</strong> attribute can only be set to one or more of the following:</p>
<ul class="simple">
<li><p>EVENT_TRACE_FLAG_ALPC</p></li>
<li><p>EVENT_TRACE_FLAG_CSWITCH</p></li>
<li><p>EVENT_TRACE_FLAG_DBGPRINT</p></li>
<li><p>EVENT_TRACE_FLAG_DISK_FILE_IO</p></li>
<li><p>EVENT_TRACE_FLAG_DISK_IO</p></li>
<li><p>EVENT_TRACE_FLAG_DISK_IO_INIT</p></li>
<li><p>EVENT_TRACE_FLAG_DISPATCHER</p></li>
<li><p>EVENT_TRACE_FLAG_DPC</p></li>
<li><p>EVENT_TRACE_FLAG_DRIVER</p></li>
<li><p>EVENT_TRACE_FLAG_FILE_IO</p></li>
<li><p>EVENT_TRACE_FLAG_FILE_IO_INIT</p></li>
<li><p>EVENT_TRACE_FLAG_IMAGE_LOAD</p></li>
<li><p>EVENT_TRACE_FLAG_INTERRUPT</p></li>
<li><p>EVENT_TRACE_FLAG_JOB</p></li>
<li><p>EVENT_TRACE_FLAG_MEMORY_HARD_FAULTS</p></li>
<li><p>EVENT_TRACE_FLAG_MEMORY_PAGE_FAULTS</p></li>
<li><p>EVENT_TRACE_FLAG_NETWORK_TCPIP</p></li>
<li><p>EVENT_TRACE_FLAG_NO_SYSCONFIG</p></li>
<li><p>EVENT_TRACE_FLAG_PROCESS</p></li>
<li><p>EVENT_TRACE_FLAG_PROCESS_COUNTERS</p></li>
<li><p>EVENT_TRACE_FLAG_PROFILE</p></li>
<li><p>EVENT_TRACE_FLAG_REGISTRY</p></li>
<li><p>EVENT_TRACE_FLAG_SPLIT_IO</p></li>
<li><p>EVENT_TRACE_FLAG_SYSTEMCALL</p></li>
<li><p>EVENT_TRACE_FLAG_THREAD</p></li>
<li><p>EVENT_TRACE_FLAG_VAMAP</p></li>
<li><p>EVENT_TRACE_FLAG_VIRTUAL_ALLOC</p></li>
</ul>
<p><strong>Typical case: profile a Python script that uses CUDA</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --trace=cuda,cudnn,cublas,osrt,nvtx
--delay=60 python my_dnn_script.py
</pre></div>
</div>
<p>Effect: Launch a Python script and start profiling it 60 seconds after the launch, tracing CUDA, cuDNN, cuBLAS, OS runtime APIs, and NVTX as well as collecting thread schedule information.</p>
<p><strong>Typical case: profile an app that uses Vulkan</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --trace=vulkan,osrt,nvtx
--delay=60 ./myapp
</pre></div>
</div>
<p>Effect: Launch an app and start profiling it 60 seconds after the launch, tracing Vulkan, OS runtime APIs, and NVTX as well as collecting CPU sampling and thread schedule information.</p>
</section>
<section id="example-interactive-cli-command-sequences">
<h3>Example Interactive CLI Command Sequences<a class="headerlink" href="#example-interactive-cli-command-sequences" title="Permalink to this heading"></a></h3>
<p><strong>Collect from beginning of application, end manually</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys start --stop-on-exit=false
nsys launch --trace=cuda,nvtx --sample=none <application> [application-arguments]
nsys stop
</pre></div>
</div>
<p>Effect: Create interactive CLI process and set it up to begin collecting as soon as an application is launched. Launch the application, set up to allow tracing of CUDA and NVTX as well as collection of thread schedule information. Stop only when explicitly requested. Generate the report#.nsys-rep in the default location.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If you start a collection and fail to stop the collection (or if you are allowing it to stop on exit, and the application runs for too long) your system’s storage space may be filled with collected data causing significant issues for the system. Nsight Systems will collect a different amount of data/sec depending on options, but in general Nsight Systems does not support runs of more than 5 minutes duration.</p>
</div>
<p><strong>Run application, begin collection manually, run until process ends</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch -w true <application> [application-arguments]
nsys start
</pre></div>
</div>
<p>Effect: Create interactive CLI and launch an application set up for default analysis. Send application output to the terminal. No data is collected until you manually start collection at area of interest. Profile until the application ends. Generate the report#.nsys-rep in the default location.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If you launch an application and that application and any descendants exit before start is called Nsight Systems will create a fully formed .nsys-rep file containing no data.</p>
</div>
<p><strong>Run application, start/stop collection using cudaProfilerStart/Stop</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys start -c cudaProfilerApi
nsys launch -w true <application> [application-arguments]
</pre></div>
</div>
<p>Effect: Create interactive CLI process and set it up to begin collecting as soon as a cudaProfileStart() is detected. Launch application for default analysis, sending application output to the terminal. Stop collection at next call to cudaProfilerStop, when the user calls <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stop</span></code>, or when the root process terminates. Generate the report#.nsys-rep in the default location.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If you call <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">launch</span></code> before <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">start</span> <span class="pre">-c</span> <span class="pre">cudaProfilerApi</span></code> and the code contains a large number of short duration cudaProfilerStart/Stop pairs, Nsight Systems may be unable to process them correctly, causing a fault. This will be corrected in a future version.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The Nsight Systems CLI does not support multiple calls to the cudaProfilerStart/Stop API at this time.</p>
</div>
<p><strong>Run application, start/stop collection using NVTX</strong></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys start -c nvtx
nsys launch -w true -p MESSAGE@DOMAIN <application> [application-arguments]
</pre></div>
</div>
<p>Effect: Create interactive CLI process and set it up to begin collecting as soon as an NVTX range with given message in given domain (capture range) is opened. Launch application for default analysis, sending application output to the terminal. Stop collection when all capture ranges are closed, when the user calls <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stop</span></code>, or when the root process terminates. Generate the report#.nsys-rep in the default location.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The Nsight Systems CLI only triggers the profiling session for the first capture range.</p>
</div>
<p>NVTX capture range can be specified:</p>
<ul>
<li><p>Message@Domain: All ranges with given message in given domain are capture ranges. For example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch -w true -p profiler@service ./app
</pre></div>
</div>
<p>This would make the profiling start when the first range with message “profiler” is opened in domain “service”.</p>
</li>
<li><p>Message@*: All ranges with given message in all domains are capture ranges. For example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch -w true -p profiler@* ./app
</pre></div>
</div>
<p>This would make the profiling start when the first range with message “profiler” is opened in any domain.</p>
</li>
<li><p>Message: All ranges with given message in default domain are capture ranges. For example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch -w true -p profiler ./app
</pre></div>
</div>
<p>This would make the profiling start when the first range with message “profiler” is opened in the default domain.</p>
</li>
<li><p>By default only messages, provided by NVTX registered strings are considered to avoid additional overhead. To enable non-registered strings check please launch your application with <code class="docutils literal notranslate"><span class="pre">NSYS_NVTX_PROFILER_REGISTER_ONLY=0</span></code> environment:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch -w true -p profiler@service -e NSYS_NVTX_PROFILER_REGISTER_ONLY=0 ./app
</pre></div>
</div>
</li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The separator ‘@’ can be escaped with backslash ‘\’. If multiple separators without escape character are specified, only the last one is applied, all others are discarded.</p>
</div>
<p><strong>Run application, start/stop collection multiple times</strong></p>
<p>The interactive CLI supports multiple sequential collections per launch.</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys launch <application> [application-arguments]
nsys start
nsys stop
nsys start
nsys stop
nsys shutdown --kill sigkill
</pre></div>
</div>
<p>Effect: Create interactive CLI and launch an application set up for default analysis. Send application output to the terminal. No data is collected until the start command is executed. Collect data from start until stop requested, generate report#.qstrm in the current working directory. Collect data from second start until the second stop request, generate report#.nsys-rep (incremented by one) in the current working directory. Shutdown the interactive CLI and send sigkill to the target application’s process group.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Calling <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">cancel</span></code> after <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">start</span></code> will cancel the collection without generating a report.</p>
</div>
</section>
<section id="example-stats-command-sequences">
<h3>Example Stats Command Sequences<a class="headerlink" href="#example-stats-command-sequences" title="Permalink to this heading"></a></h3>
<p><strong>Display default statistics</strong></p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">report1.nsys-rep</span></code></p>
<p>Effect: Export an SQLite file named report1.sqlite from report1.nsys-rep (assuming it does not already exist). Print the default reports in column format to the console.</p>
<p>Note: The following two command sequences should present very similar information:</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre">--stats=true</span> <span class="pre"><application></span></code></p>
<p>or</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre"><application></span></code></p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">report1.nsys-rep</span></code></p>
<p><strong>Display specific data from a report</strong></p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">--report</span> <span class="pre">cuda_gpu_trace</span> <span class="pre">report1.nsys-rep</span></code></p>
<p>Effect: Export an SQLite file named report1.sqlite from report1.nsys-rep (assuming it does not already exist). Print the report generated by the <code class="docutils literal notranslate"><span class="pre">cuda_gpu_trace</span></code> script to the console in column format.</p>
<p><strong>Generate multiple reports, in multiple formats, output multiple places</strong></p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">--report</span> <span class="pre">cuda_gpu_trace</span> <span class="pre">--report</span> <span class="pre">cuda_gpu_kern_sum</span> <span class="pre">--report</span> <span class="pre">cuda_api_sum</span> <span class="pre">--format</span> <span class="pre">csv,column</span> <span class="pre">--output</span> <span class="pre">.,-</span> <span class="pre">report1.nsys-rep</span></code></p>
<p>Effect: Export an SQLite file named report1.sqlite from report1.nsys-rep (assuming it does not already exist). Generate three reports. The first, the <code class="docutils literal notranslate"><span class="pre">cuda_gpu_trace</span></code> report, will be output to the file report1_cuda_gpu_trace.csv in CSV format. The other two reports, <code class="docutils literal notranslate"><span class="pre">cuda_gpu_kern_sum</span></code> and <code class="docutils literal notranslate"><span class="pre">cuda_api_sum</span></code>, will be output to the console as columns of data. Although three reports were given, only two formats and outputs are given. To reconcile this, both the list of formats and outputs is expanded to match the list of reports by repeating the last element.</p>
<p><strong>Submit report data to a command</strong></p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span> <span class="pre">--report</span> <span class="pre">cuda_api_sum</span> <span class="pre">--format</span> <span class="pre">table</span> <span class="pre">\</span> <span class="pre">--output</span> <span class="pre">@“grep</span> <span class="pre">-E</span> <span class="pre">(-|Name|cudaFree”</span> <span class="pre">test.sqlite</span></code></p>
<p>Effect: Open test.sqlite and run the <code class="docutils literal notranslate"><span class="pre">cuda_api_sum</span></code> script on that file. Generate table data and feed that into the command <code class="docutils literal notranslate"><span class="pre">grep</span> <span class="pre">-E</span> <span class="pre">(-|Name|cudaFree)</span></code>. The grep command will filter out everything but the header, formatting, and the <code class="docutils literal notranslate"><span class="pre">cudaFree</span></code> data, and display the results to the console.</p>
<p>Note: When the output name starts with @, it is defined as a command. The command is run, and the output of the report is piped to the command’s stdin (standard-input). The command’s stdout and stderr remain attached to the console, so any output will be displayed directly to the console.</p>
<p>Be aware there are some limitations in how the command string is parsed. No shell expansions (including *, ?, [], and ~) are supported. The command cannot be piped to another command, nor redirected to a file using shell syntax. The command and command arguments are split on whitespace, and no quotes (within the command syntax) are supported. For commands that require complex command line syntax, it is suggested that the command be put into a shell script file, and the script designated as the output command</p>
</section>
<section id="example-output-from-stats-option">
<h3>Example Output from <code class="docutils literal notranslate"><span class="pre">--stats</span></code> Option<a class="headerlink" href="#example-output-from-stats-option" title="Permalink to this heading"></a></h3>
<p>The <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">stats</span></code> command can be used post analysis to generate specific or personalized reports. For a default fixed set of summary statistics to be automatically generated, you can use the <code class="docutils literal notranslate"><span class="pre">--stats</span></code> option with the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> or <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">start</span></code> command to generate a fixed set of useful summary statistics.</p>
<p>If your run traces CUDA, these include CUDA API, Kernel, and Memory Operation statistics:</p>
<img alt="CUDA Statistics" class="image" src="../_images/3407-profile-cuda-summary.png" />
<p>If your run traces OS runtime events or NVTX push-pop ranges:</p>
<img alt="OS runtime and NVTX Statistics" class="image" src="../_images/3407-profile-osrt-nvtx-summary.png" />
<p>If your run traces graphics debug markers these include DX11 debug markers, DX12 debug markers, Vulkan debug markers or KHR debug markers:</p>
<img alt="Graphics Vulkan debug markers Statistics" class="image" src="../_images/ue4_infiltrator_vulkan_markers-summary.png" />
<p>Recipes for these statistics as well as documentation on how to create your own metrics will be available in a future version of the tool.</p>
</section>
<section id="importing-and-viewing-command-line-results-files">
<h3>Importing and Viewing Command Line Results Files<a class="headerlink" href="#importing-and-viewing-command-line-results-files" title="Permalink to this heading"></a></h3>
<p>The CLI generates a .qdstrm file. The .qdstrm file is an intermediate result file, not intended for multiple imports. It needs to be processed, either by importing it into the GUI or by using the standalone QdstrmImporter to generate an optimized .nsys-rep file. Use this .nsys-rep file when re-opening the result on the same machine, opening the result on a different machine, or sharing results with teammates.</p>
<p>This version of Nsight Systems will attempt to automatically convert the .qdstrm file to a .nsys-rep file with the same name after the run finishes if the required libraries are available. The ability to turn off auto-conversion will be added in a later version.</p>
<p><strong>Import Into the GUI</strong></p>
<p>The CLI and host GUI versions must match to import a .qdstrm file successfully. The host GUI is backward compatible only with .nsys-rep files.</p>
<p>Copy the .qdstrm file you are interested in viewing to a system where the Nsight Systems host GUI is installed. Launch the Nsight Systems GUI. Select <strong>File->Import…</strong> and choose the .qdstrm file you wish to open.</p>
<blockquote>
<div><img alt="Import qdstrm" class="image" src="../_images/import_qdstrm.png" />
</div></blockquote>
<p>The import of really large, multi-gigabyte, .qdstrm files may take up all of the memory on the host computer and lock up the system. This will be fixed in a later version.</p>
<p><strong>Importing Windows ETL files</strong></p>
<p>For Windows targets, ETL files captured with Xperf or the <code class="docutils literal notranslate"><span class="pre">log.cmd</span></code> command supplied with GPUView in the Windows Performance Toolkit can be imported to create reports as if they were captured with Nsight Systems’s “WDDM trace” and “Custom ETW trace” features. Simply choose the .etl file from the Import dialog to convert it to a .nsys-rep file.</p>
<p><strong>Create .nsys-rep Using QdstrmImporter</strong></p>
<p>The CLI and QdstrmImporter versions must match to convert a .qdstrm file into a .nsys-rep file. This .nsys-rep file can then be opened in the same version or more recent versions of the GUI.</p>
<p>To run QdstrmImporter on the host system, find the QdstrmImporter binary in the Host-x86_64 directory in your installation. QdstrmImporter is available for all host platforms. See options below.</p>
<p>To run QdstrmImporter on the target system, copy the Linux Host-x86_64 directory to the target Linux system or install Nsight Systems for Linux host directly on the target. The Windows or macOS host QdstrmImporter will not work on a Linux Target. See options below.</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Parameter</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>-h</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td></td>
<td><p>Help message providing information about available options and their parameters.</p></td>
</tr>
<tr class="row-odd"><td><p>-v</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--version</span></code></p></td>
<td></td>
<td><p>Output QdstrmImporter version information</p></td>
</tr>
<tr class="row-even"><td><p>-i</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--input-file</span></code></p></td>
<td><p>filename or path</p></td>
<td><p>Import .qdstrm file from this location.</p></td>
</tr>
<tr class="row-odd"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output-file</span></code></p></td>
<td><p>filename or path</p></td>
<td><p>Provide a different file name or path for the resulting .nsys-rep file. Default is the same name and path as the .qdstrm file</p></td>
</tr>
</tbody>
</table>
</section>
<section id="using-the-cli-to-analyze-mpi-codes">
<h3>Using the CLI to Analyze MPI Codes<a class="headerlink" href="#using-the-cli-to-analyze-mpi-codes" title="Permalink to this heading"></a></h3>
<section id="tracing-mpi-api-calls">
<h4>Tracing MPI API calls<a class="headerlink" href="#tracing-mpi-api-calls" title="Permalink to this heading"></a></h4>
<p>The Nsight Systems CLI has built-in API trace support for Open MPI and MPICH based MPI implementations via <code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code>. It traces a subset of the MPI API, including blocking and non-blocking point-to-point and collective communication as well as MPI one-sided communication, file I/O and pack operations (see <a class="reference external" href="index.html#mpi-trace__mpi_traced_apis">MPI functions traced</a>).</p>
<p>If you require more control over the list of traced APIs or if you are using a different MPI implementation, you can use the <a class="reference external" href="https://github.com/NVIDIA/cuda-profiler/tree/master/nvtx_pmpi_wrappers">NVTX wrappers for MPI</a> on GitHub. Choose an NVTX domain name other than “MPI”, since it is filtered out by Nsight Systems when MPI tracing is not enabled. Use the NVTX-instrumented MPI wrapper library as follows:</p>
<p><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre">-e</span> <span class="pre">LD_PRELOAD=${PATH_TO_YOUR_NVTX_MPI_LIB}</span> <span class="pre">--trace=nvtx</span></code></p>
</section>
<section id="using-the-cli-to-profile-applications-launched-with-mpirun">
<h4>Using the CLI to Profile Applications Launched with mpirun<a class="headerlink" href="#using-the-cli-to-profile-applications-launched-with-mpirun" title="Permalink to this heading"></a></h4>
<p>The Nsight Systems CLI supports concurrent use of the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> command. Each instance will create a separate report file. You cannot use multiple instances of the interactive CLI concurrently, or use the interactive CLI concurrently with <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> in this version.</p>
<p>Nsight Systems can be used to profile applications launched with <code class="docutils literal notranslate"><span class="pre">mpirun</span></code> or <code class="docutils literal notranslate"><span class="pre">mpiexec</span></code>. Since concurrent use of the CLI is supported only when using the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> command, Nsight Systems cannot profile each node from the GUI or from the interactive CLI.</p>
<p><strong>Profile all MPI ranks on a single node:</strong><code class="docutils literal notranslate"><span class="pre">nsys</span></code> can be prefixed before <code class="docutils literal notranslate"><span class="pre">mpirun/mpiexec</span></code>. Only a single report file will be created.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">nsys</span><span class="w"> </span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">mpirun</span><span class="w"> </span><span class="p">[</span><span class="n">mpirun</span><span class="w"> </span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Profile multi-node runs:</strong><code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> has to be prefixed before the program to be profiled. One report file will be created for each MPI rank. This works also for single-node runs.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">mpirun</span><span class="w"> </span><span class="p">[</span><span class="n">mpirun</span><span class="w"> </span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="p">[</span><span class="n">nsys</span><span class="w"> </span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>You can use <code class="docutils literal notranslate"><span class="pre">%q{OMPI_COMM_WORLD_RANK}</span></code> (Open MPI), <code class="docutils literal notranslate"><span class="pre">%q{PMI_RANK}</span></code> (MPICH) or <code class="docutils literal notranslate"><span class="pre">%q{SLURM_PROCID}</span></code> (Slurm) with the <code class="docutils literal notranslate"><span class="pre">-o</span></code> option to appropriately name the report files.</p>
<p><strong>Profile a single MPI process or a subset of MPI processes:</strong> Use a wrapper script similar to the following script (called “profile_rank0.sh”).</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp">#!/bin/bash</span>
<span class="cp"># Use $PMI_RANK for MPICH and $SLURM_PROCID with srun.</span>
<span class="k">if</span><span class="w"> </span><span class="p">[</span><span class="w"> </span><span class="n">$OMPI_COMM_WORLD_RANK</span><span class="w"> </span><span class="o">-</span><span class="n">eq</span><span class="w"> </span><span class="mi">0</span><span class="w"> </span><span class="p">];</span><span class="w"> </span><span class="n">then</span><span class="w"></span>
<span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">-</span><span class="n">e</span><span class="w"> </span><span class="n">NSYS_MPI_STORE_TEAMS_PER_RANK</span><span class="o">=</span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="n">t</span><span class="w"> </span><span class="n">mpi</span><span class="w"> </span><span class="s">"$@"</span><span class="w"></span>
<span class="k">else</span><span class="w"></span>
<span class="w"> </span><span class="s">"$@"</span><span class="w"></span>
<span class="n">fi</span><span class="w"></span>
</pre></div>
</div>
<p>The script runs nsys on rank 0 only. Add appropriate profiling options to the script and execute it with <code class="docutils literal notranslate"><span class="pre">mpirun</span> <span class="pre">[mpirun</span> <span class="pre">options]</span> <span class="pre">./profile_rank0.sh</span> <span class="pre">./myapp</span> <span class="pre">[app</span> <span class="pre">options]</span></code>.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If only a subset of MPI ranks is profiled, set the environment variable <code class="docutils literal notranslate"><span class="pre">NSYS_MPI_STORE_TEAMS_PER_RANK=1</span></code> to store all members of custom MPI communicators per MPI rank. Otherwise, the execution might hang or fail with an MPI error.</p>
</div>
<p><strong>Avoid redundant GPU and NIC metrics collection:</strong> If multiple instances of <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span></code> are executed concurrently on the same node and GPU and/or NIC metrics collection is enabled, each process will collect metrics for all available NICs and tries to collect GPU metrics for the specified devices. This can be avoided with a simple bash script similar to the following:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp">#!/bin/bash</span>
<span class="cp"># Use $SLURM_LOCALID with srun.</span>
<span class="k">if</span><span class="w"> </span><span class="p">[</span><span class="w"> </span><span class="n">$OMPI_COMM_WORLD_LOCAL_RANK</span><span class="w"> </span><span class="o">-</span><span class="n">eq</span><span class="w"> </span><span class="mi">0</span><span class="w"> </span><span class="p">];</span><span class="w"> </span><span class="n">then</span><span class="w"></span>
<span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">nic</span><span class="o">-</span><span class="n">metrics</span><span class="o">=</span><span class="nb">true</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="o">=</span><span class="n">all</span><span class="w"> </span><span class="s">"$@"</span><span class="w"></span>
<span class="k">else</span><span class="w"></span>
<span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="s">"$@"</span><span class="w"></span>
<span class="n">fi</span><span class="w"></span>
</pre></div>
</div>
<p>This above script will collect NIC and GPU metrics only for one rank, the node-local rank 0. Alternatively, if one rank per GPU is used, the GPU metrics devices can be specified based on the node-local rank in a wrapper script as follows:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp">#!/bin/bash</span>
<span class="cp"># Use $SLURM_LOCALID with srun.</span>
<span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">-</span><span class="n">e</span><span class="w"> </span><span class="n">CUDA_VISIBLE_DEVICES</span><span class="o">=</span><span class="n">$</span><span class="p">{</span><span class="n">OMPI_COMM_WORLD_LOCAL_RANK</span><span class="p">}</span><span class="w"> </span>\
<span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="o">=</span><span class="n">$</span><span class="p">{</span><span class="n">OMPI_COMM_WORLD_LOCAL_RANK</span><span class="p">}</span><span class="w"> </span><span class="s">"$@"</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
</section>
<section id="profiling-from-the-gui">
<h2>Profiling from the GUI<a class="headerlink" href="#profiling-from-the-gui" title="Permalink to this heading"></a></h2>
<section id="profiling-linux-targets-from-the-gui">
<h3>Profiling Linux Targets from the GUI<a class="headerlink" href="#profiling-linux-targets-from-the-gui" title="Permalink to this heading"></a></h3>
<section id="connecting-to-the-target-device">
<h4>Connecting to the Target Device<a class="headerlink" href="#connecting-to-the-target-device" title="Permalink to this heading"></a></h4>
<p>Nsight Systems provides a simple interface to profile on localhost or manage multiple connections to Linux or Windows based devices via SSH. The network connections manager can be launched through the device selection dropdown:</p>
<p>On x86_64:</p>
<blockquote>
<div><img alt="Empty device list" class="image" src="../_images/device-list-empty-x86.png" />
</div></blockquote>
<p>On Tegra:</p>
<blockquote>
<div><img alt="Empty device list" class="image" src="../_images/device-list-empty.png" />
</div></blockquote>
<p>The dialog has simple controls that allow adding, removing, and modifying connections:</p>
<blockquote>
<div><img alt="Network connection" class="image" src="../_images/network-connection.png" />
</div></blockquote>
<p><strong>Security notice</strong>: SSH is only used to establish the initial connection to a target device, perform checks, and upload necessary files. The actual profiling commands and data are transferred through a raw, unencrypted socket. Nsight Systems should not be used in a network setup where attacker-in-the-middle attack is possible, or where untrusted parties may have network access to the target device.</p>
<p>While connecting to the target device, you will be prompted to input the user’s password. Please note that if you choose to remember the password, it will be stored in plain text in the configuration file on the host. Stored passwords are bound to the public key fingerprint of the remote device.</p>
<p>The <strong>No authentication</strong> option is useful for devices configured for passwordless login using <code class="docutils literal notranslate"><span class="pre">root</span></code> username. To enable such a configuration, edit the file <code class="docutils literal notranslate"><span class="pre">/etc/ssh/sshd_config</span></code> on the target and specify the following option:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">PermitRootLogin</span><span class="w"> </span><span class="n">yes</span><span class="w"></span>
</pre></div>
</div>
<p>Then set empty password using <code class="docutils literal notranslate"><span class="pre">passwd</span></code> and restart the SSH service with <code class="docutils literal notranslate"><span class="pre">service</span> <span class="pre">ssh</span> <span class="pre">restart</span></code>.</p>
<p><strong>Open ports</strong>: The Nsight Systems daemon requires port 22 and port 45555 to be open for listening. You can confirm that these ports are open with the following command:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">firewall</span><span class="o">-</span><span class="n">cmd</span><span class="w"> </span><span class="o">--</span><span class="n">list</span><span class="o">-</span><span class="n">ports</span><span class="w"> </span><span class="o">--</span><span class="n">permanent</span><span class="w"></span>
<span class="n">sudo</span><span class="w"> </span><span class="n">firewall</span><span class="o">-</span><span class="n">cmd</span><span class="w"> </span><span class="o">--</span><span class="n">reload</span><span class="w"></span>
</pre></div>
</div>
<p>To open a port use the following command, skip <code class="docutils literal notranslate"><span class="pre">--permanent</span></code> option to open only for this session:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">firewall</span><span class="o">-</span><span class="n">cmd</span><span class="w"> </span><span class="o">--</span><span class="n">permanent</span><span class="w"> </span><span class="o">--</span><span class="n">add</span><span class="o">-</span><span class="n">port</span><span class="w"> </span><span class="mi">45555</span><span class="o">/</span><span class="n">tcp</span><span class="w"></span>
<span class="n">sudo</span><span class="w"> </span><span class="n">firewall</span><span class="o">-</span><span class="n">cmd</span><span class="w"> </span><span class="o">--</span><span class="n">reload</span><span class="w"></span>
</pre></div>
</div>
<p>Likewise, if you are running on a cloud system, you must open port 22 and port 45555 for ingress.</p>
<p><strong>Kernel Version Number</strong> - To check for the version number of the kernel support of Nsight Systems on a target device, run the following command on the remote device:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">cat</span><span class="w"> </span><span class="o">/</span><span class="n">proc</span><span class="o">/</span><span class="n">quadd</span><span class="o">/</span><span class="n">version</span><span class="w"></span>
</pre></div>
</div>
<p>Minimal supported version is 1.82.</p>
<p>Additionally, presence of Netcat command (<code class="docutils literal notranslate"><span class="pre">nc</span></code>) is required on the target device. For example, on Ubuntu this package can be installed using the following command:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">apt</span><span class="o">-</span><span class="n">get</span><span class="w"> </span><span class="n">install</span><span class="w"> </span><span class="n">netcat</span><span class="o">-</span><span class="n">openbsd</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="system-wide-profiling-options">
<h4>System-Wide Profiling Options<a class="headerlink" href="#system-wide-profiling-options" title="Permalink to this heading"></a></h4>
</section>
<section id="target-sampling-options">
<h4>Target Sampling Options<a class="headerlink" href="#target-sampling-options" title="Permalink to this heading"></a></h4>
<p>Target sampling behavior is somewhat different for Nsight Systems Workstation Edition and Nsight Systems Embedded Platforms Edition.</p>
</section>
<section id="hotkey-trace-start-stop">
<h4>Hotkey Trace Start/Stop<a class="headerlink" href="#hotkey-trace-start-stop" title="Permalink to this heading"></a></h4>
<p>Nsight Systems Workstation Edition can use hotkeys to control profiling. Press the hotkey to start and/or stop a trace session from within the target application’s graphic window. This is useful when tracing games and graphic applications that use fullscreen display. In these scenarios switching to Nsight Systems’ UI would unnecessarily introduce the window manager’s footprint into the trace. To enable the use of Hotkey check the Hotkey checkbox in the project settings page:</p>
<blockquote>
<div><img alt="Hotkey checkbox" class="image" src="../_images/hotkey-checkbox.png" />
</div></blockquote>
<p>The default hotkey is F12.</p>
</section>
<section id="launching-processes">
<h4>Launching Processes<a class="headerlink" href="#launching-processes" title="Permalink to this heading"></a></h4>
<p>Nsight Systems can launch new processes for profiling on target devices. Profiler ensures that all environment variables are set correctly to successfully collect trace information</p>
<blockquote>
<div><img alt="Process: Launch" class="image" src="../_images/process-launch.png" />
</div></blockquote>
<p>The <strong>Edit arguments…</strong> link will open an editor window, where every command line argument is edited on a separate line. This is convenient when arguments contain spaces or quotes.</p>
</section>
</section>
<section id="profiling-windows-targets-from-the-gui">
<h3>Profiling Windows Targets from the GUI<a class="headerlink" href="#profiling-windows-targets-from-the-gui" title="Permalink to this heading"></a></h3>
<p>Profiling on Windows devices is similar to the profiling on Linux devices. Please refer to the <a class="reference external" href="index.html#profiling-linux">Profiling Linux Targets from the GUI</a> section for the detailed documentation and connection information. The major differences on the platforms are listed below:</p>
<section id="remoting-to-a-windows-based-machine">
<h4>Remoting to a Windows Based Machine<a class="headerlink" href="#remoting-to-a-windows-based-machine" title="Permalink to this heading"></a></h4>
<p>To perform remote profiling to a target Windows based machines, <a class="reference external" href="https://docs.microsoft.com/en-us/windows-server/administration/openssh/openssh_install_firstuse">install and configure an OpenSSH Server on the target machine.</a></p>
</section>
<section id="id1">
<h4>Hotkey Trace Start/Stop<a class="headerlink" href="#id1" title="Permalink to this heading"></a></h4>
<p>Nsight Systems Workstation Edition can use hotkeys to control profiling. Press the hotkey to start and/or stop a trace session from within the target application’s graphic window. This is useful when tracing games and graphic applications that use fullscreen display. In these scenarios switching to Nsight Systems’ UI would unnecessarily introduce the window manager’s footprint into the trace. To enable the use of Hotkey check the Hotkey checkbox in the project settings page:</p>
<blockquote>
<div><img alt="Hotkey checkbox" class="image" src="../_images/hotkey-checkbox.png" />
</div></blockquote>
<p>The default hotkey is F12.</p>
<p><strong>Changing the Default Hotkey Binding</strong> - A different hotkey binding can be configured by setting the HotKeyIntValue configuration field in the config.ini file.</p>
<p>Set the decimal numeric identifier of the hotkey you would like to use for triggering start/stop from the target app graphics window. The default value is 123 which corresponds to 0x7B, or the F12 key.</p>
<p>Virtual key identifiers are detailed in <a class="reference external" href="https://docs.microsoft.com/en-us/windows/win32/inputdev/virtual-key-codes">MSDN’s Virtual-Key Codes</a>.</p>
<p>Note that you must convert the hexadecimal values detailed in this page to their decimal counterpart before using them in the file. For example, to use the F1 key as a start/stop trace hotkey, use the following settings in the config.ini file:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">HotKeyIntValue</span><span class="o">=</span><span class="mi">112</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="target-sampling-options-on-windows">
<h4>Target Sampling Options on Windows<a class="headerlink" href="#target-sampling-options-on-windows" title="Permalink to this heading"></a></h4>
<blockquote>
<div><img alt="Target sampling options" class="image" src="../_images/target-sampling-options-windows.png" />
</div></blockquote>
<p>Nsight Systems can sample one process tree. Sampling here means interrupting each processor periodically. The sampling rate is defined in the project settings and is either 100Hz, 1KHz (default value), 2Khz, 4KHz, or 8KHz.</p>
<blockquote>
<div><img alt="Thread activity option" class="image" src="../_images/collect_thread_activity.png" />
</div></blockquote>
<p>On Windows, Nsight Systems can collect thread activity of one process tree. Collecting thread activity means that each thread context switch event is logged and (optionally) a backtrace is collected at the point that the thread is scheduled back for execution. Thread states are displayed on the timeline.</p>
<p>If it was collected, the thread backtrace is displayed when hovering over a region where the thread execution is blocked.</p>
</section>
<section id="symbol-locations">
<h4>Symbol Locations<a class="headerlink" href="#symbol-locations" title="Permalink to this heading"></a></h4>
<p>Symbol resolution happens on host, and therefore does not affect performance of profiling on the target.</p>
<p>Press the <strong>Symbol locations…</strong> button to open the <strong>Configure debug symbols location</strong> dialog.</p>
<blockquote>
<div><img alt="Configure debug symbols location" class="image" src="../_images/symbols_location_dialog.png" />
</div></blockquote>
<p>Use this dialog to specify:</p>
<ul class="simple">
<li><p>Paths of PDB files</p></li>
<li><p>Symbols servers</p></li>
<li><p>The location of the local symbol cache</p></li>
</ul>
<p>To use a symbol server:</p>
<ol class="arabic">
<li><p>Install <strong>Debugging Tools for Windows</strong>, a part of the <a class="reference external" href="https://developer.microsoft.com/en-us/windows/downloads/windows-10-sdk">Windows 10 SDK</a>.</p></li>
<li><p>Add the symbol server URL using the <strong>Add Server</strong> button.</p>
<p>Information about Microsoft’s public symbol server, which enables getting Windows operating system related debug symbols can be found <a class="reference external" href="https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/microsoft-public-symbols">here</a>.</p>
</li>
</ol>
</section>
</section>
<section id="profiling-qnx-targets-from-the-gui">
<h3>Profiling QNX Targets from the GUI<a class="headerlink" href="#profiling-qnx-targets-from-the-gui" title="Permalink to this heading"></a></h3>
<p>Profiling on QNX devices is similar to the profiling on Linux devices. Please refer to the <a class="reference external" href="index.html#profiling-linux">Profiling Linux Targets from the GUI</a> section for the detailed documentation. The major differences on the platforms are listed below:</p>
<ul>
<li><p>Backtrace sampling is not supported. Instead backtraces are collected for long OS runtime libraries calls. Please refer to the <a class="reference external" href="topics/os-runtime-trace">OS Runtime Libraries Trace</a> section for the detailed documentation.</p></li>
<li><p>CUDA support is limited to CUDA 9.0+</p></li>
<li><p>Filesystem on QNX device might be mounted read-only. In that case Nsight Systems is not able to install target-side binaries, required to run the profiling session. Please make sure that target filesystem is writable before connecting to QNX target. For example, make sure the following command works:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">echo</span><span class="w"> </span><span class="n">XX</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="o">/</span><span class="n">xx</span><span class="w"> </span><span class="o">&&</span><span class="w"> </span><span class="n">ls</span><span class="w"> </span><span class="o">-</span><span class="n">l</span><span class="w"> </span><span class="o">/</span><span class="n">xx</span><span class="w"></span>
</pre></div>
</div>
</li>
</ul>
</section>
</section>
<section id="container-support-on-linux-servers">
<h2>Container Support on Linux Servers<a class="headerlink" href="#container-support-on-linux-servers" title="Permalink to this heading"></a></h2>
<p><strong>Collecting data within a Docker container</strong></p>
<p>While examples in this section use Docker container semantics, other containers work much the same.</p>
<p>The following information assumes the reader is knowledgeable regarding Docker containers. For further information about Docker use in general, see the <a class="reference external" href="https://docs.docker.com">Docker documentation</a>.</p>
<p>We strongly recommend using the CLI to profile in a container. Best container practice is to split services across containers when they do not require colocation. The Nsight Systems GUI is not needed to profile and brings in many dependencies, so the CLI is recommended. If you wish the GUI can be in a separate side-car container you use after to view your report. All you need is a shared folder between the containers. See section on GUI VNC Container below for more information.</p>
<p><strong>Enable Docker Collection</strong></p>
<p>When starting the Docker to perform a Nsight Systems collection, additional steps are required to enable the <code class="docutils literal notranslate"><span class="pre">perf_event_open</span></code> system call. This is required in order to utilize the Linux kernel’s perf subsystem which provides sampling information to Nsight Systems.</p>
<p>There are three ways to enable the <code class="docutils literal notranslate"><span class="pre">perf_event_open</span></code> syscall. You can enable it by using the <code class="docutils literal notranslate"><span class="pre">--privileged=true</span></code> switch, adding <code class="docutils literal notranslate"><span class="pre">--cap-add=SYS_ADMIN</span></code> switch to your docker run command file, or you can enable it by setting the seccomp security profile if your system meets the requirements.</p>
<p>Secure computing mode (seccomp) is a feature of the Linux kernel that can be used to restrict an application’s access. This feature is available only if the kernel is enabled with seccomp support. To check for seccomp support:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">grep</span><span class="w"> </span><span class="n">CONFIG_SECCOMP</span><span class="o">=</span><span class="w"> </span><span class="o">/</span><span class="n">boot</span><span class="o">/</span><span class="n">config</span><span class="o">-</span><span class="n">$</span><span class="p">(</span><span class="n">uname</span><span class="w"> </span><span class="o">-</span><span class="n">r</span><span class="p">)</span><span class="w"></span>
</pre></div>
</div>
<p>The official Docker documentation says:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="s">"Seccomp profiles require seccomp 2.2.1 which is not available on Ubuntu 14.04, Debian Wheezy, or Debian Jessie. To use seccomp on these distributions, you must download the latest static Linux binaries (rather than packages)."</span><span class="w"></span>
</pre></div>
</div>
<p>Download the default seccomp profile file, default.json, relevant to your Docker version. If <code class="docutils literal notranslate"><span class="pre">perf_event_open</span></code> is already listed in the file as guarded by <code class="docutils literal notranslate"><span class="pre">CAP_SYS_ADMIN</span></code>, then remove the <code class="docutils literal notranslate"><span class="pre">perf_event_open</span></code> line. Add the following lines under “syscalls” and save the resulting file as <code class="docutils literal notranslate"><span class="pre">default_with_perf.json</span></code>.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="p">{</span><span class="w"></span>
<span class="w"> </span><span class="s">"name"</span><span class="o">:</span><span class="w"> </span><span class="s">"perf_event_open"</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s">"action"</span><span class="o">:</span><span class="w"> </span><span class="s">"SCMP_ACT_ALLOW"</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s">"args"</span><span class="o">:</span><span class="w"> </span><span class="p">[]</span><span class="w"></span>
<span class="p">},</span><span class="w"></span>
</pre></div>
</div>
<p>Then you will be able to use the following switch when starting the Docker to apply the new seccomp profile.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">security</span><span class="o">-</span><span class="n">opt</span><span class="w"> </span><span class="n">seccomp</span><span class="o">=</span><span class="n">default_with_perf</span><span class="p">.</span><span class="n">json</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Launch Docker Collection</strong></p>
<p>Here is an example command that has been used to launch a Docker for testing with Nsight Systems:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">nvidia</span><span class="o">-</span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">--</span><span class="n">network</span><span class="o">=</span><span class="n">host</span><span class="w"> </span><span class="o">--</span><span class="n">security</span><span class="o">-</span><span class="n">opt</span><span class="w"> </span><span class="n">seccomp</span><span class="o">=</span><span class="n">default_with_perf</span><span class="p">.</span><span class="n">json</span><span class="w"> </span><span class="o">--</span><span class="n">rm</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">caffe</span><span class="o">-</span><span class="n">demo2</span><span class="w"> </span><span class="n">bash</span><span class="w"></span>
</pre></div>
</div>
<p>There is a known issue where Docker collections terminate prematurely with older versions of the driver and the CUDA Toolkit. If collection is ending unexpectedly, please update to the latest versions.</p>
<p>After the Docker has been started, use the Nsight Systems CLI to launch a collection within the Docker. The resulting .qdstrm file can be imported into the Nsight Systems host like any other CLI result.</p>
<section id="gui-vnc-container">
<h3>GUI VNC container<a class="headerlink" href="#gui-vnc-container" title="Permalink to this heading"></a></h3>
<p>Nsight Systems provides a build script to build a self isolated Docker container with the Nsight Systems GUI and VNC server.</p>
<p>You can find the build.py script in the <code class="docutils literal notranslate"><span class="pre">host-linux-x64/Scripts/VncContainer</span></code> directory (or similar on other architectures) under your Nsight Systems installation directory. You will need to have <a class="reference external" href="https://www.docker.com">Docker</a>, and Python 3.5 or later.</p>
<p><strong>Available Parameters</strong></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short Name</p></th>
<th class="head"><p>Full Name</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--vnc-password</span></code></p></td>
<td><p>(optional) Default password for VNC access (at least 6 characters). If it is specified and empty - will be asked during the build. Can be changed when running a container.</p></td>
</tr>
<tr class="row-odd"><td><p>-aba</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--additional-build-arguments</span></code></p></td>
<td><p>(optional) Additional arguments, which will be passed to the “docker build” command.</p></td>
</tr>
<tr class="row-even"><td><p>-hd</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nsys-host-directory</span></code></p></td>
<td><p>(optional) The directory with Nsight Systems host binaries (with GUI).</p></td>
</tr>
<tr class="row-odd"><td><p>-td</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--nsys-target-directory</span></code></p></td>
<td><p>(optional, repeatable) The directory with Nsight Systems target binaries (can be specified multiple times).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--tigervnc</span></code></p></td>
<td><p>(optional) Use TigerVNC instead of x11vnc.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--http</span></code></p></td>
<td><p>(optional) Install noVNC in the Docker container for HTTP access.</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--rdp</span></code></p></td>
<td><p>(optional) Install xRDP in the Docker for RDP access.</p></td>
</tr>
<tr class="row-odd"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--geometry</span></code></p></td>
<td><p>(optional) Default VNC server resolution in the format WidthxHeight (default 1920x1080).</p></td>
</tr>
<tr class="row-even"><td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--build-directory</span></code></p></td>
<td><p>(optional) The directory to save temporary files (with the write access for the current user). By default, script or tmp directory will be used.</p></td>
</tr>
</tbody>
</table>
<p><strong>Ports</strong></p>
<p>These ports can be published from the container to provide access to the Docker container:</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Port</p></th>
<th class="head"><p>Purpose</p></th>
<th class="head"><p>Condition</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>TCP 5900</p></td>
<td><p>Port for VNC access</p></td>
<td></td>
</tr>
<tr class="row-odd"><td><p>TCP 80 (optional)</p></td>
<td><p>Port for HTTP access to noVNC server</p></td>
<td><p>Container is build with <code class="docutils literal notranslate"><span class="pre">--http</span></code> parameter</p></td>
</tr>
<tr class="row-even"><td><p>TCP 3389 (optional)</p></td>
<td><p>Port for RDP access</p></td>
<td><p>Container is build with <code class="docutils literal notranslate"><span class="pre">--rdp</span></code> parameter</p></td>
</tr>
</tbody>
</table>
<p><strong>Volumes</strong></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Docker folder</p></th>
<th class="head"><p>Purpose</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>/mnt/host</p></td>
<td><p>Root path for shared folders</p></td>
<td><p>Folder owned by the Docker user (inner content can be accessed from Nsight Systems GUI)</p></td>
</tr>
<tr class="row-odd"><td><p>/mnt/host/Projects</p></td>
<td></td>
<td><p>Folder with projects and reports, created by Nsight Systems UI in container</p></td>
</tr>
<tr class="row-even"><td><p>/mnt/host/logs</p></td>
<td><p>Folder with inner services logs</p></td>
<td><p>May be useful to send reports to developers</p></td>
</tr>
</tbody>
</table>
<p><strong>Environment variables</strong></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Variable Name</p></th>
<th class="head"><p>Purpose</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>VNC_PASSWORD</p></td>
<td><p>Password for VNC access (at least 6 characters)</p></td>
</tr>
<tr class="row-odd"><td><p>NSYS_WINDOW_WIDTH</p></td>
<td><p>Width of VNC server display (in pixels)</p></td>
</tr>
<tr class="row-even"><td><p>NSYS_WINDOW_HEIGHT</p></td>
<td><p>Height of VNC server display (in pixels)</p></td>
</tr>
</tbody>
</table>
<p><strong>Examples</strong></p>
<p>With VNC access on port 5916:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">5916</span><span class="o">:</span><span class="mi">5900</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="o">-</span><span class="n">vnc</span><span class="o">:</span><span class="mf">1.0</span><span class="w"></span>
</pre></div>
</div>
<p>With VNC access on port 5916 and HTTP access on port 8080:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">5916</span><span class="o">:</span><span class="mi">5900</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">8080</span><span class="o">:</span><span class="mi">80</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="o">-</span><span class="n">vnc</span><span class="o">:</span><span class="mf">1.0</span><span class="w"></span>
</pre></div>
</div>
<p>With VNC access on port 5916, HTTP access on port 8080 and RDP access on port 33890:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">5916</span><span class="o">:</span><span class="mi">5900</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">8080</span><span class="o">:</span><span class="mi">80</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">33890</span><span class="o">:</span><span class="mi">3389</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="o">-</span><span class="n">vnc</span><span class="o">:</span><span class="mf">1.0</span><span class="w"></span>
</pre></div>
</div>
<p>With VNC access on port 5916, shared “HOME” folder from the host, VNC server resolution 3840x2160, and custom VNC password</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">5916</span><span class="o">:</span><span class="mi">5900</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">v</span><span class="w"> </span><span class="n">$HOME</span><span class="o">:/</span><span class="n">mnt</span><span class="o">/</span><span class="n">host</span><span class="o">/</span><span class="n">home</span><span class="w"> </span><span class="o">-</span><span class="n">e</span><span class="w"> </span><span class="n">NSYS_WINDOW_WIDTH</span><span class="o">=</span><span class="mi">3840</span><span class="w"> </span><span class="o">-</span><span class="n">e</span><span class="w"> </span><span class="n">NSYS_WINDOW_HEIGHT</span><span class="o">=</span><span class="mi">2160</span><span class="w"> </span><span class="o">-</span><span class="n">e</span><span class="w"> </span><span class="n">VNC_PASSWORD</span><span class="o">=</span><span class="mi">7654321</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="o">-</span><span class="n">vnc</span><span class="o">:</span><span class="mf">1.0</span><span class="w"></span>
</pre></div>
</div>
<p>With VNC access on port 5916, shared “HOME” folder from the host, and the projects folder to access reports created by Nsight Systems GUI in container</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">docker</span><span class="w"> </span><span class="n">run</span><span class="w"> </span><span class="o">-</span><span class="n">p</span><span class="w"> </span><span class="mi">5916</span><span class="o">:</span><span class="mi">5900</span><span class="o">/</span><span class="n">tcp</span><span class="w"> </span><span class="o">-</span><span class="n">v</span><span class="w"> </span><span class="n">$HOME</span><span class="o">:/</span><span class="n">mnt</span><span class="o">/</span><span class="n">host</span><span class="o">/</span><span class="n">home</span><span class="w"> </span><span class="o">-</span><span class="n">v</span><span class="w"> </span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">NsysProjects</span><span class="o">:/</span><span class="n">mnt</span><span class="o">/</span><span class="n">host</span><span class="o">/</span><span class="n">Projects</span><span class="w"> </span><span class="o">-</span><span class="n">ti</span><span class="w"> </span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="o">-</span><span class="n">vnc</span><span class="o">:</span><span class="mf">1.0</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="gui-webrtc-container">
<h3>GUI WebRTC container<a class="headerlink" href="#gui-webrtc-container" title="Permalink to this heading"></a></h3>
<p>Instructions for creating a self-isolated Docker container for accessing Nsight Systems through browser using WebRTC.</p>
<p><strong>Prerequisites</strong></p>
<ul class="simple">
<li><p>x86_64 Linux</p></li>
<li><p><a class="reference external" href="https://docs.docker.com/engine/install">Docker</a></p></li>
<li><p>Internet access for downloading Ubuntu packages inside the container.</p></li>
</ul>
<p><strong>Build</strong></p>
<p>To build the docker container use the follwing command:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo ./setup/build-docker.sh
</pre></div>
</div>
<p>The above command will create a docker image, which can be run using ./start-nsys.sh</p>
<p><strong>Build environment variables</strong></p>
<p>Following environment variables can be used to configure build parameters.</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Variable</p></th>
<th class="head"><p>Description</p></th>
<th class="head"><p>Default Value</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>USERNAME</p></td>
<td><p>User name for NVIDIA Nsight Systems GUI. Password can be set on container start</p></td>
<td><p>nvidia</p></td>
</tr>
</tbody>
</table>
<p><strong>Additional docker build arguments</strong></p>
<p>Additional <a class="reference external" href="https://docs.docker.com/engine/reference/commandline/build">Docker Build</a> arguments may be passed to the build-docker.sh. For example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo ./setup/build-docker.sh --network=host
</pre></div>
</div>
<p><strong>Run</strong></p>
<p>To run the docker container:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo ./start-nsys.sh
</pre></div>
</div>
<p>At the end of <code class="docutils literal notranslate"><span class="pre">start-nsys.sh</span></code> it will provide you with a URL to connect to the WebRTC client. It will look something like <code class="docutils literal notranslate"><span class="pre">http://$HOST_IP:8080/</span></code>. You can use this address in your browser to access Nsight Systems GUI interface.</p>
<p><strong>Additional docker run arguments</strong></p>
<p>Additional <a class="reference external" href="https://docs.docker.com/engine/reference/commandline/run">Docker Run</a> arguments may be passed to the <code class="docutils literal notranslate"><span class="pre">start-nsys.sh.</span></code> These argument can be used to mount host directories with Nsight Systems reports to the docker container. For example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo ./start-nsys.sh -v $HOME:/mnt/host/home -v /myawesomereports:/mnt/host/myawesomereports
</pre></div>
</div>
<p><strong>Runtime environment variables</strong></p>
<p>Runtime environment variables can be used to configure runtime parameters.</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Variable</p></th>
<th class="head"><p>Description</p></th>
<th class="head"><p>Default Value</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>PASSWORD</p></td>
<td><p>Password for WebUI. Username can be set only on the build step</p></td>
<td><p>nvidia</p></td>
</tr>
<tr class="row-odd"><td><p>HOST_IP</p></td>
<td><p>IP of the server that will be sent to client. This IP should be accessible from the client side to establish client/server connection.</p></td>
<td><p>The IP address of the first available network interface.</p></td>
</tr>
<tr class="row-even"><td><p>HTTP_PORT</p></td>
<td><p>Port for HTTP access to Nsight Systems user interface.</p></td>
<td><p>8080</p></td>
</tr>
<tr class="row-odd"><td><p>CONNECTION_UDP_PORT</p></td>
<td><p>UDP port which will be used for handling the incoming connection.</p></td>
<td><p>8081</p></td>
</tr>
<tr class="row-even"><td><p>FALLBACK_CONNECTION_TCP_PORT</p></td>
<td><p>TCP port which will be used for handling the incoming connection in case of connection failure over TCP (can be the same port number as CONNECTION_UDP_PORT).</p></td>
<td><p>8081</p></td>
</tr>
<tr class="row-odd"><td><p>SCREEN</p></td>
<td><p>Resolution and refresh rate of the screen used for rendering.</p></td>
<td><p>1920x1080@30</p></td>
</tr>
<tr class="row-even"><td><p>USE_OPENH264_BUILD_CACHE</p></td>
<td><p>Setting this option to false disables caching of openh264 binaries. It should be reenabled on each container start.</p></td>
<td><p>true</p></td>
</tr>
<tr class="row-odd"><td><p>OPENH264_BUILD_CACHE_VOLUME_NAME</p></td>
<td><p>Directory or docker volume name for openh264 binaries cache.</p></td>
<td><p>nvidia-devtools-streamer-openh264-volume</p></td>
</tr>
</tbody>
</table>
<p><strong>Video encoding</strong></p>
<p>By default, the container uses the VP8 codec for video streaming. For an improved experience, the H.264 codec can be enabled.</p>
<ul>
<li><p>If internet is available to download the required libraries:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo docker exec nvidia-devtools-streamer /setup/enable-h264-streaming.sh
</pre></div>
</div>
<p>If USE_OPENH264_BUILD_CACHE was not set to false, openh264 binaries will be cached in OPENH264_BUILD_CACHE_VOLUME_NAME and H.264 codec will be used during future launches of the container.</p>
<p>Currently, only software encoding is supported.</p>
</li>
<li><p>If internet is not available:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo -- sh -c 'MY_IMAGE_NAME=my-openh264-nsys-streamer:1.0 USE_OPENH264_BUILD_CACHE=false ./start-nsys.sh && docker exec nvidia-devtools-streamer /setup/enable-h264-streaming.sh && docker commit nvidia-devtools-streamer $MY_IMAGE_NAME && docker save -o my-openh264-nsys-streamer.tar $MY_IMAGE_NAME'
</pre></div>
</div>
<p>As a result, <code class="docutils literal notranslate"><span class="pre">my-openh264-nsys-streamer.tar</span></code> will contain the image with enabled H.264 codec. This file should be transferred to the target machine without internet access. Then, on a machine without internet access, the container can be started using the following command:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo -- sh -c 'CONTAINER_IMAGE=my-openh264-nsys-streamer:1.0 USE_OPENH264_BUILD_CACHE=false docker load -i my-openh264-nsys-streamer.tar && ./start-nsys.sh'
</pre></div>
</div>
</li>
</ul>
<p><strong>Volumes</strong></p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Docker folder</p></th>
<th class="head"><p>Purpose</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>/mnt/host/logs</p></td>
<td><p>Folder with inner services logs</p></td>
<td><p>May be useful to send reports to NVIDIA developer</p></td>
</tr>
</tbody>
</table>
<p><strong>Example</strong></p>
<p>To run the container on 10.10.10.10 network interface, using 8000 HTTP port, 8888 connection port, without caching openh264 binaries:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ sudo HOST_IP=10.10.10.10 HTTP_PORT=8000 CONNECTION_UDP_PORT=8888 USE_OPENH264_BUILD_CACHE=false ./start-nsys.sh
</pre></div>
</div>
</section>
</section>
<section id="migrating-from-nvidia-nvprof">
<h2>Migrating from NVIDIA nvprof<a class="headerlink" href="#migrating-from-nvidia-nvprof" title="Permalink to this heading"></a></h2>
<section id="using-the-product-name-cli-nvprof-command">
<h3>Using the Nsight Systems CLI nvprof Command<a class="headerlink" href="#using-the-product-name-cli-nvprof-command" title="Permalink to this heading"></a></h3>
<p>The <code class="docutils literal notranslate"><span class="pre">nvprof</span></code> command of the Nsight Systems CLI is intended to help former nvprof users transition to nsys. Many nvprof switches are not supported by nsys, often because they are now part of NVIDIA Nsight Compute.</p>
<p>The full nvprof documentation can be found at <a class="reference external" href="https://docs.nvidia.com/cuda/profiler-users-guide">https://docs.nvidia.com/cuda/profiler-users-guide</a>.</p>
<p>The nvprof transition guide for Nsight Compute can be found at <a class="reference external" href="https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html#nvprof-guide">https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html#nvprof-guide</a>.</p>
<p>Any nvprof switch not listed below is not supported by the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">nvprof</span></code> command. No additional nsys functionality is available through this command. New features will not be added to this command in the future.</p>
</section>
<section id="cli-nvprof-command-switch-options">
<h3>CLI nvprof Command Switch Options<a class="headerlink" href="#cli-nvprof-command-switch-options" title="Permalink to this heading"></a></h3>
<p>After choosing the <code class="docutils literal notranslate"><span class="pre">nvprof</span></code> command switch, the following options are available. When you are ready to move to using Nsight Systems CLI directly, see <a class="reference external" href="index.html#cli-profiling">Command Line Options</a> documentation for the nsys switch(es) given below. Note that the nsys implementation and output may vary from nvprof.</p>
<p><strong>Usage</strong>.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="n">nvprof</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Switch</p></th>
<th class="head"><p>Parameters (Default in Bold)</p></th>
<th class="head"><p>nsys switch</p></th>
<th class="head"><p>Switch Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">--annotate-mpi</span></code></p></td>
<td><p><strong>off</strong>, openmpi, mpich</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace=mpi</span></code> AND <code class="docutils literal notranslate"><span class="pre">--mpi-impl</span></code></p></td>
<td><p>Automatically annotate MPI calls with NVTX markers. Specify the MPI implementation installed on your machine. Only OpenMPI and MPICH implementations are supported.</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">--cpu-thread-tracing</span></code></p></td>
<td><p>on, <strong>off</strong></p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace=osrt</span></code></p></td>
<td><p>Collect information about CPU thread API activity.</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">--profile-api-trace</span></code></p></td>
<td><p>none, runtime, driver,<strong>all</strong></p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace=cuda</span></code></p></td>
<td><p>Turn on/off CUDA runtime and driver API tracing. For Nsight Systems there is no separate CUDA runtime and CUDA driver trace, so selecting <code class="docutils literal notranslate"><span class="pre">runtime</span></code> or <code class="docutils literal notranslate"><span class="pre">driver</span></code> is equivalent to selecting <code class="docutils literal notranslate"><span class="pre">all</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">--profile-from-start</span></code></p></td>
<td><p><strong>on</strong>, off</p></td>
<td><p>if off use <code class="docutils literal notranslate"><span class="pre">--capture-range=cudaProfilerApi</span></code></p></td>
<td><p>Enable/disable profiling from the start of the application. If disabled, the application can use {cu,cuda}Profiler{Start,Stop} to turn on/off profiling.</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">-t</span></code> <code class="docutils literal notranslate"><span class="pre">--timeout</span></code></p></td>
<td><p><nanoseconds> default=<strong>0</strong></p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--duration=seconds</span></code></p></td>
<td><p>If greater than 0, stop the collection and kill the launched application after timeout seconds. nvprof started counting when the CUDA driver is initialized. nsys starts counting immediately.</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">--cpu-profiling</span></code></p></td>
<td><p>on, <strong>off</strong></p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sampling=cpu</span></code></p></td>
<td><p>Turn on/off CPU profiling</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">--openacc-profiling</span></code></p></td>
<td><p><strong>on</strong>, off</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--trace=openacc</span></code> to turn on</p></td>
<td><p>Enable/disable recording information from the OpenACC profiling interface. Note: OpenACC profiling interface depends on the presence of the OpenACC runtime. For supported runtimes, see <strong>CUDA Trace</strong> section of documentation</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">-o</span></code> <code class="docutils literal notranslate"><span class="pre">--export-profile</span></code></p></td>
<td><p><filename></p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output={filename}</span></code> and/or <code class="docutils literal notranslate"><span class="pre">--export=sqlite</span></code></p></td>
<td><p>Export named file to be imported or opened in the Nsight Systems GUI. %q{ENV_VAR} in string will be replaced with the set value of the environment variable. If not set this is an error. %h in the string is replaced with the system hostname. %% in the string is replaced with %. %p in the string is not supported currently. Any other character following % is illegal. The default is report1, with the number incrementing to avoid overwriting files, in users working directory.</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">-f</span></code> <code class="docutils literal notranslate"><span class="pre">--force-overwrite</span></code></p></td>
<td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force-overwrite=true</span></code></p></td>
<td><p>Force overwriting all output files with same name.</p></td>
</tr>
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">-h</span></code> <code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td><p>Print Nsight Systems CLI help</p></td>
</tr>
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">-V</span></code> <code class="docutils literal notranslate"><span class="pre">--version</span></code></p></td>
<td></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--version</span></code></p></td>
<td><p>Print Nsight Systems CLI version information</p></td>
</tr>
</tbody>
</table>
</section>
<section id="next-steps">
<h3>Next Steps<a class="headerlink" href="#next-steps" title="Permalink to this heading"></a></h3>
<p>NVIDIA Visual Profiler (NVVP) and NVIDIA nvprof are deprecated. New GPUs and features will not be supported by those tools. We encourage you to make the move to Nsight Systems now. For additional information, suggestions, and rationale, see the blog series in <a class="reference external" href="../additional-info/index.html#other-resources">Other Resources</a>.</p>
</section>
</section>
<section id="direct3d-trace">
<h2>Direct3D Trace<a class="headerlink" href="#direct3d-trace" title="Permalink to this heading"></a></h2>
<p>Nsight Systems has the ability to trace both the Direct3D 11 API and the Direct3D 12 API on Windows targets.</p>
<section id="d3d11-api-trace">
<h3>D3D11 API trace<a class="headerlink" href="#d3d11-api-trace" title="Permalink to this heading"></a></h3>
<p>Nsight Systems can capture information about Direct3D 11 API calls made by the profiled process. This includes capturing the execution time of D3D11 API functions, performance markers, and frame durations.</p>
<blockquote>
<div><img alt="D3D11 API Trace" class="image" src="../_images/DX11_API_Row.png" />
</div></blockquote>
<p><strong>SLI Trace</strong></p>
<p>Trace SLI queries and peer-to-peer transfers of D3D11 applications. Requires SLI hardware and an active SLI profile definition in the NVIDIA console.</p>
</section>
<section id="d3d12-api-trace">
<h3>D3D12 API Trace<a class="headerlink" href="#d3d12-api-trace" title="Permalink to this heading"></a></h3>
<p>Direct3D 12 is a low-overhead 3D graphics and compute API for Microsoft Windows. Information about Direct3D 12 can be found at the <a class="reference external" href="https://docs.microsoft.com/en-us/windows/desktop/direct3d12/directx-12-programming-guide">Direct3D 12 Programming Guide</a>.</p>
<p>Nsight Systems can capture information about Direct3D 12 usage by the profiled process. This includes capturing the execution time of D3D12 API functions, corresponding workloads executed on the GPU, performance markers, and frame durations.</p>
<img alt="D3D12 overview picture" class="image" src="../_images/dx12_overview.png" />
<p>The Command List Creation row displays time periods when command lists were being created. This enables developers to improve their application’s multi-threaded command list creation. Command list creation time period is measured between the call to <code class="docutils literal notranslate"><span class="pre">ID3D12GraphicsCommandList::Reset</span></code> and the call to <code class="docutils literal notranslate"><span class="pre">ID3D12GraphicsCommandList::Close</span></code>.</p>
<img alt="D3D12 commandlist creation" class="image" src="../_images/dx12_commandlist_creation.png" />
<p>The GPU row shows a compressed view of the D3D12 queue activity, color-coded by the queue type. Expanding it will show the individual queues and their corresponding API calls.</p>
<img alt="D3D12 GPU aggregated" class="image" src="../_images/dx12_commandlist_creation.png" />
<p>A Command Queue row is displayed for each D3D12 command queue created by the profiled application. The row’s header displays the queue’s running index and its type (Direct, Compute, Copy).</p>
<img alt="D3D12 command queue overview" class="image" src="../_images/dx12_commandqueues.png" />
<p>The DX12 API Memory Ops row displays all API memory operations and non-persistent resource mappings. Event ranges in the row are color-coded by the heap type they belong to (Default, Readback, Upload, Custom, or CPU-Visible VRAM), with usage warnings highlighted in yellow. A breakdown of the operations can be found by expanding the row to show rows for each individual heap type.</p>
<p>The following operations and warnings are shown:</p>
<ul class="simple">
<li><p>Calls to <code class="docutils literal notranslate"><span class="pre">ID3D12Device::CreateCommittedResource</span></code>, <code class="docutils literal notranslate"><span class="pre">ID3D12Device4::CreateCommittedResource1</span></code>, and <code class="docutils literal notranslate"><span class="pre">ID3D12Device8::CreateCommittedResource2</span></code></p>
<ul>
<li><p>A warning will be reported if <code class="docutils literal notranslate"><span class="pre">D3D12_HEAP_FLAG_CREATE_NOT_ZEROED</span></code> is not set in the method’s <code class="docutils literal notranslate"><span class="pre">HeapFlags</span></code> parameter</p></li>
</ul>
</li>
<li><p>Calls to <code class="docutils literal notranslate"><span class="pre">ID3D12Device::CreateHeap</span></code> and <code class="docutils literal notranslate"><span class="pre">ID3D12Device4::CreateHeap1</span></code></p>
<ul>
<li><p>A warning will be reported if <code class="docutils literal notranslate"><span class="pre">D3D12_HEAP_FLAG_CREATE_NOT_ZEROED</span></code> is not set in the <code class="docutils literal notranslate"><span class="pre">Flags</span></code> field of the method’s <code class="docutils literal notranslate"><span class="pre">pDesc</span></code> parameter</p></li>
</ul>
</li>
<li><p>Calls to <code class="docutils literal notranslate"><span class="pre">ID3D12Resource::ReadFromSubResource</span></code></p>
<ul>
<li><p>A warning will be reported if the read is to a <code class="docutils literal notranslate"><span class="pre">D3D12_CPU_PAGE_PROPERTY_WRITE_COMBINE</span></code> CPU page or from a <code class="docutils literal notranslate"><span class="pre">D3D12_HEAP_TYPE_UPLOAD</span></code> resource</p></li>
</ul>
</li>
<li><p>Calls to <code class="docutils literal notranslate"><span class="pre">ID3D12Resource::WriteToSubResource</span></code></p>
<ul>
<li><p>A warning will be reported if the write is from a <code class="docutils literal notranslate"><span class="pre">D3D12_CPU_PAGE_PROPERTY_WRITE_BACK</span></code> CPU page or to a <code class="docutils literal notranslate"><span class="pre">D3D12_HEAP_TYPE_READBACK</span></code> resource</p></li>
</ul>
</li>
<li><p>Calls to <code class="docutils literal notranslate"><span class="pre">ID3D12Resource::Map</span></code> and <code class="docutils literal notranslate"><span class="pre">ID3D12Resource::Unmap</span></code> will be matched into [Map, Unmap] ranges for non-persistent mappings. If a mapping range is nested, only the most external range (reference count = 1) will be shown.</p></li>
</ul>
<img alt="D3D12 memory operations and usage warning" class="image" src="../_images/dx12_memory_operations.png" />
<p>The API row displays time periods where <code class="docutils literal notranslate"><span class="pre">ID3D12CommandQueue::ExecuteCommandLists</span></code> was called. The GPU Workload row displays time periods where workloads were executed by the GPU. The workload’s type (Graphics, Compute, Copy, etc.) is displayed on the bar representing the workload’s GPU execution.</p>
<img alt="D3D12 API and Workload" class="image" src="../_images/dx12_api_and_related_workload.png" />
<p>In addition, you can see the PIX command queue CPU-side performance markers, GPU-side performance markers and the GPU Command List performance markers, each in their row.</p>
<img alt="D3D12 CPU marker" class="image" src="../_images/dx12_cpu_marker.png" />
<img alt="D3D12 GPU marker" class="image" src="../_images/dx12_gpu_marker.png" />
<img alt="D3D12 commandlist marker" class="image" src="../_images/dx12_commandlist_marker.png" />
<p>Clicking on a GPU workload highlights the corresponding <code class="docutils literal notranslate"><span class="pre">ID3D12CommandQueue::ExecuteCommandLists</span></code>, <code class="docutils literal notranslate"><span class="pre">ID3D12GraphicsCommandList::Reset</span></code> and <code class="docutils literal notranslate"><span class="pre">ID3D12GraphicsCommandList::Close</span> <span class="pre">API</span></code> calls, and vice versa.</p>
<img alt="D3D12 correlation" class="image" src="../_images/dx12_correlation.png" />
<p>Detecting which CPU thread was blocked by a fence can be difficult in complex apps that run tens of CPU threads. The timeline view displays the 3 operations involved:</p>
<ul class="simple">
<li><p>The CPU thread pushing a signal command and fence value into the command queue. This is displayed on the DX12 Synchronization sub-row of the calling thread.</p></li>
<li><p>The GPU executing that command, setting the fence value and signaling the fence. This is displayed on the GPU Queue Synchronization sub-row.</p></li>
<li><p>The CPU thread calling a Win32 wait API to block-wait until the fence is signaled. This is displayed on the Thread’s OS runtime libraries row.</p></li>
</ul>
<p>Clicking one of these will highlight it and the corresponding other two calls.</p>
<img alt="D3D12 fence sync" class="image" src="../_images/d3d12_fence_synchronization.png" />
</section>
</section>
<section id="wddm-queues">
<h2>WDDM Queues<a class="headerlink" href="#wddm-queues" title="Permalink to this heading"></a></h2>
<p>The Windows Display Driver Model (WDDM) architecture uses queues to send work packets from the CPU to the GPU. Each D3D device in each process is associated with one or more contexts. Graphics, compute, and copy commands that the profiled application uses are associated with a context, batched in a command buffer, and pushed into the relevant queue associated with that context.</p>
<p>Nsight Systems can capture the state of these queues during the trace session.</p>
<p>Enabling the “Collect additional range of ETW events” option will also capture extended DxgKrnl events from the <code class="docutils literal notranslate"><span class="pre">Microsoft-Windows-DxgKrnl</span></code> provider, such as context status, allocations, sync wait, signal events, etc.</p>
<blockquote>
<div><img alt="WDDM Queues" class="image" src="../_images/WDDM_Queues.png" />
</div></blockquote>
<p>A command buffer in a WDDM queues may have one the following types:</p>
<ul class="simple">
<li><p>Render</p></li>
<li><p>Deferred</p></li>
<li><p>System</p></li>
<li><p>MMIOFlip</p></li>
<li><p>Wait</p></li>
<li><p>Signal</p></li>
<li><p>Device</p></li>
<li><p>Software</p></li>
</ul>
<p>It may also be marked as a Present buffer, indicating that the application has finished rendering and requests to display the source surface.</p>
<p>See the Microsoft documentation for the WDDM architecture and the DXGKETW_QUEUE_PACKET_TYPE enumeration.</p>
<p>To retain the .etl trace files captured, so that they can be viewed in other tools (e.g. GPUView), change the “Save ETW log files in project folder” option under “Profile Behavior” in Nsight Systems’s global Options dialog. The .etl files will appear in the same folder as the .nsys-rep file, accessible by right-clicking the report in the Project Explorer and choosing “Show in Folder…”. Data collected from each ETW provider will appear in its own .etl file, and an additional .etl file named “Report XX-Merged-*.etl”, containing the events from all captured sources, will be created as well.</p>
</section>
<section id="wddm-hw-scheduler">
<h2>WDDM HW Scheduler<a class="headerlink" href="#wddm-hw-scheduler" title="Permalink to this heading"></a></h2>
<p>When GPU Hardware Scheduling is enabled in Windows 10 or newer version, the Windows Display Driver Model (WDDM) uses the DxgKrnl ETW provider to expose report of NVIDIA GPUs’ hardware scheduling context switches.</p>
<p>Nsight Systems can capture these context switch events, and display under the GPUs in the timeline rows titled WDDM HW Scheduler - [HW Queue type]. The ranges under each queue will show the process name and PID assoicated with the GPU work during the time period.</p>
<p>The events will be captured if GPU Hardware Scheduling is enabled in the Windows System Display settings, and “Collect WDDM Trace” is enabled in the Nsight Systems Project Settings.</p>
<blockquote>
<div><img alt="WDDM HW Scheduler row for 3D HW Queue" class="image" src="../_images/wddm-hw-scheduler.png" />
</div></blockquote>
</section>
<section id="vulkan-api-trace">
<h2>Vulkan API Trace<a class="headerlink" href="#vulkan-api-trace" title="Permalink to this heading"></a></h2>
<section id="vulkan-overview">
<h3>Vulkan Overview<a class="headerlink" href="#vulkan-overview" title="Permalink to this heading"></a></h3>
<p>Vulkan is a low-overhead, cross-platform 3D graphics and compute API, targeting a wide variety of devices from PCs to mobile phones and embedded platforms. The Vulkan API is defined by the Khronos Group. Information about Vulkan and the Khronos Group can be found at the <a class="reference external" href="https://www.khronos.org/vulkan/">Khronos Vulkan Site</a>.</p>
<p>Nsight Systems can capture information about Vulkan usage by the profiled process. This includes capturing the execution time of Vulkan API functions, corresponding GPU workloads, debug util labels, and frame durations. Vulkan profiling is supported on both Windows and x86 Linux operating systems.</p>
<img alt="Vulkan overview picture" class="image" src="../_images/vulkan_overview.png" />
<p>The Command Buffer Creation row displays time periods when command buffers were being created. This enables developers to improve their application’s multi-threaded command buffer creation. Command buffer creation time period is measured between the call to <code class="docutils literal notranslate"><span class="pre">vkBeginCommandBuffer</span></code> and the call to <code class="docutils literal notranslate"><span class="pre">vkEndCommandBuffer</span></code>.</p>
<img alt="Vulkan command buffer creation" class="image" src="../_images/vulkan_command_buffer_creation.png" />
<p>A Queue row is displayed for each Vulkan queue created by the profiled application. The API sub-row displays time periods where <code class="docutils literal notranslate"><span class="pre">vkQueueSubmit</span></code> was called. The GPU Workload sub-row displays time periods where workloads were executed by the GPU.</p>
<img alt="Vulkan API and Workload" class="image" src="../_images/vulkan_api_and_related_workload.png" />
<p>In addition, you can see <a class="reference external" href="https://github.com/KhronosGroup/Vulkan-Docs/blob/master/appendices/VK_EXT_debug_utils.txt">Vulkan debug util labels</a> on both the CPU and the GPU.</p>
<img alt="Vulkan CPU marker" class="image" src="../_images/vulkan_cpu_marker.png" />
<p>Clicking on a GPU workload highlights the corresponding <code class="docutils literal notranslate"><span class="pre">vkQueueSubmit</span></code> call, and vice versa.</p>
<img alt="Vulkan correlation" class="image" src="../_images/vulkan_correlation.png" />
<p>The Vulkan Memory Operations row contains an aggregation of all the Vulkan host-side memory operations, such as host-blocking writes and reads or non-persistent map-unmap ranges.</p>
<p>The row is separated into sub-rows by heap index and memory type - the tooltip for each row and the ranges inside show the heap flags and the memory property flags.</p>
<img alt="Vulkan Memory Operations" class="image" src="../_images/vulkan_memory_operations.png" />
<img alt="Vulkan Memory Operations" class="image" src="../_images/vulkan_memory_operations_row.png" />
</section>
<section id="pipeline-creation-feedback">
<h3>Pipeline Creation Feedback<a class="headerlink" href="#pipeline-creation-feedback" title="Permalink to this heading"></a></h3>
<p>When tracing target application calls to Vulkan pipeline creation APIs, Nsight Systems leverages the Pipeline Creation Feedback extension to collect more details about the duration of individual pipeline creation stages.</p>
<p>See <a class="reference external" href="https://www.khronos.org/registry/vulkan/specs/1.1-extensions/html/vkspec.html#VkPipelineCreationFeedbackEXT">Pipeline Creation Feedback extension</a> for details about this extension.</p>
<p>Vulkan pipeline creation feedback is available on NVIDIA driver release 435 or later.</p>
<img alt="Vulkan pipeline creation feedback" class="image" src="../_images/vulkan-pipeline-creation-feedback.png" />
</section>
<section id="vulkan-gpu-trace-notes">
<h3>Vulkan GPU Trace Notes<a class="headerlink" href="#vulkan-gpu-trace-notes" title="Permalink to this heading"></a></h3>
<ul class="simple">
<li><p>Vulkan GPU trace is available only when tracing apps that use NVIDIA GPUs.</p></li>
<li><p>The endings of Vulkan Command Buffers execution ranges on Compute and Transfer queues may appear earlier on the timeline than their actual occurrence.</p></li>
</ul>
</section>
</section>
<section id="stutter-analysis">
<h2>Stutter Analysis<a class="headerlink" href="#stutter-analysis" title="Permalink to this heading"></a></h2>
<p class="title sectiontitle rubric" id="stutter-analysis-overview">Stutter Analysis Overview</p>
<p>Nsight Systems on Windows targets displays stutter analysis visualization aids for profiled graphics applications that use either OpenGL, D3D11, D3D12 or Vulkan, as detailed below in the following sections.</p>
<section id="fps-overview">
<h3>FPS Overview<a class="headerlink" href="#fps-overview" title="Permalink to this heading"></a></h3>
<p>The Frame Duration section displays frame durations on both the CPU and the GPU.</p>
<blockquote>
<div><img alt="FPS overview" class="image" src="../_images/fps_overview.png" />
</div></blockquote>
<p>The frame duration row displays live FPS statistics for the current timeline viewport. Values shown are:</p>
<ol class="arabic simple">
<li><p>Number of CPU frames shown of the total number captured</p></li>
<li><p>Average, minimal, and maximal CPU frame time of the currently displayed time range</p></li>
<li><p>Average FPS value for the currently displayed frames</p></li>
<li><p>The 99th percentile value of the frame lengths (such that only 1% of the frames in the range are longer than this value).</p></li>
</ol>
<p>The values will update automatically when scrolling, zooming or filtering the timeline view.</p>
<blockquote>
<div><img alt="FPS stutter row" class="image" src="../_images/stutter_row.png" />
</div></blockquote>
<p>The stutter row highlights frames that are significantly longer than the other frames in their immediate vicinity.</p>
<p>The stutter row uses an algorithm that compares the duration of each frame to the median duration of the surrounding 19 frames. Duration difference under 4 milliseconds is never considered a stutter, to avoid cluttering the display with frames whose absolute stutter is small and not noticeable to the user.</p>
<p>For example, if the stutter threshold is set at 20%:</p>
<ol class="arabic simple">
<li><p>Median duration is 10 ms. Frame with 13 ms time will not be reported (relative difference > 20%, absolute difference < 4 ms)</p></li>
<li><p>Median duration is 60 ms. Frame with 71 ms time will not be reported (relative difference < 20%, absolute difference > 4 ms)</p></li>
<li><p>Median duration is 60 ms. Frame with 80 ms is a stutter (relative difference > 20%, absolute difference > 4 ms, both conditions met)</p></li>
</ol>
<p><strong>OSC detection</strong></p>
<p>The “19 frame window median” algorithm by itself may not work well with some cases of “oscillation” (consecutive fast and slow frames), resulting in some false positives. The median duration is not meaningful in cases of oscillation and can be misleading.</p>
<p>To address the issue and identify if oscillating frames, the following method is applied:</p>
<ol class="arabic simple">
<li><p>For every frame, calculate the median duration, 1st and 3rd quartiles of 19-frames window.</p></li>
<li><p>Calculate the delta and ratio between 1st and 3rd quartiles.</p></li>
<li><p>If the 90th percentile of 3rd - 1st quartile delta array > 4 ms AND the 90th percentile of 3rd/1st quartile array > 1.2 (120%) then mark the results with “OSC” text.</p></li>
</ol>
<p>Right-clicking the Frame Duration row caption lets you choose the target frame rate (30, 60, 90 or custom frames per second).</p>
<img alt="FPS pick" class="image" src="../_images/fps_pick.png" />
<p>By clicking the Customize FPS Display option, a customization dialog pops up. In the dialog, you can now define the frame duration threshold to customize the view of the potentially problematic frames. In addition, you can define the threshold for the stutter analysis frames.</p>
<img alt="fps_customizations" class="image" src="../_images/fps_customizations.png" />
<p>Frame duration bars are color coded:</p>
<ul class="simple">
<li><p>Green, the frame duration is shorter than required by the target FPS ratio.</p></li>
<li><p>Yellow, duration is slightly longer than required by the target FPS rate.</p></li>
<li><p>Red, duration far exceeds that required to maintain the target FPS rate.</p></li>
</ul>
<img alt="Bad FPS" class="image" src="../_images/bad_fps.png" />
<p>The CPU Frame Duration row displays the CPU frame duration measured between the ends of consecutive frame boundary calls:</p>
<ul class="simple">
<li><p>The OpenGL frame boundaries are <code class="docutils literal notranslate"><span class="pre">eglSwapBuffers/glXSwapBuffers/SwapBuffers</span></code> calls.</p></li>
<li><p>The D3D11 and D3D12 frame boundaries are <code class="docutils literal notranslate"><span class="pre">IDXGISwapChainX::Present</span></code> calls.</p></li>
<li><p>The Vulkan frame boundaries are <code class="docutils literal notranslate"><span class="pre">vkQueuePresentKHR</span></code> calls.</p></li>
</ul>
<p>The timing of the actual calls to the frame boundary calls can be seen in the blue bar at the bottom of the CPU frame duration row</p>
<p>The GPU Frame Duration row displays the time measured between</p>
<ul class="simple">
<li><p>The start time of the first GPU workload execution of this frame.</p></li>
<li><p>The start time of the first GPU workload execution of the next frame.</p></li>
</ul>
<p><strong>Reflex SDK</strong></p>
<p>NVIDIA Reflex SDK is a series of NVAPI calls that allow applications to integrate the Ultra Low Latency driver feature more directly into their game to further optimize synchronization between simulation and rendering stages and lower the latency between user input and final image rendering. For more details about Reflex SDK, see <a class="reference external" href="https://developer.nvidia.com/reflex/">Reflex SDK Site</a>.</p>
<p>Nsight Systems will automatically capture NVAPI functions when either Direct3D 11, Direct3D 12, or Vulkan API trace are enabled.</p>
<p>The Reflex SDK row displays timeline ranges for the following types of latency markers:</p>
<ul class="simple">
<li><p>RenderSubmit.</p></li>
<li><p>Simulation.</p></li>
<li><p>Present.</p></li>
<li><p>Driver.</p></li>
<li><p>OS Render Queue.</p></li>
<li><p>GPU Render.</p></li>
</ul>
<img alt="Reflex SDK" class="image" src="../_images/reflex_sdk.png" />
<p><strong>Performance Warnings row</strong></p>
<p>This row shows performance warnings and common pitfalls that are automatically detected based on the enabled capture types. Warnings are reported for:</p>
<ul class="simple">
<li><p>ETW performance warnings</p></li>
<li><p>Vulkan calls to <code class="docutils literal notranslate"><span class="pre">vkQueueSubmit</span></code> and D3D12 calls to <code class="docutils literal notranslate"><span class="pre">ID3D12CommandQueue::ExecuteCommandList</span></code> that take a longer time to execute than the total time of the GPU workloads they generated</p></li>
<li><p><a class="reference external" href="topics/d3d12-trace">D3D12 Memory Operation warnings</a></p></li>
<li><p>Usage of Vulkan API functions that may adversely affect performance</p></li>
<li><p>Creation of a Vulkan device with memory zeroing, whether by physical device default or manually</p></li>
<li><p>Vulkan command buffer barrier which can be combined or removed, such as subsequent barriers or read-to-read barriers</p></li>
</ul>
<img alt="Performance Warnings row" class="image" src="../_images/fps_performance_warnings.png" />
</section>
<section id="frame-health">
<h3>Frame Health<a class="headerlink" href="#frame-health" title="Permalink to this heading"></a></h3>
<p>The Frame Health row displays actions that took significantly a longer time during the current frame, compared to the median time of the same actions executed during the surrounding 19-frames. This is a great tool for detecting the reason for frame time stuttering. Such actions may be: shader compilation, present, memory mapping, and more. Nsight Systems measures the accumulated time of such actions in each frame. For example: calculating the accumulated time of shader compilations in each frame and comparing it to the accumulated time of shader compilations in the surrounding 19 frames.</p>
<p>Example of a Vulkan frame health row:</p>
<img alt="Frame Health Vulkan" class="image" src="../_images/vulkan_frame_health.png" />
<img alt="Frame Health DX12" class="image" src="../_images/dx12_frame_health.png" />
</section>
<section id="gpu-memory">
<h3>GPU Memory<a class="headerlink" href="#gpu-memory" title="Permalink to this heading"></a></h3>
<p>Each GPU has two rows detailing its memory utilization: <strong>GPU VRAM</strong>, showing the memory consumed on the device, and <strong>GPU WDDM SYSMEM</strong>, showing the memory consumed on the host computer RAM.</p>
<img alt="Memory Utilization rows" class="image" src="../_images/memory_utilization.png" />
<p>These rows show a green-colored line graph for the memory budget for this memory segment, and an orange-colored line graph for the actual amount of memory used. Note that these graphs are scaled to fit the highest value enconutered, as indicated by the “Y axis” value in the row header. You can use the vertical zoom slider in the top-right of the timeline view to make the row taller and view the graph in more detail.</p>
<img alt="Vertical Zoom slider" class="image" src="../_images/vertical-scroll.png" />
<p>Note that the value in the GPU VRAM row is not the same as the CUDA kernel memory allocation graph, see <a class="reference external" href="topics/cuda-gpu-memory-graph">CUDA GPU Memory Graph</a> for that functionality.</p>
<p>The GPU VRAM row also has several child rows, accessed by expanding the row in the tree view</p>
<img alt="GPU VRAM row expanded" class="image" src="../_images/memory_utilization_gpu_vram.png" />
<p><strong>VidMm Device Suspension</strong></p>
<p>This row displays time ranges when the GPU memory manager suspended all memory transfer operations, pending the completion of a single memory transfer.</p>
<p><strong>Demoted Memory</strong></p>
<p>This row displays the amount of VRAM that was demoted from GPU local memory to non-local memory (possibly due to exceeding the VRAM budget) as a blue-colored line graph. High amounts of demoted memory could be indicative of video memory leaks or poor memory management. Note that the Demoted memory row is scaled to its highest value, similar to the GPU VRAM and GPU WDDM SYSMEM rows.</p>
<p><strong>Resource Allocations</strong></p>
<img alt="Resource Allocations row" class="image" src="../_images/memory_utilization_resource_allocations.png" />
<p>This row shows markers indicating resource allocation events. VRAM resources are shown as green markers while SYSMEM resources are shown in gray. Hovering over a marker or selecting it in the <a class="reference external" href="topics/events-view">Events view</a> will display all the allocation parameters as well as the call stack that led to the allocation event.</p>
<p><strong>Resource Migrations</strong></p>
<img alt="Resource Migrations row" class="image" src="../_images/memory_utilization_resource_migrations.png" />
<p>This row displays a breakdown of resources’ movement between VRAM and SYSMEM, focusing on resource evictions. The main row shows a timeline of total evicted resource memory and count as a red-colored line graph.</p>
<p>Each child row displays a timeline of the status of each resource, as reflected by WDDM events related to it. If the object has been named using PIX or <code class="docutils literal notranslate"><span class="pre">ID3D11Object::SetName</span></code> / <code class="docutils literal notranslate"><span class="pre">ID3D12Object::SetName</span></code>, the name will be shown in the row title. Whether named or not, the row title will also show the resource dimensions, format, priority, and the memory size migrated. If the resource was migrated in parts using subresources, the row can be expanded to show the status for each subresource at any given time.</p>
<p>Expanding the row for a resource will show the individual WDDM events relevant to it and the call stacks that led to each event.</p>
<p>By default, the resources are sorted by Relevance (most / largest migrations). Right-clicking the main Resource Migrations row header allows choosing between the following sorting options:</p>
<ul class="simple">
<li><p>Relevance</p></li>
<li><p>Name</p></li>
<li><p>Format</p></li>
<li><p>Priority</p></li>
<li><p>Earliest allocation timestamp (order of appearance on the host)</p></li>
<li><p>Earliest migration timestamp (order of appearance on the device)</p></li>
</ul>
<p>The top 5 resources are shown initially. If more than 5 resources exist, a row showing the number of hidden resources and buttons allowing to show more or fewer of them will appear below them. Right-click this row and select “show all” or “show all collapsed” to display all the resources at once.</p>
<p><strong>Memory Transfer</strong></p>
<img alt="Memory Transfer row" class="image" src="../_images/memory_utilization_memory_transfer.png" />
<p>This row shows an overview of all memory transfer operations. Device-to-host transfers are shown in orange, host-to-device transfers are shown in green, discarded device memory is shown in light green, and unknown events are shown in dark gray. The height of each event marker corresponds to the amount of memory that the event affected. Hovering over the marker will show the exact amount.</p>
<p>Expanding the row will show a breakdown of the events by each specific type.</p>
</section>
<section id="vertical-synchronization">
<h3>Vertical Synchronization<a class="headerlink" href="#vertical-synchronization" title="Permalink to this heading"></a></h3>
<p>The VSYNC rows display when the monitor’s vertical synchronizations occur.</p>
<img alt="Vertical Synchronization" class="image" src="../_images/vsync.png" />
</section>
</section>
<section id="openmp-trace">
<h2>OpenMP Trace<a class="headerlink" href="#openmp-trace" title="Permalink to this heading"></a></h2>
<p>Nsight Systems for Linux is capable of capturing information about OpenMP events. This functionality is built on the OpenMP Tools Interface (OMPT), full support is available only for runtime libraries supporting tools interface defined in OpenMP 5.0 or greater.</p>
<p>As an example, LLVM OpenMP runtime library partially implements tools interface. If you use PGI compiler <= 20.4 to build your OpenMP applications, add -mp=libomp switch to use LLVM OpenMP runtime and enable OMPT based tracing. If you use Clang, make sure the LLVM OpenMP runtime library you link to was compiled with tools interface enabled.</p>
<img alt="OpenMP trace selection" class="image" src="../_images/select-openmp-trace.png" />
<p>Only a subset of the OMPT callbacks are processed:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ompt_callback_parallel_begin</span><span class="w"></span>
<span class="n">ompt_callback_parallel_end</span><span class="w"></span>
<span class="n">ompt_callback_sync_region</span><span class="w"></span>
<span class="n">ompt_callback_task_create</span><span class="w"></span>
<span class="n">ompt_callback_task_schedule</span><span class="w"></span>
<span class="n">ompt_callback_implicit_task</span><span class="w"></span>
<span class="n">ompt_callback_master</span><span class="w"></span>
<span class="n">ompt_callback_reduction</span><span class="w"></span>
<span class="n">ompt_callback_task_create</span><span class="w"></span>
<span class="n">ompt_callback_cancel</span><span class="w"></span>
<span class="n">ompt_callback_mutex_acquire</span><span class="p">,</span><span class="w"> </span><span class="n">ompt_callback_mutex_acquired</span><span class="w"></span>
<span class="n">ompt_callback_mutex_acquired</span><span class="p">,</span><span class="w"> </span><span class="n">ompt_callback_mutex_released</span><span class="w"></span>
<span class="n">ompt_callback_mutex_released</span><span class="w"></span>
<span class="n">ompt_callback_work</span><span class="w"></span>
<span class="n">ompt_callback_dispatch</span><span class="w"></span>
<span class="n">ompt_callback_flush</span><span class="w"></span>
</pre></div>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The raw OMPT events are used to generate ranges indicating the runtime of OpenMP operations and constructs.</p>
</div>
<p>Example screenshot:</p>
<img alt="OpenMP API trace" class="image" src="../_images/openmp-trace.png" />
</section>
<section id="os-runtime-libraries-trace">
<h2>OS Runtime Libraries Trace<a class="headerlink" href="#os-runtime-libraries-trace" title="Permalink to this heading"></a></h2>
<p>On Linux, OS runtime libraries can be traced to gather information about low-level userspace APIs. This traces the system call wrappers and thread synchronization interfaces exposed by the C runtime and POSIX Threads (pthread) libraries. This does not perform a complete runtime library API trace, but instead focuses on the functions that can take a long time to execute, or could potentially cause your thread be unscheduled from the CPU while waiting for an event to complete. OS runtime trace is not available for Windows targets.</p>
<p>OS runtime tracing complements and enhances sampling information by:</p>
<ol class="arabic simple">
<li><p>Visualizing when the process is communicating with the hardware, controlling resources, performing multi-threading synchronization or interacting with the kernel scheduler.</p></li>
<li><p>Adding additional thread states by correlating how OS runtime libraries traces affect the thread scheduling:</p>
<ul class="simple">
<li><p><strong>Waiting</strong> — the thread is not scheduled on a CPU, it is inside of an OS runtime libraries trace and is believed to be waiting on the firmware to complete a request.</p></li>
<li><p><strong>In OS runtime library function</strong> — the thread is scheduled on a CPU and inside of an OS runtime libraries trace. If the trace represents a system call, the process is likely running in kernel mode.</p></li>
</ul>
</li>
<li><p>Collecting backtraces for long OS runtime libraries call. This provides a way to gather blocked-state backtraces, allowing you to gain more context about why the thread was blocked so long, yet avoiding unnecessary overhead for short events.</p></li>
</ol>
<blockquote>
<div><img alt="OS runtime libraries row" class="image" src="../_images/os-runtime-row-x86.png" />
</div></blockquote>
<p>To enable OS runtime libraries tracing from Nsight Systems:</p>
<p><strong>CLI</strong> — Use the <code class="docutils literal notranslate"><span class="pre">-t</span></code>, <code class="docutils literal notranslate"><span class="pre">--trace</span></code> option with the <code class="docutils literal notranslate"><span class="pre">osrt</span></code> parameter. See <a class="reference external" href="topics/cli-profiling">Command Line Options</a> for more information.</p>
<p><strong>GUI</strong> — Select the <strong>Collect OS runtime libraries trace</strong> checkbox.</p>
<blockquote>
<div><img alt="Configure OS runtime libraries trace" class="image" src="../_images/configure-os-runtime-trace.png" />
</div></blockquote>
<p>You can also use <strong>Skip if shorter than</strong>. This will skip calls shorter than the given threshold. Enabling this option will improve performances as well as reduce noise on the timeline. We strongly encourage you to skip OS runtime libraries call shorter than 1 μs.</p>
<section id="locking-a-resource">
<h3>Locking a Resource<a class="headerlink" href="#locking-a-resource" title="Permalink to this heading"></a></h3>
<p>The functions listed below receive a special treatment. If the tool detects that the resource is already acquired by another thread and will induce a blocking call, we always trace it. Otherwise, it will never be traced.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">pthread_mutex_lock</span><span class="w"></span>
<span class="n">pthread_rwlock_rdlock</span><span class="w"></span>
<span class="n">pthread_rwlock_wrlock</span><span class="w"></span>
<span class="n">pthread_spin_lock</span><span class="w"></span>
<span class="n">sem_wait</span><span class="w"></span>
</pre></div>
</div>
<p>Note that even if a call is determined as potentially blocking, there is a chance that it may not actually block after a few cycles have elapsed. The call will still be traced in this scenario.</p>
</section>
<section id="limitations">
<h3>Limitations<a class="headerlink" href="#limitations" title="Permalink to this heading"></a></h3>
<ul class="simple">
<li><p>Nsight Systems only traces syscall wrappers exposed by the C runtime. It is not able to trace syscall invoked through assembly code.</p></li>
<li><p>Additional thread states, as well as backtrace collection on long calls, are only enabled if sampling is turned on.</p></li>
<li><p>It is not possible to configure the depth and duration threshold when collecting backtraces. Currently, only OS runtime libraries calls longer than 80 μs will generate a backtrace with a maximum of 24 frames. This limitation will be removed in a future version of the product.</p></li>
<li><p>It is required to compile your application and libraries with the <code class="docutils literal notranslate"><span class="pre">-funwind-tables</span></code> compiler flag in order for Nsight Systems to unwind the backtraces correctly.</p></li>
</ul>
</section>
<section id="os-runtime-libraries-trace-filters">
<h3>OS Runtime Libraries Trace Filters<a class="headerlink" href="#os-runtime-libraries-trace-filters" title="Permalink to this heading"></a></h3>
<p>The OS runtime libraries tracing is limited to a select list of functions. It also depends on the version of the C runtime linked to the application.</p>
</section>
<section id="os-runtime-default-function-list">
<h3>OS Runtime Default Function List<a class="headerlink" href="#os-runtime-default-function-list" title="Permalink to this heading"></a></h3>
<p><strong>Libc system call wrappers</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">accept</span><span class="w"></span>
<span class="n">accept4</span><span class="w"></span>
<span class="n">acct</span><span class="w"></span>
<span class="n">alarm</span><span class="w"></span>
<span class="n">arch_prctl</span><span class="w"></span>
<span class="n">bind</span><span class="w"></span>
<span class="n">bpf</span><span class="w"></span>
<span class="n">brk</span><span class="w"></span>
<span class="n">chroot</span><span class="w"></span>
<span class="n">clock_nanosleep</span><span class="w"></span>
<span class="n">connect</span><span class="w"></span>
<span class="n">copy_file_range</span><span class="w"></span>
<span class="n">creat</span><span class="w"></span>
<span class="n">creat64</span><span class="w"></span>
<span class="n">dup</span><span class="w"></span>
<span class="n">dup2</span><span class="w"></span>
<span class="n">dup3</span><span class="w"></span>
<span class="n">epoll_ctl</span><span class="w"></span>
<span class="n">epoll_pwait</span><span class="w"></span>
<span class="n">epoll_wait</span><span class="w"></span>
<span class="n">fallocate</span><span class="w"></span>
<span class="n">fallocate64</span><span class="w"></span>
<span class="n">fcntl</span><span class="w"></span>
<span class="n">fdatasync</span><span class="w"></span>
<span class="n">flock</span><span class="w"></span>
<span class="n">fork</span><span class="w"></span>
<span class="n">fsync</span><span class="w"></span>
<span class="n">ftruncate</span><span class="w"></span>
<span class="n">futex</span><span class="w"></span>
<span class="n">ioctl</span><span class="w"></span>
<span class="n">ioperm</span><span class="w"></span>
<span class="n">iopl</span><span class="w"></span>
<span class="n">kill</span><span class="w"></span>
<span class="n">killpg</span><span class="w"></span>
<span class="n">listen</span><span class="w"></span>
<span class="n">membarrier</span><span class="w"></span>
<span class="n">mlock</span><span class="w"></span>
<span class="n">mlock2</span><span class="w"></span>
<span class="n">mlockall</span><span class="w"></span>
<span class="n">mmap</span><span class="w"></span>
<span class="n">mmap64</span><span class="w"></span>
<span class="n">mount</span><span class="w"></span>
<span class="n">move_pages</span><span class="w"></span>
<span class="n">mprotect</span><span class="w"></span>
<span class="n">mq_notify</span><span class="w"></span>
<span class="n">mq_open</span><span class="w"></span>
<span class="n">mq_receive</span><span class="w"></span>
<span class="n">mq_send</span><span class="w"></span>
<span class="n">mq_timedreceive</span><span class="w"></span>
<span class="n">mq_timedsend</span><span class="w"></span>
<span class="n">mremap</span><span class="w"></span>
<span class="n">msgctl</span><span class="w"></span>
<span class="n">msgget</span><span class="w"></span>
<span class="n">msgrcv</span><span class="w"></span>
<span class="n">msgsnd</span><span class="w"></span>
<span class="n">msync</span><span class="w"></span>
<span class="n">munmap</span><span class="w"></span>
<span class="n">nanosleep</span><span class="w"></span>
<span class="n">nfsservctl</span><span class="w"></span>
<span class="n">open</span><span class="w"></span>
<span class="n">open64</span><span class="w"></span>
<span class="n">openat</span><span class="w"></span>
<span class="n">openat64</span><span class="w"></span>
<span class="n">pause</span><span class="w"></span>
<span class="n">pipe</span><span class="w"></span>
<span class="n">pipe2</span><span class="w"></span>
<span class="n">pivot_root</span><span class="w"></span>
<span class="n">poll</span><span class="w"></span>
<span class="n">ppoll</span><span class="w"></span>
<span class="n">prctl</span><span class="w"></span>
<span class="n">pread</span><span class="w"></span>
<span class="n">pread64</span><span class="w"></span>
<span class="n">preadv</span><span class="w"></span>
<span class="n">preadv2</span><span class="w"></span>
<span class="n">preadv64</span><span class="w"></span>
<span class="n">process_vm_readv</span><span class="w"></span>
<span class="n">process_vm_writev</span><span class="w"></span>
<span class="n">pselect6</span><span class="w"></span>
<span class="n">ptrace</span><span class="w"></span>
<span class="n">pwrite</span><span class="w"></span>
<span class="n">pwrite64</span><span class="w"></span>
<span class="n">pwritev</span><span class="w"></span>
<span class="n">pwritev2</span><span class="w"></span>
<span class="n">pwritev64</span><span class="w"></span>
<span class="n">read</span><span class="w"></span>
<span class="n">readv</span><span class="w"></span>
<span class="n">reboot</span><span class="w"></span>
<span class="n">recv</span><span class="w"></span>
<span class="n">recvfrom</span><span class="w"></span>
<span class="n">recvmmsg</span><span class="w"></span>
<span class="n">recvmsg</span><span class="w"></span>
<span class="n">rt_sigaction</span><span class="w"></span>
<span class="n">rt_sigqueueinfo</span><span class="w"></span>
<span class="n">rt_sigsuspend</span><span class="w"></span>
<span class="n">rt_sigtimedwait</span><span class="w"></span>
<span class="n">sched_yield</span><span class="w"></span>
<span class="n">seccomp</span><span class="w"></span>
<span class="n">select</span><span class="w"></span>
<span class="n">semctl</span><span class="w"></span>
<span class="n">semget</span><span class="w"></span>
<span class="n">semop</span><span class="w"></span>
<span class="n">semtimedop</span><span class="w"></span>
<span class="n">send</span><span class="w"></span>
<span class="n">sendfile</span><span class="w"></span>
<span class="n">sendfile64</span><span class="w"></span>
<span class="n">sendmmsg</span><span class="w"></span>
<span class="n">sendmsg</span><span class="w"></span>
<span class="n">sendto</span><span class="w"></span>
<span class="n">shmat</span><span class="w"></span>
<span class="n">shmctl</span><span class="w"></span>
<span class="n">shmdt</span><span class="w"></span>
<span class="n">shmget</span><span class="w"></span>
<span class="n">shutdown</span><span class="w"></span>
<span class="n">sigaction</span><span class="w"></span>
<span class="n">sigsuspend</span><span class="w"></span>
<span class="n">sigtimedwait</span><span class="w"></span>
<span class="n">socket</span><span class="w"></span>
<span class="n">socketpair</span><span class="w"></span>
<span class="n">splice</span><span class="w"></span>
<span class="n">swapoff</span><span class="w"></span>
<span class="n">swapon</span><span class="w"></span>
<span class="n">sync</span><span class="w"></span>
<span class="n">sync_file_range</span><span class="w"></span>
<span class="n">syncfs</span><span class="w"></span>
<span class="n">tee</span><span class="w"></span>
<span class="n">tgkill</span><span class="w"></span>
<span class="n">tgsigqueueinfo</span><span class="w"></span>
<span class="n">tkill</span><span class="w"></span>
<span class="n">truncate</span><span class="w"></span>
<span class="n">umount2</span><span class="w"></span>
<span class="n">unshare</span><span class="w"></span>
<span class="n">uselib</span><span class="w"></span>
<span class="n">vfork</span><span class="w"></span>
<span class="n">vhangup</span><span class="w"></span>
<span class="n">vmsplice</span><span class="w"></span>
<span class="n">wait</span><span class="w"></span>
<span class="n">wait3</span><span class="w"></span>
<span class="n">wait4</span><span class="w"></span>
<span class="n">waitid</span><span class="w"></span>
<span class="n">waitpid</span><span class="w"></span>
<span class="n">write</span><span class="w"></span>
<span class="n">writev</span><span class="w"></span>
<span class="n">_sysctl</span><span class="w"></span>
</pre></div>
</div>
<p><strong>POSIX Threads</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">pthread_barrier_wait</span><span class="w"></span>
<span class="n">pthread_cancel</span><span class="w"></span>
<span class="n">pthread_cond_broadcast</span><span class="w"></span>
<span class="n">pthread_cond_signal</span><span class="w"></span>
<span class="n">pthread_cond_timedwait</span><span class="w"></span>
<span class="n">pthread_cond_wait</span><span class="w"></span>
<span class="n">pthread_create</span><span class="w"></span>
<span class="n">pthread_join</span><span class="w"></span>
<span class="n">pthread_kill</span><span class="w"></span>
<span class="n">pthread_mutex_lock</span><span class="w"></span>
<span class="n">pthread_mutex_timedlock</span><span class="w"></span>
<span class="n">pthread_mutex_trylock</span><span class="w"></span>
<span class="n">pthread_rwlock_rdlock</span><span class="w"></span>
<span class="n">pthread_rwlock_timedrdlock</span><span class="w"></span>
<span class="n">pthread_rwlock_timedwrlock</span><span class="w"></span>
<span class="n">pthread_rwlock_tryrdlock</span><span class="w"></span>
<span class="n">pthread_rwlock_trywrlock</span><span class="w"></span>
<span class="n">pthread_rwlock_wrlock</span><span class="w"></span>
<span class="n">pthread_spin_lock</span><span class="w"></span>
<span class="n">pthread_spin_trylock</span><span class="w"></span>
<span class="n">pthread_timedjoin_np</span><span class="w"></span>
<span class="n">pthread_tryjoin_np</span><span class="w"></span>
<span class="n">pthread_yield</span><span class="w"></span>
<span class="n">sem_timedwait</span><span class="w"></span>
<span class="n">sem_trywait</span><span class="w"></span>
<span class="n">sem_wait</span><span class="w"></span>
</pre></div>
</div>
<p><strong>I/O</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">aio_fsync</span><span class="w"></span>
<span class="n">aio_fsync64</span><span class="w"></span>
<span class="n">aio_suspend</span><span class="w"></span>
<span class="n">aio_suspend64</span><span class="w"></span>
<span class="n">fclose</span><span class="w"></span>
<span class="n">fcloseall</span><span class="w"></span>
<span class="n">fflush</span><span class="w"></span>
<span class="n">fflush_unlocked</span><span class="w"></span>
<span class="n">fgetc</span><span class="w"></span>
<span class="n">fgetc_unlocked</span><span class="w"></span>
<span class="n">fgets</span><span class="w"></span>
<span class="n">fgets_unlocked</span><span class="w"></span>
<span class="n">fgetwc</span><span class="w"></span>
<span class="n">fgetwc_unlocked</span><span class="w"></span>
<span class="n">fgetws</span><span class="w"></span>
<span class="n">fgetws_unlocked</span><span class="w"></span>
<span class="n">flockfile</span><span class="w"></span>
<span class="n">fopen</span><span class="w"></span>
<span class="n">fopen64</span><span class="w"></span>
<span class="n">fputc</span><span class="w"></span>
<span class="n">fputc_unlocked</span><span class="w"></span>
<span class="n">fputs</span><span class="w"></span>
<span class="n">fputs_unlocked</span><span class="w"></span>
<span class="n">fputwc</span><span class="w"></span>
<span class="n">fputwc_unlocked</span><span class="w"></span>
<span class="n">fputws</span><span class="w"></span>
<span class="n">fputws_unlocked</span><span class="w"></span>
<span class="n">fread</span><span class="w"></span>
<span class="n">fread_unlocked</span><span class="w"></span>
<span class="n">freopen</span><span class="w"></span>
<span class="n">freopen64</span><span class="w"></span>
<span class="n">ftrylockfile</span><span class="w"></span>
<span class="n">fwrite</span><span class="w"></span>
<span class="n">fwrite_unlocked</span><span class="w"></span>
<span class="n">getc</span><span class="w"></span>
<span class="n">getc_unlocked</span><span class="w"></span>
<span class="n">getdelim</span><span class="w"></span>
<span class="n">getline</span><span class="w"></span>
<span class="n">getw</span><span class="w"></span>
<span class="n">getwc</span><span class="w"></span>
<span class="n">getwc_unlocked</span><span class="w"></span>
<span class="n">lockf</span><span class="w"></span>
<span class="n">lockf64</span><span class="w"></span>
<span class="n">mkfifo</span><span class="w"></span>
<span class="n">mkfifoat</span><span class="w"></span>
<span class="n">posix_fallocate</span><span class="w"></span>
<span class="n">posix_fallocate64</span><span class="w"></span>
<span class="n">putc</span><span class="w"></span>
<span class="n">putc_unlocked</span><span class="w"></span>
<span class="n">putwc</span><span class="w"></span>
<span class="n">putwc_unlocked</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Miscellaneous</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">forkpty</span><span class="w"></span>
<span class="n">popen</span><span class="w"></span>
<span class="n">posix_spawn</span><span class="w"></span>
<span class="n">posix_spawnp</span><span class="w"></span>
<span class="n">sigwait</span><span class="w"></span>
<span class="n">sigwaitinfo</span><span class="w"></span>
<span class="n">sleep</span><span class="w"></span>
<span class="n">system</span><span class="w"></span>
<span class="n">usleep</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
<section id="nvtx-trace">
<h2>NVTX Trace<a class="headerlink" href="#nvtx-trace" title="Permalink to this heading"></a></h2>
<p>The NVIDIA Tools Extension Library (NVTX) is a powerful mechanism that allows users to manually instrument their application. Nsight Systems can then collect the information and present it on the timeline.</p>
<p>Nsight Systems supports version 3.0 of the NVTX specification.</p>
<p>The following features are supported:</p>
<ul>
<li><p>Domains</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainCreate</span><span class="p">(),</span><span class="w"> </span><span class="n">nvtxDomainDestroy</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainRegisterString</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Push-pop ranges (nested ranges that start and end in the same thread).</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxRangePush</span><span class="p">(),</span><span class="w"> </span><span class="n">nvtxRangePushEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxRangePop</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainRangePushEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainRangePop</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Start-end ranges (ranges that are global to the process and are not restricted to a single thread)</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxRangeStart</span><span class="p">(),</span><span class="w"> </span><span class="n">nvtxRangeStartEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxRangeEnd</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainRangeStartEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainRangeEnd</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Marks</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxMark</span><span class="p">(),</span><span class="w"> </span><span class="n">nvtxMarkEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainMarkEx</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Thread names</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxNameOsThread</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Categories</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxNameCategory</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvtxDomainNameCategory</span><span class="p">()</span><span class="w"></span>
</pre></div>
</div>
</li>
</ul>
<p>To learn more about specific features of NVTX, please refer to the NVTX header file: <code class="docutils literal notranslate"><span class="pre">nvToolsExt.h</span></code> or the <a class="reference external" href="https://nvidia.github.io/NVTX/">NVTX documentation</a>.</p>
<p>To use NVTX in your application, follow these steps:</p>
<ol class="arabic">
<li><p>Add <code class="docutils literal notranslate"><span class="pre">#include</span> <span class="pre">"nvtx3/nvToolsExt.h"</span></code> in your source code. The nvtx3 directory is located in the Nsight Systems package in the Target-<architecture>/nvtx/include directory and is available via github at <a class="reference external" href="http://github.com/NVIDIA/NVTX">http://github.com/NVIDIA/NVTX</a>.</p></li>
<li><p>Add the following compiler flag: <code class="docutils literal notranslate"><span class="pre">-ldl</span></code></p></li>
<li><p>Add calls to the NVTX API functions. For example, try adding <code class="docutils literal notranslate"><span class="pre">nvtxRangePush("main")</span></code> in the beginning of the <code class="docutils literal notranslate"><span class="pre">main()</span></code> function, and <code class="docutils literal notranslate"><span class="pre">nvtxRangePop()</span></code> just before the return statement in the end.</p>
<p>For convenience in C++ code, consider adding a wrapper that implements RAII (resource acquisition is initialization) pattern, which would guarantee that every range gets closed.</p>
</li>
<li><p>In the project settings, select the <strong>Collect NVTX trace</strong> checkbox.</p></li>
</ol>
<p>In addition, by enabling the “Insert NVTX Marker hotkey” option it is possible to add NVTX markers to a running non-console applications by pressing the F11 key. These will appear in the report under the NVTX Domain named “HotKey markers”.</p>
<p>Typically calls to NVTX functions can be left in the source code even if the application is not being built for profiling purposes, since the overhead is very low when the profiler is not attached.</p>
<p>NVTX is not intended to annotate very small pieces of code that are being called very frequently. A good rule of thumb to use: if code being annotated usually takes less than 1 microsecond to execute, adding an NVTX range around this code should be done carefully.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Range annotations should be matched carefully. If many ranges are opened but not closed, Nsight Systems has no meaningful way to visualize it. A rule of thumb is to not have more than a couple dozen ranges open at any point in time. Nsight Systems does not support reports with many unclosed ranges.</p>
</div>
<p><strong>NVTX Domains and Categories</strong></p>
<p>NVTX domains enable scoping of annotations. Unless specified differently, all events and annotations are in the default domain. Additionally, categories can be used to group events.</p>
<p>Nsight Systems gives the user the ability to include or exclude NVTX events from a particular domain. This can be especially useful if you are profiling across multiple libraries and are only interested in nvtx events from some of them.</p>
<img alt="NVTX domain selection screen" class="image" src="../_images/nvtx-domain.png" />
<p>This functionality is also available from the CLI. See the CLI documentation for <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-include</span></code> and <code class="docutils literal notranslate"><span class="pre">--nvtx-domain-exclude</span></code> for more details.</p>
<p>Categories that are set in by the user will be recognized and displayed in the GUI.</p>
<img alt="NVTX screenshot with domains and categories" class="image" src="../_images/nvtx-screenshot.png" />
</section>
<section id="cuda-trace">
<h2>CUDA Trace<a class="headerlink" href="#cuda-trace" title="Permalink to this heading"></a></h2>
<p>Nsight Systems is capable of capturing information about CUDA execution in the profiled process.</p>
<p>The following information can be collected and presented on the timeline in the report:</p>
<ul class="simple">
<li><p>CUDA API trace — trace of CUDA Runtime and CUDA Driver calls made by the application.</p>
<ul>
<li><p>CUDA Runtime calls typically start with <code class="docutils literal notranslate"><span class="pre">cuda</span></code> prefix (e.g. <code class="docutils literal notranslate"><span class="pre">cudaLaunch</span></code>).</p></li>
<li><p>CUDA Driver calls typically start with <code class="docutils literal notranslate"><span class="pre">cu</span></code> prefix (e.g. <code class="docutils literal notranslate"><span class="pre">cuDeviceGetCount</span></code>).</p></li>
</ul>
</li>
<li><p>CUDA workload trace — trace of activity happening on the GPU, which includes memory operations (e.g., Host-to-Device memory copies) and kernel executions. Within the threads that use the CUDA API, additional child rows will appear in the timeline tree.</p></li>
<li><p>On Nsight Systems Workstation Edition, cuDNN and cuBLAS API tracing and OpenACC tracing.</p></li>
</ul>
<img alt="CUDA thread rows" class="image" src="../_images/cuda-thread-rows.png" />
<p>Near the bottom of the timeline row tree, the GPU node will appear and contain a CUDA node. Within the CUDA node, each CUDA context used within the process will be shown along with its corresponding CUDA streams. Steams will contain memory operations and kernel launches on the GPU. Kernel launches are represented by blue, while memory transfers are displayed in red.</p>
<img alt="CUDA GPU rows" class="image" src="../_images/cuda-gpu-rows.png" />
<p>The easiest way to capture CUDA information is to launch the process from Nsight Systems, and it will setup the environment for you. To do so, simply set up a normal launch and select the <strong>Collect CUDA trace</strong> checkbox.</p>
<p>For Nsight Systems Workstation Edition this looks like:</p>
<blockquote>
<div><img alt="Configure CUDA trace" class="image" src="../_images/configure-cuda-trace-x86.png" />
</div></blockquote>
<p>For Nsight Systems Embedded Platforms Edition this looks like:</p>
<blockquote>
<div><img alt="Configure CUDA trace" class="image" src="../_images/configure-cuda-trace.png" />
</div></blockquote>
<p>Additional configuration parameters are available:</p>
<ul>
<li><p><strong>Collect backtraces for API calls longer than X seconds</strong> - turns on collection of CUDA API backtraces and sets the minimum time a CUDA API event must take before its backtraces are collected. Setting this value too low can cause high application overhead and seriously increase the size of your results file.</p></li>
<li><p><strong>Flush data periodically</strong> — specifies the period after which an attempt to flush CUDA trace data will be made. Normally, in order to collect full CUDA trace, the application needs to finalize the device used for CUDA work (call <code class="docutils literal notranslate"><span class="pre">cudaDeviceReset()</span></code>, and then let the application gracefully exit (as opposed to crashing).</p>
<p>This option allows flushing CUDA trace data even before the device is finalized. However, it might introduce additional overhead to a random CUDA Driver or CUDA Runtime API call.</p>
</li>
<li><p><strong>Skip some API calls</strong> — avoids tracing insignificant CUDA Runtime API calls (namely, <code class="docutils literal notranslate"><span class="pre">cudaConfigureCall()</span></code>, <code class="docutils literal notranslate"><span class="pre">cudaSetupArgument()</span></code>, <code class="docutils literal notranslate"><span class="pre">cudaHostGetDevicePointers()</span></code>). Not tracing these functions allows Nsight Systems to significantly reduce the profiling overhead, without losing any interesting data. (See CUDA Trace Filters, below)</p></li>
<li><p><strong>Collect GPU Memory Usage</strong> - collects information used to generate a graph of CUDA allocated memory across time. Note that this will increase overhead. See section on <strong>CUDA GPU Memory Allocation Graph</strong> below.</p></li>
<li><p><strong>Collect Unified Memory CPU page faults</strong> - collects information on page faults that occur when CPU code tries to access a memory page that resides on the device. See section on <strong>Unified Memory CPU Page Faults</strong> in the <strong>Unified Memory Transfer Trace</strong> documentation below.</p></li>
<li><p><strong>Collect Unified Memory GPU page faults</strong> - collects information on page faults that occur when GPU code tries to access a memory page that resides on the CPU. See section on <strong>Unified Memory GPU Page Faults</strong> in the <strong>Unified Memory Transfer Trace</strong> documentation below.</p></li>
<li><p><strong>Collect CUDA Graph trace</strong> - by default, CUDA tracing will collect and expose information on a whole graph basis. The user can opt to collect on a node per node basis. See section on <strong>CUDA Graph Trace</strong> below.</p></li>
<li><p>For Nsight Systems Workstation Edition, <strong>Collect cuDNN trace</strong>, <strong>Collect cuBLAS trace</strong>, <strong>Collect OpenACC trace</strong> - selects which (if any) extra libraries that depend on CUDA to trace.</p>
<p>OpenACC versions 2.0, 2.5, and 2.6 are supported when using PGI runtime version 15.7 or greater and not compiling statically. In order to differentiate constructs, a PGI runtime of 16.1 or later is required. Note that Nsight Systems Workstation Edition does not support the GCC implementation of OpenACC at this time.</p>
</li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If your application crashes before all collected CUDA trace data has been copied out, some or all data might be lost and not present in the report.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Nsight Systems will not have information about CUDA events that were still in device buffers when analysis terminated. It is a good idea, if using cudaProfilerAPI to control analysis to call cudaDeviceReset before ending analysis.</p>
</div>
<section id="cuda-gpu-memory-allocation-graph">
<h3>CUDA GPU Memory Allocation Graph<a class="headerlink" href="#cuda-gpu-memory-allocation-graph" title="Permalink to this heading"></a></h3>
<p>When the <strong>Collect GPU Memory Usage</strong> option is selected from the <strong>Collect CUDA trace</strong> option set, Nsight Systems will track CUDA GPU memory allocations and deallocations and present a graph of this information in the timeline. This is not the same as the GPU memory graph generated during stutter analysis on the Windows target (see <a class="reference external" href="topics/gpu-memory-utilization">Stutter Memory Trace</a>)</p>
<p>Below, in the report on the left, memory is allocated and freed during the collection. In the report on the right, memory is allocated, but not freed during the collection.</p>
<img alt="CUDA memory allocation graphs where the memory is or is not released during application run" class="image" src="../_images/memory_allocation_graph1.png" />
<p>Here is another example, where allocations are happening on multiple GPUs</p>
<img alt="CUDA memory usage graph on multiple threads" class="image" src="../_images/memory_allocation_graph2.png" />
</section>
<section id="unified-memory-transfer-trace">
<h3>Unified Memory Transfer Trace<a class="headerlink" href="#unified-memory-transfer-trace" title="Permalink to this heading"></a></h3>
<p>For Nsight Systems Workstation Edition, Unified Memory (also called Managed Memory) transfer trace is enabled automatically in Nsight Systems when CUDA trace is selected. It incurs no overhead in programs that do not perform any Unified Memory transfers. Data is displayed in the Managed Memory area of the timeline:</p>
<img alt="UVM trace" class="image" src="../_images/uvm.png" />
<p><strong>HtoD transfer</strong> indicates the CUDA kernel accessed managed memory that was residing on the host, so the kernel execution paused and transferred the data to the device. Heavy traffic here will incur performance penalties in CUDA kernels, so consider using manual cudaMemcpy operations from pinned host memory instead.</p>
<p><strong>PtoP transfer</strong> indicates the CUDA kernel accessed managed memory that was residing on a different device, so the kernel execution paused and transferred the data to this device. Heavy traffic here will incur performance penalties, so consider using manual cudaMemcpyPeer operations to transfer from other devices’ memory instead. The row showing these events is for the destination device - the source device is shown in the tooltip for each transfer event.</p>
<p><strong>DtoH transfer</strong> indicates the CPU accessed managed memory that was residing on a CUDA device, so the CPU execution paused and transferred the data to system memory. Heavy traffic here will incur performance penalties in CPU code, so consider using manual cudaMemcpy operations from pinned host memory instead.</p>
<p>Some Unified Memory transfers are highlighted with red to indicate potential performance issues:</p>
<img alt="Unified Memory transfer migration cause highlight" class="image" src="../_images/uvm_migration_cause_hl.png" />
<p>Transfers with the following migration causes are highlighted:</p>
<ul>
<li><p><strong>Coherence</strong></p>
<p>Unified Memory migration occurred to guarantee data coherence. SMs (streaming multiprocessors) stop until the migration completes.</p>
</li>
<li><p><strong>Eviction</strong></p>
<p>Unified Memory migrated to the CPU because it was evicted to make room for another block of memory on the GPU. This happens due to memory overcommitment which is available on Linux with Compute Capability ≥ 6.</p>
</li>
</ul>
<section id="unified-memory-cpu-page-faults">
<h4>Unified Memory CPU Page Faults<a class="headerlink" href="#unified-memory-cpu-page-faults" title="Permalink to this heading"></a></h4>
<p>The Unified Memory CPU page faults feature in Nsight Systems tracks the page faults that occur when CPU code tries to access a memory page that resides on the device.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Collecting Unified Memory CPU page faults can cause overhead of up to 70% in testing. Please use this functionality only when needed.</p>
</div>
<img alt="Unified Memory CPU Page Faults on timeline" class="image" src="../_images/uvm-cpu-page-faults.png" />
</section>
<section id="unified-memory-gpu-page-faults">
<h4>Unified Memory GPU Page Faults<a class="headerlink" href="#unified-memory-gpu-page-faults" title="Permalink to this heading"></a></h4>
<p>The Unified Memory GPU page faults feature in Nsight Systems tracks the page faults that occur when GPU code tries to access a memory page that resides on the host.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Collecting Unified Memory GPU page faults can cause overhead of up to 70% in testing. Please use this functionality only when needed.</p>
</div>
<img alt="Unified Memory GPU Page Faults on timeline" class="image" src="../_images/uvm-gpu-page-faults.png" />
</section>
</section>
<section id="cuda-graph-trace">
<h3>CUDA Graph Trace<a class="headerlink" href="#cuda-graph-trace" title="Permalink to this heading"></a></h3>
<p>Nsight Systems is capable of capturing information about CUDA graphs in your application at either the graph or node granularity. This can be set in the CLI using the <code class="docutils literal notranslate"><span class="pre">--cuda-graph-trace</span></code> option, or in the GUI by setting the appropriate drop down.</p>
<blockquote>
<div><img alt="Configure CUDA graph trace" class="image" src="../_images/cudagraph-selector.png" />
</div></blockquote>
<p>When CUDA graph trace is set to <code class="docutils literal notranslate"><span class="pre">graph</span></code>, the users sees each graph as one item on the timeline:</p>
<blockquote>
<div><img alt="CUDA Graph trace at the graph level" class="image" src="../_images/cudagraph-graphlevel.png" />
</div></blockquote>
<p>When CUDA graph trace is set to <code class="docutils literal notranslate"><span class="pre">node</span></code>, the users sees each graph as a set of nodes on the timeline:</p>
<blockquote>
<div><img alt="CUDA Graph trace at the node level" class="image" src="../_images/cudagraph-nodelevel.png" />
</div></blockquote>
<p>Tracing CUDA graphs at the graph level rather than the tracing the underlying nodes results in significantly less overhead. This option is only available with CUDA driver 515.43 or higher.</p>
</section>
<section id="cuda-python-backtrace">
<h3>CUDA Python Backtrace<a class="headerlink" href="#cuda-python-backtrace" title="Permalink to this heading"></a></h3>
<p>Nsight Systems for Arm server (SBSA) platforms and x86 Linux targets, is capable of capturing Python backtrace information when CUDA backtrace is being captured.</p>
<p>To enable CUDA Python backtrace from Nsight Systems:</p>
<p><strong>CLI</strong> — Set <code class="docutils literal notranslate"><span class="pre">--python-backtrace=cuda</span></code>.</p>
<p><strong>GUI</strong> — Select the <strong>Collect Python backtrace for selected API calls</strong> checkbox.</p>
<blockquote>
<div><img alt="Configure CUDA Python Backtrace" class="image" src="../_images/configure-cuda-python-backtrace.png" />
</div></blockquote>
<p>Example screenshot:</p>
<img alt="CUDA Python Backtrace" class="image" src="../_images/cuda-python-backtrace.png" />
</section>
<section id="cuda-default-function-list-for-cli">
<h3>CUDA Default Function List for CLI<a class="headerlink" href="#cuda-default-function-list-for-cli" title="Permalink to this heading"></a></h3>
<p><strong>CUDA Runtime API</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">cudaBindSurfaceToArray</span><span class="w"></span>
<span class="n">cudaBindTexture</span><span class="w"></span>
<span class="n">cudaBindTexture2D</span><span class="w"></span>
<span class="n">cudaBindTextureToArray</span><span class="w"></span>
<span class="n">cudaBindTextureToMipmappedArray</span><span class="w"></span>
<span class="n">cudaConfigureCall</span><span class="w"></span>
<span class="n">cudaCreateSurfaceObject</span><span class="w"></span>
<span class="n">cudaCreateTextureObject</span><span class="w"></span>
<span class="n">cudaD3D10MapResources</span><span class="w"></span>
<span class="n">cudaD3D10RegisterResource</span><span class="w"></span>
<span class="n">cudaD3D10UnmapResources</span><span class="w"></span>
<span class="n">cudaD3D10UnregisterResource</span><span class="w"></span>
<span class="n">cudaD3D9MapResources</span><span class="w"></span>
<span class="n">cudaD3D9MapVertexBuffer</span><span class="w"></span>
<span class="n">cudaD3D9RegisterResource</span><span class="w"></span>
<span class="n">cudaD3D9RegisterVertexBuffer</span><span class="w"></span>
<span class="n">cudaD3D9UnmapResources</span><span class="w"></span>
<span class="n">cudaD3D9UnmapVertexBuffer</span><span class="w"></span>
<span class="n">cudaD3D9UnregisterResource</span><span class="w"></span>
<span class="n">cudaD3D9UnregisterVertexBuffer</span><span class="w"></span>
<span class="n">cudaDestroySurfaceObject</span><span class="w"></span>
<span class="n">cudaDestroyTextureObject</span><span class="w"></span>
<span class="n">cudaDeviceReset</span><span class="w"></span>
<span class="n">cudaDeviceSynchronize</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerAcquireFrame</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerConnect</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerConnectWithFlags</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerDisconnect</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerReleaseFrame</span><span class="w"></span>
<span class="n">cudaEGLStreamConsumerReleaseFrame</span><span class="w"></span>
<span class="n">cudaEGLStreamProducerConnect</span><span class="w"></span>
<span class="n">cudaEGLStreamProducerDisconnect</span><span class="w"></span>
<span class="n">cudaEGLStreamProducerReturnFrame</span><span class="w"></span>
<span class="n">cudaEventCreate</span><span class="w"></span>
<span class="n">cudaEventCreateFromEGLSync</span><span class="w"></span>
<span class="n">cudaEventCreateWithFlags</span><span class="w"></span>
<span class="n">cudaEventDestroy</span><span class="w"></span>
<span class="n">cudaEventQuery</span><span class="w"></span>
<span class="n">cudaEventRecord</span><span class="w"></span>
<span class="n">cudaEventRecord_ptsz</span><span class="w"></span>
<span class="n">cudaEventSynchronize</span><span class="w"></span>
<span class="n">cudaFree</span><span class="w"></span>
<span class="n">cudaFreeArray</span><span class="w"></span>
<span class="n">cudaFreeHost</span><span class="w"></span>
<span class="n">cudaFreeMipmappedArray</span><span class="w"></span>
<span class="n">cudaGLMapBufferObject</span><span class="w"></span>
<span class="n">cudaGLMapBufferObjectAsync</span><span class="w"></span>
<span class="n">cudaGLRegisterBufferObject</span><span class="w"></span>
<span class="n">cudaGLUnmapBufferObject</span><span class="w"></span>
<span class="n">cudaGLUnmapBufferObjectAsync</span><span class="w"></span>
<span class="n">cudaGLUnregisterBufferObject</span><span class="w"></span>
<span class="n">cudaGraphicsD3D10RegisterResource</span><span class="w"></span>
<span class="n">cudaGraphicsD3D11RegisterResource</span><span class="w"></span>
<span class="n">cudaGraphicsD3D9RegisterResource</span><span class="w"></span>
<span class="n">cudaGraphicsEGLRegisterImage</span><span class="w"></span>
<span class="n">cudaGraphicsGLRegisterBuffer</span><span class="w"></span>
<span class="n">cudaGraphicsGLRegisterImage</span><span class="w"></span>
<span class="n">cudaGraphicsMapResources</span><span class="w"></span>
<span class="n">cudaGraphicsUnmapResources</span><span class="w"></span>
<span class="n">cudaGraphicsUnregisterResource</span><span class="w"></span>
<span class="n">cudaGraphicsVDPAURegisterOutputSurface</span><span class="w"></span>
<span class="n">cudaGraphicsVDPAURegisterVideoSurface</span><span class="w"></span>
<span class="n">cudaHostAlloc</span><span class="w"></span>
<span class="n">cudaHostRegister</span><span class="w"></span>
<span class="n">cudaHostUnregister</span><span class="w"></span>
<span class="n">cudaLaunch</span><span class="w"></span>
<span class="n">cudaLaunchCooperativeKernel</span><span class="w"></span>
<span class="n">cudaLaunchCooperativeKernelMultiDevice</span><span class="w"></span>
<span class="n">cudaLaunchCooperativeKernel_ptsz</span><span class="w"></span>
<span class="n">cudaLaunchKernel</span><span class="w"></span>
<span class="n">cudaLaunchKernel_ptsz</span><span class="w"></span>
<span class="n">cudaLaunch_ptsz</span><span class="w"></span>
<span class="n">cudaMalloc</span><span class="w"></span>
<span class="n">cudaMalloc3D</span><span class="w"></span>
<span class="n">cudaMalloc3DArray</span><span class="w"></span>
<span class="n">cudaMallocArray</span><span class="w"></span>
<span class="n">cudaMallocHost</span><span class="w"></span>
<span class="n">cudaMallocManaged</span><span class="w"></span>
<span class="n">cudaMallocMipmappedArray</span><span class="w"></span>
<span class="n">cudaMallocPitch</span><span class="w"></span>
<span class="n">cudaMemGetInfo</span><span class="w"></span>
<span class="n">cudaMemPrefetchAsync</span><span class="w"></span>
<span class="n">cudaMemPrefetchAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy</span><span class="w"></span>
<span class="n">cudaMemcpy2D</span><span class="w"></span>
<span class="n">cudaMemcpy2DArrayToArray</span><span class="w"></span>
<span class="n">cudaMemcpy2DArrayToArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy2DAsync</span><span class="w"></span>
<span class="n">cudaMemcpy2DAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy2DFromArray</span><span class="w"></span>
<span class="n">cudaMemcpy2DFromArrayAsync</span><span class="w"></span>
<span class="n">cudaMemcpy2DFromArrayAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy2DFromArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy2DToArray</span><span class="w"></span>
<span class="n">cudaMemcpy2DToArrayAsync</span><span class="w"></span>
<span class="n">cudaMemcpy2DToArrayAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy2DToArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy2D_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy3D</span><span class="w"></span>
<span class="n">cudaMemcpy3DAsync</span><span class="w"></span>
<span class="n">cudaMemcpy3DAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy3DPeer</span><span class="w"></span>
<span class="n">cudaMemcpy3DPeerAsync</span><span class="w"></span>
<span class="n">cudaMemcpy3DPeerAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpy3DPeer_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy3D_ptds</span><span class="w"></span>
<span class="n">cudaMemcpyArrayToArray</span><span class="w"></span>
<span class="n">cudaMemcpyArrayToArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpyAsync</span><span class="w"></span>
<span class="n">cudaMemcpyAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpyFromArray</span><span class="w"></span>
<span class="n">cudaMemcpyFromArrayAsync</span><span class="w"></span>
<span class="n">cudaMemcpyFromArrayAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpyFromArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpyFromSymbol</span><span class="w"></span>
<span class="n">cudaMemcpyFromSymbolAsync</span><span class="w"></span>
<span class="n">cudaMemcpyFromSymbolAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpyFromSymbol_ptds</span><span class="w"></span>
<span class="n">cudaMemcpyPeer</span><span class="w"></span>
<span class="n">cudaMemcpyPeerAsync</span><span class="w"></span>
<span class="n">cudaMemcpyToArray</span><span class="w"></span>
<span class="n">cudaMemcpyToArrayAsync</span><span class="w"></span>
<span class="n">cudaMemcpyToArrayAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpyToArray_ptds</span><span class="w"></span>
<span class="n">cudaMemcpyToSymbol</span><span class="w"></span>
<span class="n">cudaMemcpyToSymbolAsync</span><span class="w"></span>
<span class="n">cudaMemcpyToSymbolAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemcpyToSymbol_ptds</span><span class="w"></span>
<span class="n">cudaMemcpy_ptds</span><span class="w"></span>
<span class="n">cudaMemset</span><span class="w"></span>
<span class="n">cudaMemset2D</span><span class="w"></span>
<span class="n">cudaMemset2DAsync</span><span class="w"></span>
<span class="n">cudaMemset2DAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemset2D_ptds</span><span class="w"></span>
<span class="n">cudaMemset3D</span><span class="w"></span>
<span class="n">cudaMemset3DAsync</span><span class="w"></span>
<span class="n">cudaMemset3DAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemset3D_ptds</span><span class="w"></span>
<span class="n">cudaMemsetAsync</span><span class="w"></span>
<span class="n">cudaMemsetAsync_ptsz</span><span class="w"></span>
<span class="n">cudaMemset_ptds</span><span class="w"></span>
<span class="n">cudaPeerRegister</span><span class="w"></span>
<span class="n">cudaPeerUnregister</span><span class="w"></span>
<span class="n">cudaStreamAddCallback</span><span class="w"></span>
<span class="n">cudaStreamAddCallback_ptsz</span><span class="w"></span>
<span class="n">cudaStreamAttachMemAsync</span><span class="w"></span>
<span class="n">cudaStreamAttachMemAsync_ptsz</span><span class="w"></span>
<span class="n">cudaStreamCreate</span><span class="w"></span>
<span class="n">cudaStreamCreateWithFlags</span><span class="w"></span>
<span class="n">cudaStreamCreateWithPriority</span><span class="w"></span>
<span class="n">cudaStreamDestroy</span><span class="w"></span>
<span class="n">cudaStreamQuery</span><span class="w"></span>
<span class="n">cudaStreamQuery_ptsz</span><span class="w"></span>
<span class="n">cudaStreamSynchronize</span><span class="w"></span>
<span class="n">cudaStreamSynchronize_ptsz</span><span class="w"></span>
<span class="n">cudaStreamWaitEvent</span><span class="w"></span>
<span class="n">cudaStreamWaitEvent_ptsz</span><span class="w"></span>
<span class="n">cudaThreadSynchronize</span><span class="w"></span>
<span class="n">cudaUnbindTexture</span><span class="w"></span>
</pre></div>
</div>
<p><strong>CUDA Primary API</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">cu64Array3DCreate</span><span class="w"></span>
<span class="n">cu64ArrayCreate</span><span class="w"></span>
<span class="n">cu64D3D9MapVertexBuffer</span><span class="w"></span>
<span class="n">cu64GLMapBufferObject</span><span class="w"></span>
<span class="n">cu64GLMapBufferObjectAsync</span><span class="w"></span>
<span class="n">cu64MemAlloc</span><span class="w"></span>
<span class="n">cu64MemAllocPitch</span><span class="w"></span>
<span class="n">cu64MemFree</span><span class="w"></span>
<span class="n">cu64MemGetInfo</span><span class="w"></span>
<span class="n">cu64MemHostAlloc</span><span class="w"></span>
<span class="n">cu64Memcpy2D</span><span class="w"></span>
<span class="n">cu64Memcpy2DAsync</span><span class="w"></span>
<span class="n">cu64Memcpy2DUnaligned</span><span class="w"></span>
<span class="n">cu64Memcpy3D</span><span class="w"></span>
<span class="n">cu64Memcpy3DAsync</span><span class="w"></span>
<span class="n">cu64MemcpyAtoD</span><span class="w"></span>
<span class="n">cu64MemcpyDtoA</span><span class="w"></span>
<span class="n">cu64MemcpyDtoD</span><span class="w"></span>
<span class="n">cu64MemcpyDtoDAsync</span><span class="w"></span>
<span class="n">cu64MemcpyDtoH</span><span class="w"></span>
<span class="n">cu64MemcpyDtoHAsync</span><span class="w"></span>
<span class="n">cu64MemcpyHtoD</span><span class="w"></span>
<span class="n">cu64MemcpyHtoDAsync</span><span class="w"></span>
<span class="n">cu64MemsetD16</span><span class="w"></span>
<span class="n">cu64MemsetD16Async</span><span class="w"></span>
<span class="n">cu64MemsetD2D16</span><span class="w"></span>
<span class="n">cu64MemsetD2D16Async</span><span class="w"></span>
<span class="n">cu64MemsetD2D32</span><span class="w"></span>
<span class="n">cu64MemsetD2D32Async</span><span class="w"></span>
<span class="n">cu64MemsetD2D8</span><span class="w"></span>
<span class="n">cu64MemsetD2D8Async</span><span class="w"></span>
<span class="n">cu64MemsetD32</span><span class="w"></span>
<span class="n">cu64MemsetD32Async</span><span class="w"></span>
<span class="n">cu64MemsetD8</span><span class="w"></span>
<span class="n">cu64MemsetD8Async</span><span class="w"></span>
<span class="n">cuArray3DCreate</span><span class="w"></span>
<span class="n">cuArray3DCreate_v2</span><span class="w"></span>
<span class="n">cuArrayCreate</span><span class="w"></span>
<span class="n">cuArrayCreate_v2</span><span class="w"></span>
<span class="n">cuArrayDestroy</span><span class="w"></span>
<span class="n">cuBinaryFree</span><span class="w"></span>
<span class="n">cuCompilePtx</span><span class="w"></span>
<span class="n">cuCtxCreate</span><span class="w"></span>
<span class="n">cuCtxCreate_v2</span><span class="w"></span>
<span class="n">cuCtxDestroy</span><span class="w"></span>
<span class="n">cuCtxDestroy_v2</span><span class="w"></span>
<span class="n">cuCtxSynchronize</span><span class="w"></span>
<span class="n">cuD3D10CtxCreate</span><span class="w"></span>
<span class="n">cuD3D10CtxCreateOnDevice</span><span class="w"></span>
<span class="n">cuD3D10CtxCreate_v2</span><span class="w"></span>
<span class="n">cuD3D10MapResources</span><span class="w"></span>
<span class="n">cuD3D10RegisterResource</span><span class="w"></span>
<span class="n">cuD3D10UnmapResources</span><span class="w"></span>
<span class="n">cuD3D10UnregisterResource</span><span class="w"></span>
<span class="n">cuD3D11CtxCreate</span><span class="w"></span>
<span class="n">cuD3D11CtxCreateOnDevice</span><span class="w"></span>
<span class="n">cuD3D11CtxCreate_v2</span><span class="w"></span>
<span class="n">cuD3D9CtxCreate</span><span class="w"></span>
<span class="n">cuD3D9CtxCreateOnDevice</span><span class="w"></span>
<span class="n">cuD3D9CtxCreate_v2</span><span class="w"></span>
<span class="n">cuD3D9MapResources</span><span class="w"></span>
<span class="n">cuD3D9MapVertexBuffer</span><span class="w"></span>
<span class="n">cuD3D9MapVertexBuffer_v2</span><span class="w"></span>
<span class="n">cuD3D9RegisterResource</span><span class="w"></span>
<span class="n">cuD3D9RegisterVertexBuffer</span><span class="w"></span>
<span class="n">cuD3D9UnmapResources</span><span class="w"></span>
<span class="n">cuD3D9UnmapVertexBuffer</span><span class="w"></span>
<span class="n">cuD3D9UnregisterResource</span><span class="w"></span>
<span class="n">cuD3D9UnregisterVertexBuffer</span><span class="w"></span>
<span class="n">cuEGLStreamConsumerAcquireFrame</span><span class="w"></span>
<span class="n">cuEGLStreamConsumerConnect</span><span class="w"></span>
<span class="n">cuEGLStreamConsumerConnectWithFlags</span><span class="w"></span>
<span class="n">cuEGLStreamConsumerDisconnect</span><span class="w"></span>
<span class="n">cuEGLStreamConsumerReleaseFrame</span><span class="w"></span>
<span class="n">cuEGLStreamProducerConnect</span><span class="w"></span>
<span class="n">cuEGLStreamProducerDisconnect</span><span class="w"></span>
<span class="n">cuEGLStreamProducerPresentFrame</span><span class="w"></span>
<span class="n">cuEGLStreamProducerReturnFrame</span><span class="w"></span>
<span class="n">cuEventCreate</span><span class="w"></span>
<span class="n">cuEventCreateFromEGLSync</span><span class="w"></span>
<span class="n">cuEventCreateFromNVNSync</span><span class="w"></span>
<span class="n">cuEventDestroy</span><span class="w"></span>
<span class="n">cuEventDestroy_v2</span><span class="w"></span>
<span class="n">cuEventQuery</span><span class="w"></span>
<span class="n">cuEventRecord</span><span class="w"></span>
<span class="n">cuEventRecord_ptsz</span><span class="w"></span>
<span class="n">cuEventSynchronize</span><span class="w"></span>
<span class="n">cuGLCtxCreate</span><span class="w"></span>
<span class="n">cuGLCtxCreate_v2</span><span class="w"></span>
<span class="n">cuGLInit</span><span class="w"></span>
<span class="n">cuGLMapBufferObject</span><span class="w"></span>
<span class="n">cuGLMapBufferObjectAsync</span><span class="w"></span>
<span class="n">cuGLMapBufferObjectAsync_v2</span><span class="w"></span>
<span class="n">cuGLMapBufferObjectAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuGLMapBufferObject_v2</span><span class="w"></span>
<span class="n">cuGLMapBufferObject_v2_ptds</span><span class="w"></span>
<span class="n">cuGLRegisterBufferObject</span><span class="w"></span>
<span class="n">cuGLUnmapBufferObject</span><span class="w"></span>
<span class="n">cuGLUnmapBufferObjectAsync</span><span class="w"></span>
<span class="n">cuGLUnregisterBufferObject</span><span class="w"></span>
<span class="n">cuGraphicsD3D10RegisterResource</span><span class="w"></span>
<span class="n">cuGraphicsD3D11RegisterResource</span><span class="w"></span>
<span class="n">cuGraphicsD3D9RegisterResource</span><span class="w"></span>
<span class="n">cuGraphicsEGLRegisterImage</span><span class="w"></span>
<span class="n">cuGraphicsGLRegisterBuffer</span><span class="w"></span>
<span class="n">cuGraphicsGLRegisterImage</span><span class="w"></span>
<span class="n">cuGraphicsMapResources</span><span class="w"></span>
<span class="n">cuGraphicsMapResources_ptsz</span><span class="w"></span>
<span class="n">cuGraphicsUnmapResources</span><span class="w"></span>
<span class="n">cuGraphicsUnmapResources_ptsz</span><span class="w"></span>
<span class="n">cuGraphicsUnregisterResource</span><span class="w"></span>
<span class="n">cuGraphicsVDPAURegisterOutputSurface</span><span class="w"></span>
<span class="n">cuGraphicsVDPAURegisterVideoSurface</span><span class="w"></span>
<span class="n">cuInit</span><span class="w"></span>
<span class="n">cuLaunch</span><span class="w"></span>
<span class="n">cuLaunchCooperativeKernel</span><span class="w"></span>
<span class="n">cuLaunchCooperativeKernelMultiDevice</span><span class="w"></span>
<span class="n">cuLaunchCooperativeKernel_ptsz</span><span class="w"></span>
<span class="n">cuLaunchGrid</span><span class="w"></span>
<span class="n">cuLaunchGridAsync</span><span class="w"></span>
<span class="n">cuLaunchKernel</span><span class="w"></span>
<span class="n">cuLaunchKernel_ptsz</span><span class="w"></span>
<span class="n">cuLinkComplete</span><span class="w"></span>
<span class="n">cuLinkCreate</span><span class="w"></span>
<span class="n">cuLinkCreate_v2</span><span class="w"></span>
<span class="n">cuLinkDestroy</span><span class="w"></span>
<span class="n">cuMemAlloc</span><span class="w"></span>
<span class="n">cuMemAllocHost</span><span class="w"></span>
<span class="n">cuMemAllocHost_v2</span><span class="w"></span>
<span class="n">cuMemAllocManaged</span><span class="w"></span>
<span class="n">cuMemAllocPitch</span><span class="w"></span>
<span class="n">cuMemAllocPitch_v2</span><span class="w"></span>
<span class="n">cuMemAlloc_v2</span><span class="w"></span>
<span class="n">cuMemFree</span><span class="w"></span>
<span class="n">cuMemFreeHost</span><span class="w"></span>
<span class="n">cuMemFree_v2</span><span class="w"></span>
<span class="n">cuMemGetInfo</span><span class="w"></span>
<span class="n">cuMemGetInfo_v2</span><span class="w"></span>
<span class="n">cuMemHostAlloc</span><span class="w"></span>
<span class="n">cuMemHostAlloc_v2</span><span class="w"></span>
<span class="n">cuMemHostRegister</span><span class="w"></span>
<span class="n">cuMemHostRegister_v2</span><span class="w"></span>
<span class="n">cuMemHostUnregister</span><span class="w"></span>
<span class="n">cuMemPeerRegister</span><span class="w"></span>
<span class="n">cuMemPeerUnregister</span><span class="w"></span>
<span class="n">cuMemPrefetchAsync</span><span class="w"></span>
<span class="n">cuMemPrefetchAsync_ptsz</span><span class="w"></span>
<span class="n">cuMemcpy</span><span class="w"></span>
<span class="n">cuMemcpy2D</span><span class="w"></span>
<span class="n">cuMemcpy2DAsync</span><span class="w"></span>
<span class="n">cuMemcpy2DAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpy2DAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpy2DUnaligned</span><span class="w"></span>
<span class="n">cuMemcpy2DUnaligned_v2</span><span class="w"></span>
<span class="n">cuMemcpy2DUnaligned_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpy2D_v2</span><span class="w"></span>
<span class="n">cuMemcpy2D_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpy3D</span><span class="w"></span>
<span class="n">cuMemcpy3DAsync</span><span class="w"></span>
<span class="n">cuMemcpy3DAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpy3DAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpy3DPeer</span><span class="w"></span>
<span class="n">cuMemcpy3DPeerAsync</span><span class="w"></span>
<span class="n">cuMemcpy3DPeerAsync_ptsz</span><span class="w"></span>
<span class="n">cuMemcpy3DPeer_ptds</span><span class="w"></span>
<span class="n">cuMemcpy3D_v2</span><span class="w"></span>
<span class="n">cuMemcpy3D_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyAsync</span><span class="w"></span>
<span class="n">cuMemcpyAsync_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyAtoA</span><span class="w"></span>
<span class="n">cuMemcpyAtoA_v2</span><span class="w"></span>
<span class="n">cuMemcpyAtoA_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyAtoD</span><span class="w"></span>
<span class="n">cuMemcpyAtoD_v2</span><span class="w"></span>
<span class="n">cuMemcpyAtoD_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyAtoH</span><span class="w"></span>
<span class="n">cuMemcpyAtoHAsync</span><span class="w"></span>
<span class="n">cuMemcpyAtoHAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpyAtoHAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyAtoH_v2</span><span class="w"></span>
<span class="n">cuMemcpyAtoH_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyDtoA</span><span class="w"></span>
<span class="n">cuMemcpyDtoA_v2</span><span class="w"></span>
<span class="n">cuMemcpyDtoA_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyDtoD</span><span class="w"></span>
<span class="n">cuMemcpyDtoDAsync</span><span class="w"></span>
<span class="n">cuMemcpyDtoDAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpyDtoDAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyDtoD_v2</span><span class="w"></span>
<span class="n">cuMemcpyDtoD_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyDtoH</span><span class="w"></span>
<span class="n">cuMemcpyDtoHAsync</span><span class="w"></span>
<span class="n">cuMemcpyDtoHAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpyDtoHAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyDtoH_v2</span><span class="w"></span>
<span class="n">cuMemcpyDtoH_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyHtoA</span><span class="w"></span>
<span class="n">cuMemcpyHtoAAsync</span><span class="w"></span>
<span class="n">cuMemcpyHtoAAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpyHtoAAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyHtoA_v2</span><span class="w"></span>
<span class="n">cuMemcpyHtoA_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyHtoD</span><span class="w"></span>
<span class="n">cuMemcpyHtoDAsync</span><span class="w"></span>
<span class="n">cuMemcpyHtoDAsync_v2</span><span class="w"></span>
<span class="n">cuMemcpyHtoDAsync_v2_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyHtoD_v2</span><span class="w"></span>
<span class="n">cuMemcpyHtoD_v2_ptds</span><span class="w"></span>
<span class="n">cuMemcpyPeer</span><span class="w"></span>
<span class="n">cuMemcpyPeerAsync</span><span class="w"></span>
<span class="n">cuMemcpyPeerAsync_ptsz</span><span class="w"></span>
<span class="n">cuMemcpyPeer_ptds</span><span class="w"></span>
<span class="n">cuMemcpy_ptds</span><span class="w"></span>
<span class="n">cuMemcpy_v2</span><span class="w"></span>
<span class="n">cuMemsetD16</span><span class="w"></span>
<span class="n">cuMemsetD16Async</span><span class="w"></span>
<span class="n">cuMemsetD16Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD16_v2</span><span class="w"></span>
<span class="n">cuMemsetD16_v2_ptds</span><span class="w"></span>
<span class="n">cuMemsetD2D16</span><span class="w"></span>
<span class="n">cuMemsetD2D16Async</span><span class="w"></span>
<span class="n">cuMemsetD2D16Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD2D16_v2</span><span class="w"></span>
<span class="n">cuMemsetD2D16_v2_ptds</span><span class="w"></span>
<span class="n">cuMemsetD2D32</span><span class="w"></span>
<span class="n">cuMemsetD2D32Async</span><span class="w"></span>
<span class="n">cuMemsetD2D32Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD2D32_v2</span><span class="w"></span>
<span class="n">cuMemsetD2D32_v2_ptds</span><span class="w"></span>
<span class="n">cuMemsetD2D8</span><span class="w"></span>
<span class="n">cuMemsetD2D8Async</span><span class="w"></span>
<span class="n">cuMemsetD2D8Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD2D8_v2</span><span class="w"></span>
<span class="n">cuMemsetD2D8_v2_ptds</span><span class="w"></span>
<span class="n">cuMemsetD32</span><span class="w"></span>
<span class="n">cuMemsetD32Async</span><span class="w"></span>
<span class="n">cuMemsetD32Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD32_v2</span><span class="w"></span>
<span class="n">cuMemsetD32_v2_ptds</span><span class="w"></span>
<span class="n">cuMemsetD8</span><span class="w"></span>
<span class="n">cuMemsetD8Async</span><span class="w"></span>
<span class="n">cuMemsetD8Async_ptsz</span><span class="w"></span>
<span class="n">cuMemsetD8_v2</span><span class="w"></span>
<span class="n">cuMemsetD8_v2_ptds</span><span class="w"></span>
<span class="n">cuMipmappedArrayCreate</span><span class="w"></span>
<span class="n">cuMipmappedArrayDestroy</span><span class="w"></span>
<span class="n">cuModuleLoad</span><span class="w"></span>
<span class="n">cuModuleLoadData</span><span class="w"></span>
<span class="n">cuModuleLoadDataEx</span><span class="w"></span>
<span class="n">cuModuleLoadFatBinary</span><span class="w"></span>
<span class="n">cuModuleUnload</span><span class="w"></span>
<span class="n">cuStreamAddCallback</span><span class="w"></span>
<span class="n">cuStreamAddCallback_ptsz</span><span class="w"></span>
<span class="n">cuStreamAttachMemAsync</span><span class="w"></span>
<span class="n">cuStreamAttachMemAsync_ptsz</span><span class="w"></span>
<span class="n">cuStreamBatchMemOp</span><span class="w"></span>
<span class="n">cuStreamBatchMemOp_ptsz</span><span class="w"></span>
<span class="n">cuStreamCreate</span><span class="w"></span>
<span class="n">cuStreamCreateWithPriority</span><span class="w"></span>
<span class="n">cuStreamDestroy</span><span class="w"></span>
<span class="n">cuStreamDestroy_v2</span><span class="w"></span>
<span class="n">cuStreamSynchronize</span><span class="w"></span>
<span class="n">cuStreamSynchronize_ptsz</span><span class="w"></span>
<span class="n">cuStreamWaitEvent</span><span class="w"></span>
<span class="n">cuStreamWaitEvent_ptsz</span><span class="w"></span>
<span class="n">cuStreamWaitValue32</span><span class="w"></span>
<span class="n">cuStreamWaitValue32_ptsz</span><span class="w"></span>
<span class="n">cuStreamWaitValue64</span><span class="w"></span>
<span class="n">cuStreamWaitValue64_ptsz</span><span class="w"></span>
<span class="n">cuStreamWriteValue32</span><span class="w"></span>
<span class="n">cuStreamWriteValue32_ptsz</span><span class="w"></span>
<span class="n">cuStreamWriteValue64</span><span class="w"></span>
<span class="n">cuStreamWriteValue64_ptsz</span><span class="w"></span>
<span class="n">cuSurfObjectCreate</span><span class="w"></span>
<span class="n">cuSurfObjectDestroy</span><span class="w"></span>
<span class="n">cuSurfRefCreate</span><span class="w"></span>
<span class="n">cuSurfRefDestroy</span><span class="w"></span>
<span class="n">cuTexObjectCreate</span><span class="w"></span>
<span class="n">cuTexObjectDestroy</span><span class="w"></span>
<span class="n">cuTexRefCreate</span><span class="w"></span>
<span class="n">cuTexRefDestroy</span><span class="w"></span>
<span class="n">cuVDPAUCtxCreate</span><span class="w"></span>
<span class="n">cuVDPAUCtxCreate_v2</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="cudnn-function-list-for-x86-cli">
<h3>cuDNN Function List for X86 CLI<a class="headerlink" href="#cudnn-function-list-for-x86-cli" title="Permalink to this heading"></a></h3>
<p><strong>cuDNN API functions</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">cudnnActivationBackward</span><span class="w"></span>
<span class="n">cudnnActivationBackward_v3</span><span class="w"></span>
<span class="n">cudnnActivationBackward_v4</span><span class="w"></span>
<span class="n">cudnnActivationForward</span><span class="w"></span>
<span class="n">cudnnActivationForward_v3</span><span class="w"></span>
<span class="n">cudnnActivationForward_v4</span><span class="w"></span>
<span class="n">cudnnAddTensor</span><span class="w"></span>
<span class="n">cudnnBatchNormalizationBackward</span><span class="w"></span>
<span class="n">cudnnBatchNormalizationBackwardEx</span><span class="w"></span>
<span class="n">cudnnBatchNormalizationForwardInference</span><span class="w"></span>
<span class="n">cudnnBatchNormalizationForwardTraining</span><span class="w"></span>
<span class="n">cudnnBatchNormalizationForwardTrainingEx</span><span class="w"></span>
<span class="n">cudnnCTCLoss</span><span class="w"></span>
<span class="n">cudnnConvolutionBackwardBias</span><span class="w"></span>
<span class="n">cudnnConvolutionBackwardData</span><span class="w"></span>
<span class="n">cudnnConvolutionBackwardFilter</span><span class="w"></span>
<span class="n">cudnnConvolutionBiasActivationForward</span><span class="w"></span>
<span class="n">cudnnConvolutionForward</span><span class="w"></span>
<span class="n">cudnnCreate</span><span class="w"></span>
<span class="n">cudnnCreateAlgorithmPerformance</span><span class="w"></span>
<span class="n">cudnnDestroy</span><span class="w"></span>
<span class="n">cudnnDestroyAlgorithmPerformance</span><span class="w"></span>
<span class="n">cudnnDestroyPersistentRNNPlan</span><span class="w"></span>
<span class="n">cudnnDivisiveNormalizationBackward</span><span class="w"></span>
<span class="n">cudnnDivisiveNormalizationForward</span><span class="w"></span>
<span class="n">cudnnDropoutBackward</span><span class="w"></span>
<span class="n">cudnnDropoutForward</span><span class="w"></span>
<span class="n">cudnnDropoutGetReserveSpaceSize</span><span class="w"></span>
<span class="n">cudnnDropoutGetStatesSize</span><span class="w"></span>
<span class="n">cudnnFindConvolutionBackwardDataAlgorithm</span><span class="w"></span>
<span class="n">cudnnFindConvolutionBackwardDataAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindConvolutionBackwardFilterAlgorithm</span><span class="w"></span>
<span class="n">cudnnFindConvolutionBackwardFilterAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindConvolutionForwardAlgorithm</span><span class="w"></span>
<span class="n">cudnnFindConvolutionForwardAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindRNNBackwardDataAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindRNNBackwardWeightsAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindRNNForwardInferenceAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFindRNNForwardTrainingAlgorithmEx</span><span class="w"></span>
<span class="n">cudnnFusedOpsExecute</span><span class="w"></span>
<span class="n">cudnnIm2Col</span><span class="w"></span>
<span class="n">cudnnLRNCrossChannelBackward</span><span class="w"></span>
<span class="n">cudnnLRNCrossChannelForward</span><span class="w"></span>
<span class="n">cudnnMakeFusedOpsPlan</span><span class="w"></span>
<span class="n">cudnnMultiHeadAttnBackwardData</span><span class="w"></span>
<span class="n">cudnnMultiHeadAttnBackwardWeights</span><span class="w"></span>
<span class="n">cudnnMultiHeadAttnForward</span><span class="w"></span>
<span class="n">cudnnOpTensor</span><span class="w"></span>
<span class="n">cudnnPoolingBackward</span><span class="w"></span>
<span class="n">cudnnPoolingForward</span><span class="w"></span>
<span class="n">cudnnRNNBackwardData</span><span class="w"></span>
<span class="n">cudnnRNNBackwardDataEx</span><span class="w"></span>
<span class="n">cudnnRNNBackwardWeights</span><span class="w"></span>
<span class="n">cudnnRNNBackwardWeightsEx</span><span class="w"></span>
<span class="n">cudnnRNNForwardInference</span><span class="w"></span>
<span class="n">cudnnRNNForwardInferenceEx</span><span class="w"></span>
<span class="n">cudnnRNNForwardTraining</span><span class="w"></span>
<span class="n">cudnnRNNForwardTrainingEx</span><span class="w"></span>
<span class="n">cudnnReduceTensor</span><span class="w"></span>
<span class="n">cudnnReorderFilterAndBias</span><span class="w"></span>
<span class="n">cudnnRestoreAlgorithm</span><span class="w"></span>
<span class="n">cudnnRestoreDropoutDescriptor</span><span class="w"></span>
<span class="n">cudnnSaveAlgorithm</span><span class="w"></span>
<span class="n">cudnnScaleTensor</span><span class="w"></span>
<span class="n">cudnnSoftmaxBackward</span><span class="w"></span>
<span class="n">cudnnSoftmaxForward</span><span class="w"></span>
<span class="n">cudnnSpatialTfGridGeneratorBackward</span><span class="w"></span>
<span class="n">cudnnSpatialTfGridGeneratorForward</span><span class="w"></span>
<span class="n">cudnnSpatialTfSamplerBackward</span><span class="w"></span>
<span class="n">cudnnSpatialTfSamplerForward</span><span class="w"></span>
<span class="n">cudnnTransformFilter</span><span class="w"></span>
<span class="n">cudnnTransformTensor</span><span class="w"></span>
<span class="n">cudnnTransformTensorEx</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
<section id="openacc-trace">
<h2>OpenACC Trace<a class="headerlink" href="#openacc-trace" title="Permalink to this heading"></a></h2>
<p>Nsight Systems for Linux x86_64 and Power targets is capable of capturing information about OpenACC execution in the profiled process.</p>
<p>OpenACC versions 2.0, 2.5, and 2.6 are supported when using PGI runtime version 15.7 or later. In order to differentiate constructs (see tooltip below), a PGI runtime of 16.0 or later is required. Note that Nsight Systems does not support the GCC implementation of OpenACC at this time.</p>
<p>Under the CPU rows in the timeline tree, each thread that uses OpenACC will show OpenACC trace information. You can click on a OpenACC API call to see correlation with the underlying CUDA API calls (highlighted in teal):</p>
<img alt="OpenACC rows" class="image" src="../_images/openacc-CUDA-correlation.png" />
<p>If the OpenACC API results in GPU work, that will also be highlighted:</p>
<img alt="OpenACC rows" class="image" src="../_images/openacc-gpu-correlation.png" />
<p>Hovering over a particular OpenACC construct will bring up a tooltip with details about that construct:</p>
<blockquote>
<div><img alt="OpenACC construct tooltip" class="image" src="../_images/openacc-tooltip.png" />
</div></blockquote>
<p>To capture OpenACC information from the Nsight Systems GUI, select the <strong>Collect OpenACC trace</strong> checkbox under <strong>Collect CUDA trace</strong> configurations. Note that turning on OpenACC tracing will also turn on CUDA tracing.</p>
<blockquote>
<div><img alt="Configure CUDA trace" class="image" src="../_images/configure-cuda-trace-x86.png" />
</div></blockquote>
<p>Please note that if your application crashes before all collected OpenACC trace data has been copied out, some or all data might be lost and not present in the report.</p>
</section>
<section id="opengl-trace">
<h2>OpenGL Trace<a class="headerlink" href="#opengl-trace" title="Permalink to this heading"></a></h2>
<p>OpenGL and OpenGL ES APIs can be traced to assist in the analysis of CPU and GPU interactions.</p>
<p>A few usage examples are:</p>
<ol class="arabic simple">
<li><p>Visualize how long <code class="docutils literal notranslate"><span class="pre">eglSwapBuffers</span></code> (or similar) is taking.</p></li>
<li><p>API trace can easily show correlations between thread state and graphics driver’s behavior, uncovering where the CPU may be waiting on the GPU.</p></li>
<li><p>Spot bubbles of opportunity on the GPU, where more GPU workload could be created.</p></li>
<li><p>Use <code class="docutils literal notranslate"><span class="pre">KHR_debug</span></code> extension to trace GL events on both the CPU and GPU.</p></li>
</ol>
<p>OpenGL trace feature in Nsight Systems consists of two different activities which will be shown in the CPU rows for those threads</p>
<ul class="simple">
<li><p><strong>CPU trace</strong>: interception of API calls that an application does to APIs (such as OpenGL, OpenGL ES, EGL, GLX, WGL, etc.).</p></li>
<li><p><strong>GPU trace</strong> (or <strong>workload trace</strong>): trace of GPU workload (activity) triggered by use of OpenGL or OpenGL ES. Since draw calls are executed back-to-back, the GPU workload trace ranges include many OpenGL draw calls and operations in order to optimize performance overhead, rather than tracing each individual operation.</p></li>
</ul>
<p>To collect GPU trace, the <code class="docutils literal notranslate"><span class="pre">glQueryCounter()</span></code> function is used to measure how much time batches of GPU workload take to complete.</p>
<blockquote>
<div><img alt="Configure OpenGL trace" class="image" src="../_images/configure-opengl-trace.png" />
</div></blockquote>
<blockquote>
<div><img alt="Configure OpenGL functions" class="image" src="../_images/configure-opengl-functions.png" />
</div></blockquote>
<p>Ranges defined by the <code class="docutils literal notranslate"><span class="pre">KHR_debug</span></code> calls are represented similarly to OpenGL API and OpenGL GPU workload trace. GPU ranges in this case represent <em>incremental draw cost</em>. They cannot fully account for GPUs that can execute multiple draw calls in parallel. In this case, Nsight Systems will not show overlapping GPU ranges.</p>
<section id="opengl-trace-using-command-line">
<h3>OpenGL Trace Using Command Line<a class="headerlink" href="#opengl-trace-using-command-line" title="Permalink to this heading"></a></h3>
<p>For general information on using the target CLI, see <a class="reference external" href="topics/cli-profiling">CLI Profiling on Linux</a>. For the CLI, the functions that are traced are set to the following list:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">glWaitSync</span><span class="w"></span>
<span class="n">glReadPixels</span><span class="w"></span>
<span class="n">glReadnPixelsKHR</span><span class="w"></span>
<span class="n">glReadnPixelsEXT</span><span class="w"></span>
<span class="n">glReadnPixelsARB</span><span class="w"></span>
<span class="n">glReadnPixels</span><span class="w"></span>
<span class="n">glFlush</span><span class="w"></span>
<span class="n">glFinishFenceNV</span><span class="w"></span>
<span class="n">glFinish</span><span class="w"></span>
<span class="n">glClientWaitSync</span><span class="w"></span>
<span class="n">glClearTexSubImage</span><span class="w"></span>
<span class="n">glClearTexImage</span><span class="w"></span>
<span class="n">glClearStencil</span><span class="w"></span>
<span class="n">glClearNamedFramebufferuiv</span><span class="w"></span>
<span class="n">glClearNamedFramebufferiv</span><span class="w"></span>
<span class="n">glClearNamedFramebufferfv</span><span class="w"></span>
<span class="n">glClearNamedFramebufferfi</span><span class="w"></span>
<span class="n">glClearNamedBufferSubDataEXT</span><span class="w"></span>
<span class="n">glClearNamedBufferSubData</span><span class="w"></span>
<span class="n">glClearNamedBufferDataEXT</span><span class="w"></span>
<span class="n">glClearNamedBufferData</span><span class="w"></span>
<span class="n">glClearIndex</span><span class="w"></span>
<span class="n">glClearDepthx</span><span class="w"></span>
<span class="n">glClearDepthf</span><span class="w"></span>
<span class="n">glClearDepthdNV</span><span class="w"></span>
<span class="n">glClearDepth</span><span class="w"></span>
<span class="n">glClearColorx</span><span class="w"></span>
<span class="n">glClearColorIuiEXT</span><span class="w"></span>
<span class="n">glClearColorIiEXT</span><span class="w"></span>
<span class="n">glClearColor</span><span class="w"></span>
<span class="n">glClearBufferuiv</span><span class="w"></span>
<span class="n">glClearBufferSubData</span><span class="w"></span>
<span class="n">glClearBufferiv</span><span class="w"></span>
<span class="n">glClearBufferfv</span><span class="w"></span>
<span class="n">glClearBufferfi</span><span class="w"></span>
<span class="n">glClearBufferData</span><span class="w"></span>
<span class="n">glClearAccum</span><span class="w"></span>
<span class="n">glClear</span><span class="w"></span>
<span class="n">glDispatchComputeIndirect</span><span class="w"></span>
<span class="n">glDispatchComputeGroupSizeARB</span><span class="w"></span>
<span class="n">glDispatchCompute</span><span class="w"></span>
<span class="n">glComputeStreamNV</span><span class="w"></span>
<span class="n">glNamedFramebufferDrawBuffers</span><span class="w"></span>
<span class="n">glNamedFramebufferDrawBuffer</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirectEXT</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirectCountARB</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirectBindlessNV</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirectBindlessCountNV</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirectAMD</span><span class="w"></span>
<span class="n">glMultiDrawElementsIndirect</span><span class="w"></span>
<span class="n">glMultiDrawElementsEXT</span><span class="w"></span>
<span class="n">glMultiDrawElementsBaseVertex</span><span class="w"></span>
<span class="n">glMultiDrawElements</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirectEXT</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirectCountARB</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirectBindlessNV</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirectBindlessCountNV</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirectAMD</span><span class="w"></span>
<span class="n">glMultiDrawArraysIndirect</span><span class="w"></span>
<span class="n">glMultiDrawArraysEXT</span><span class="w"></span>
<span class="n">glMultiDrawArrays</span><span class="w"></span>
<span class="n">glListDrawCommandsStatesClientNV</span><span class="w"></span>
<span class="n">glFramebufferDrawBuffersEXT</span><span class="w"></span>
<span class="n">glFramebufferDrawBufferEXT</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackStreamInstanced</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackStream</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackNV</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackInstancedEXT</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackInstanced</span><span class="w"></span>
<span class="n">glDrawTransformFeedbackEXT</span><span class="w"></span>
<span class="n">glDrawTransformFeedback</span><span class="w"></span>
<span class="n">glDrawTexxvOES</span><span class="w"></span>
<span class="n">glDrawTexxOES</span><span class="w"></span>
<span class="n">glDrawTextureNV</span><span class="w"></span>
<span class="n">glDrawTexsvOES</span><span class="w"></span>
<span class="n">glDrawTexsOES</span><span class="w"></span>
<span class="n">glDrawTexivOES</span><span class="w"></span>
<span class="n">glDrawTexiOES</span><span class="w"></span>
<span class="n">glDrawTexfvOES</span><span class="w"></span>
<span class="n">glDrawTexfOES</span><span class="w"></span>
<span class="n">glDrawRangeElementsEXT</span><span class="w"></span>
<span class="n">glDrawRangeElementsBaseVertexOES</span><span class="w"></span>
<span class="n">glDrawRangeElementsBaseVertexEXT</span><span class="w"></span>
<span class="n">glDrawRangeElementsBaseVertex</span><span class="w"></span>
<span class="n">glDrawRangeElements</span><span class="w"></span>
<span class="n">glDrawPixels</span><span class="w"></span>
<span class="n">glDrawElementsInstancedNV</span><span class="w"></span>
<span class="n">glDrawElementsInstancedEXT</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseVertexOES</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseVertexEXT</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseVertexBaseInstanceEXT</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseVertexBaseInstance</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseVertex</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseInstanceEXT</span><span class="w"></span>
<span class="n">glDrawElementsInstancedBaseInstance</span><span class="w"></span>
<span class="n">glDrawElementsInstancedARB</span><span class="w"></span>
<span class="n">glDrawElementsInstanced</span><span class="w"></span>
<span class="n">glDrawElementsIndirect</span><span class="w"></span>
<span class="n">glDrawElementsBaseVertexOES</span><span class="w"></span>
<span class="n">glDrawElementsBaseVertexEXT</span><span class="w"></span>
<span class="n">glDrawElementsBaseVertex</span><span class="w"></span>
<span class="n">glDrawElements</span><span class="w"></span>
<span class="n">glDrawCommandsStatesNV</span><span class="w"></span>
<span class="n">glDrawCommandsStatesAddressNV</span><span class="w"></span>
<span class="n">glDrawCommandsNV</span><span class="w"></span>
<span class="n">glDrawCommandsAddressNV</span><span class="w"></span>
<span class="n">glDrawBuffersNV</span><span class="w"></span>
<span class="n">glDrawBuffersATI</span><span class="w"></span>
<span class="n">glDrawBuffersARB</span><span class="w"></span>
<span class="n">glDrawBuffers</span><span class="w"></span>
<span class="n">glDrawBuffer</span><span class="w"></span>
<span class="n">glDrawArraysInstancedNV</span><span class="w"></span>
<span class="n">glDrawArraysInstancedEXT</span><span class="w"></span>
<span class="n">glDrawArraysInstancedBaseInstanceEXT</span><span class="w"></span>
<span class="n">glDrawArraysInstancedBaseInstance</span><span class="w"></span>
<span class="n">glDrawArraysInstancedARB</span><span class="w"></span>
<span class="n">glDrawArraysInstanced</span><span class="w"></span>
<span class="n">glDrawArraysIndirect</span><span class="w"></span>
<span class="n">glDrawArraysEXT</span><span class="w"></span>
<span class="n">glDrawArrays</span><span class="w"></span>
<span class="n">eglSwapBuffersWithDamageKHR</span><span class="w"></span>
<span class="n">eglSwapBuffers</span><span class="w"></span>
<span class="n">glXSwapBuffers</span><span class="w"></span>
<span class="n">glXQueryDrawable</span><span class="w"></span>
<span class="n">glXGetCurrentReadDrawable</span><span class="w"></span>
<span class="n">glXGetCurrentDrawable</span><span class="w"></span>
<span class="n">glGetQueryObjectuivEXT</span><span class="w"></span>
<span class="n">glGetQueryObjectuivARB</span><span class="w"></span>
<span class="n">glGetQueryObjectuiv</span><span class="w"></span>
<span class="n">glGetQueryObjectivARB</span><span class="w"></span>
<span class="n">glGetQueryObjectiv</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
<section id="custom-etw-trace">
<h2>Custom ETW Trace<a class="headerlink" href="#custom-etw-trace" title="Permalink to this heading"></a></h2>
<p>Use the custom ETW trace feature to enable and collect any manifest-based ETW log. The collected events are displayed on the timeline on dedicated rows for each event type.</p>
<p>Custom ETW is available on Windows target machines.</p>
<blockquote>
<div><img alt="Adding details of an ETW provider" class="image" src="../_images/custom-etw-add-provider.png" />
</div></blockquote>
<blockquote>
<div><img alt="Adding an ETW provider to the trace settings" class="image" src="../_images/custom-etw-trace-settings.png" />
</div></blockquote>
<img alt="Display of custom ETW trace events on the timeline" class="image" src="../_images/custom-etw-events-timeline.png" />
<p>To retain the .etl trace files captured, so that they can be viewed in other tools (e.g. GPUView), change the “Save ETW log files in project folder” option under “Profile Behavior” in Nsight Systems’s global Options dialog. The .etl files will appear in the same folder as the .nsys-rep file, accessible by right-clicking the report in the Project Explorer and choosing “Show in Folder…”. Data collected from each ETW provider will appear in its own .etl file, and an additional .etl file named “Report XX-Merged-*.etl”, containing the events from all captured sources, will be created as well.</p>
</section>
<section id="gpu-metrics">
<h2>GPU Metrics<a class="headerlink" href="#gpu-metrics" title="Permalink to this heading"></a></h2>
<section id="overview">
<h3>Overview<a class="headerlink" href="#overview" title="Permalink to this heading"></a></h3>
<p>GPU Metrics feature is intended to identify performance limiters in applications using GPU for computations and graphics. It uses periodic sampling to gather performance metrics and detailed timing statistics associated with different GPU hardware units taking advantage of specialized hardware to capture this data in a single pass with minimal overhead.</p>
<p><strong>Note:</strong> GPU Metrics will give you precise device level information, but it does not know which process or context is involved. GPU context switch trace provides less precise information, but will give you process and context information.</p>
<img alt="Example report with GPU Metrics" class="image" src="../_images/gpu-metrics-timeline.png" />
<p>These metrics provide an overview of GPU efficiency over time within compute, graphics, and input/output (IO) activities such as:</p>
<ul class="simple">
<li><p><strong>IO throughputs:</strong> PCIe, NVLink, and GPU memory bandwidth</p></li>
<li><p><strong>SM utilization:</strong> SMs activity, tensor core activity, instructions issued, warp occupancy, and unassigned warp slots</p></li>
</ul>
<p>It is designed to help users answer the common questions:</p>
<ul class="simple">
<li><p>Is my GPU idle?</p></li>
<li><p>Is my GPU full? Enough kernel grids size and streams? Are my SMs and warp slots full?</p></li>
<li><p>Am I using TensorCores?</p></li>
<li><p>Is my instruction rate high?</p></li>
<li><p>Am I possibly blocked on IO, or number of warps, etc</p></li>
</ul>
<p>Nsight Systems GPU Metrics is only available for Linux targets on x86-64 and aarch64, and for Windows targets. It requires NVIDIA Turing architecture or newer.</p>
<p>Minimum required driver versions:</p>
<ul class="simple">
<li><p>NVIDIA Turing architecture TU10x, TU11x - r440</p></li>
<li><p>NVIDIA Ampere architecture GA100 - r450</p></li>
<li><p>NVIDIA Ampere architecture GA100 MIG - r470 TRD1</p></li>
<li><p>NVIDIA Ampere architecture GA10x - r455</p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p><strong>Permissions:</strong> Elevated permissions are required. On Linux use sudo to elevate privileges. On Windows the user must run from an admin command prompt or accept the UAC escalation dialog. See <a class="reference external" href="https://developer.nvidia.com/ERR_NVGPUCTRPERM">Permissions Issues and Performance Counters</a> for more information.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p><strong>Tensor Core:</strong> If you run <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre">--gpu-metrics-device</span> <span class="pre">all</span></code>, the Tensor Core utilization can be found in the GUI under the <strong>SM instructions/Tensor Active</strong> row.</p>
<p>Please note that it is not practical to expect a CUDA kernel to reach 100% Tensor Core utilization since there are other overheads. In general, the more computation-intensive an operation is, the higher Tensor Core utilization rate the CUDA kernel can achieve.</p>
</div>
</section>
<section id="launching-gpu-metrics-from-the-cli">
<h3>Launching GPU Metrics from the CLI<a class="headerlink" href="#launching-gpu-metrics-from-the-cli" title="Permalink to this heading"></a></h3>
<p>GPU Metrics feature is controlled with 3 CLI switches:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-device=[all,</span> <span class="pre">none,</span> <span class="pre"><index>]</span></code> selects GPUs to sample (default is none)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-set=[<index>,</span> <span class="pre"><alias>]</span></code> selects metric set to use (default is the 1st suitable from the list)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--gpu-metrics-frequency=[10..200000]</span></code> selects sampling frequency in Hz (default is 10000)</p></li>
</ul>
<p>To profile with default options and sample GPU Metrics on GPU 0:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp"># Must have elevated permissions (see https:</span><span class="c1">//developer.nvidia.com/ERR_NVGPUCTRPERM) or be root (Linux) or Administrator (Windows)</span>
<span class="n">$</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="o">=</span><span class="mi">0</span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">my</span><span class="o">-</span><span class="n">app</span><span class="w"></span>
</pre></div>
</div>
<p>To list available GPUs, use:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="o">=</span><span class="n">help</span><span class="w"></span>
<span class="n">Possible</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="w"> </span><span class="n">values</span><span class="w"> </span><span class="n">are</span><span class="o">:</span><span class="w"></span>
<span class="w"> </span><span class="mi">0</span><span class="o">:</span><span class="w"> </span><span class="n">Quadro</span><span class="w"> </span><span class="n">GV100</span><span class="w"> </span><span class="n">PCI</span><span class="p">[</span><span class="mo">0000</span><span class="o">:</span><span class="mi">17</span><span class="o">:</span><span class="mf">00.0</span><span class="p">]</span><span class="w"></span>
<span class="w"> </span><span class="mi">1</span><span class="o">:</span><span class="w"> </span><span class="n">GeForce</span><span class="w"> </span><span class="n">RTX</span><span class="w"> </span><span class="mi">2070</span><span class="w"> </span><span class="n">SUPER</span><span class="w"> </span><span class="n">PCI</span><span class="p">[</span><span class="mo">0000</span><span class="o">:</span><span class="mi">65</span><span class="o">:</span><span class="mf">00.0</span><span class="p">]</span><span class="w"></span>
<span class="w"> </span><span class="nl">all</span><span class="p">:</span><span class="w"> </span><span class="n">Select</span><span class="w"> </span><span class="n">all</span><span class="w"> </span><span class="n">supported</span><span class="w"> </span><span class="n">GPUs</span><span class="w"></span>
<span class="w"> </span><span class="nl">none</span><span class="p">:</span><span class="w"> </span><span class="n">Disable</span><span class="w"> </span><span class="n">GPU</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="p">[</span><span class="n">Default</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>By default, the first <strong>metric set</strong> which supports all selected GPUs is used. But you can manually select another metric set from the list. To see available metric sets, use:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">set</span><span class="o">=</span><span class="n">help</span><span class="w"></span>
<span class="n">Possible</span><span class="w"> </span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">set</span><span class="w"> </span><span class="n">values</span><span class="w"> </span><span class="n">are</span><span class="o">:</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">tu10x</span><span class="p">]</span><span class="w"> </span><span class="n">General</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">TU10x</span><span class="w"> </span><span class="p">(</span><span class="n">any</span><span class="w"> </span><span class="n">frequency</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">tu11x</span><span class="p">]</span><span class="w"> </span><span class="n">General</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">TU11x</span><span class="w"> </span><span class="p">(</span><span class="n">any</span><span class="w"> </span><span class="n">frequency</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">2</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">ga100</span><span class="p">]</span><span class="w"> </span><span class="n">General</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">GA100</span><span class="w"> </span><span class="p">(</span><span class="n">any</span><span class="w"> </span><span class="n">frequency</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">3</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">ga10x</span><span class="p">]</span><span class="w"> </span><span class="n">General</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">GA10x</span><span class="w"> </span><span class="p">(</span><span class="n">any</span><span class="w"> </span><span class="n">frequency</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">4</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">tu10x</span><span class="o">-</span><span class="n">gfxt</span><span class="p">]</span><span class="w"> </span><span class="n">Graphics</span><span class="w"> </span><span class="n">Throughput</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">TU10x</span><span class="w"> </span><span class="p">(</span><span class="n">frequency</span><span class="w"> </span><span class="o">>=</span><span class="w"> </span><span class="mi">10</span><span class="n">kHz</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">5</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">ga10x</span><span class="o">-</span><span class="n">gfxt</span><span class="p">]</span><span class="w"> </span><span class="n">Graphics</span><span class="w"> </span><span class="n">Throughput</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">GA10x</span><span class="w"> </span><span class="p">(</span><span class="n">frequency</span><span class="w"> </span><span class="o">>=</span><span class="w"> </span><span class="mi">10</span><span class="n">kHz</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="mi">6</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="n">ga10x</span><span class="o">-</span><span class="n">gfxact</span><span class="p">]</span><span class="w"> </span><span class="n">Graphics</span><span class="w"> </span><span class="n">Async</span><span class="w"> </span><span class="n">Compute</span><span class="w"> </span><span class="n">Triage</span><span class="w"> </span><span class="n">Metrics</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">NVIDIA</span><span class="w"> </span><span class="n">GA10x</span><span class="w"> </span><span class="p">(</span><span class="n">frequency</span><span class="w"> </span><span class="o">>=</span><span class="w"> </span><span class="mi">10</span><span class="n">kHz</span><span class="p">)</span><span class="w"></span>
</pre></div>
</div>
<p>By default, <strong>sampling frequency</strong> is set to 10 kHz. But you can manually set it from 10 Hz to 200 kHz using</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">gpu</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">frequency</span><span class="o">=<</span><span class="n">value</span><span class="o">></span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="launching-gpu-metrics-from-the-gui">
<h3>Launching GPU Metrics from the GUI<a class="headerlink" href="#launching-gpu-metrics-from-the-gui" title="Permalink to this heading"></a></h3>
<p>For commands to launch GPU Metrics from the CLI with examples, see the <a class="reference external" href="topics/cli-profiling">CLI documentation</a>.</p>
<p>When launching analysis in Nsight Systems, select <strong>Collect GPU Metrics</strong>.</p>
<img alt="GPU Metrics: GUI overview" class="image" src="../_images/gpu-metrics-gui-overview.png" />
<p>Select the <strong>GPUs</strong> dropdown to pick which GPUs you wish to sample.</p>
<p>Select the <strong>Metric set:</strong> dropdown to choose which available metric set you would like to sample.</p>
<img alt="GPU Metrics: Metric sets" class="image" src="../_images/gpu-metrics-gui-metric-sets.png" />
<p>Note that metric sets for GPUs that are not being sampled will be greyed out.</p>
</section>
<section id="sampling-frequency">
<h3>Sampling frequency<a class="headerlink" href="#sampling-frequency" title="Permalink to this heading"></a></h3>
<p>Sampling frequency can be selected from the range of 10 Hz - 200 kHz. The default value is 10 kHz.</p>
<p>The maximum sampling frequency without buffer overflow events depends on GPU (SM count), GPU load intensity, and overall system load. The bigger the chip and the higher the load, the lower the maximum frequency. If you need higher frequency, you can increase it until you get “Buffer overflow” message in the Diagnostics Summary report page.</p>
<p>Each metric set has a recommended sampling frequency range in its description. These ranges are approximate. If you observe <code class="docutils literal notranslate"><span class="pre">Inconsistent</span> <span class="pre">Data</span></code> or <code class="docutils literal notranslate"><span class="pre">Missing</span> <span class="pre">Data</span></code> ranges on timeline, please try closer to the recommended frequency.</p>
</section>
<section id="available-metrics">
<h3>Available metrics<a class="headerlink" href="#available-metrics" title="Permalink to this heading"></a></h3>
<ul>
<li><p><strong>GPC Clock Frequency</strong> - <code class="docutils literal notranslate"><span class="pre">gpc__cycles_elapsed.avg.per_second</span></code></p>
<p>The average GPC clock frequency in hertz. In public documentation the GPC clock may be called the “Application” clock, “Graphic” clock, “Base” clock, or “Boost” clock.</p>
<p><strong>Note</strong>: The collection mechanism for GPC can result in a small fluctuation between samples.</p>
</li>
<li><p><strong>SYS Clock Frequency</strong> - <code class="docutils literal notranslate"><span class="pre">sys__cycles_elapsed.avg.per_second</span></code></p>
<p>The average SYS clock frequency in hertz. The GPU front end (command processor), copy engines, and the performance monitor run at the SYS clock. On Turing and NVIDIA GA100 GPUs the sampling frequency is based upon a period of SYS clocks (not time) so samples per second will vary with SYS clock. On NVIDIA GA10x GPUs the sampling frequency is based upon a fixed frequency clock. The maximum frequency scales linearly with the SYS clock.</p>
</li>
<li><p><strong>GR Active</strong> - <code class="docutils literal notranslate"><span class="pre">gr__cycles_active.sum.pct_of_peak_sustained_elapsed</span></code></p>
<p>The percentage of cycles the graphics/compute engine is active. The graphics/compute engine is active if there is any work in the graphics pipe or if the compute pipe is processing work.</p>
<p>GA100 MIG - MIG is not yet supported. This counter will report the activity of the primary GR engine.</p>
</li>
<li><p><strong>Sync Compute In Flight</strong> - <code class="docutils literal notranslate"><span class="pre">gr__dispatch_cycles_active_queue_sync.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The percentage of cycles with synchronous compute in flight.</p>
<p>CUDA: CUDA will only report synchronous queue in the case of MPS configured with 64 sub-context. Synchronous refers to work submitted in VEID=0.</p>
<p>Graphics: This will be true if any compute work submitted from the direct queue is in flight.</p>
</li>
<li><p><strong>Async Compute in Flight</strong> - <code class="docutils literal notranslate"><span class="pre">gr__dispatch_cycles_active_queue_async.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The percentage of cycles with asynchronous compute in flight.</p>
<p>CUDA: CUDA will only report all compute work as asynchronous. The one exception is if MPS is configured and all 64 sub-context are in use. 1 sub-context (VEID=0) will report as synchronous.</p>
<p>Graphics: This will be true if any compute work submitted from a compute queue is in flight.</p>
</li>
<li><p><strong>Draw Started</strong> - <code class="docutils literal notranslate"><span class="pre">fe__draw_count.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of draw calls issued to the graphics pipe to the maximum sustained rate of the graphics pipe.</p>
<p><strong>Note:</strong>The percentage will always be very low as the front end can issue draw calls significantly faster than the pipe can execute the draw call. The rendering of this row will be changed to help indicate when draw calls are being issued.</p>
</li>
<li><p><strong>Dispatch Started</strong> - <code class="docutils literal notranslate"><span class="pre">gr__dispatch_count.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of compute grid launches (dispatches) to the compute pipe to the maximum sustained rate of the compute pipe.</p>
<p><strong>Note:</strong> The percentage will always be very low as the front end can issue grid launches significantly faster than the pipe can execute the draw call. The rendering of this row will be changed to help indicate when grid launches are being issued.</p>
</li>
<li><p><strong>Vertex/Tess/Geometry Warps in Flight</strong> - <code class="docutils literal notranslate"><span class="pre">tpc__warps_active_shader_vtg_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of active vertex, geometry, tessellation, and meshlet shader warps resident on the SMs to the maximum number of warps per SM as a percentage.</p>
</li>
<li><p><strong>Pixel Warps in Flight</strong> - <code class="docutils literal notranslate"><span class="pre">tpc__warps_active_shader_ps_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of active pixel/fragment shader warps resident on the SMs to the maximum number of warps per SM as a percentage.</p>
</li>
<li><p><strong>Compute Warps in Flight</strong> - <code class="docutils literal notranslate"><span class="pre">tpc__warps_active_shader_cs_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of active compute shader warps resident on the SMs to the maximum number of warps per SM as a percentage.</p>
</li>
<li><p><strong>Active SM Unused Warp Slots</strong> - <code class="docutils literal notranslate"><span class="pre">tpc__warps_inactive_sm_active_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of inactive warp slots on the SMs to the maximum number of warps per SM as a percentage. This is an indication of how many more warps may fit on the SMs if occupancy is not limited by a resource such as max warps of a shader type, shared memory, registers per thread, or thread blocks per SM.</p>
</li>
<li><p><strong>Idle SM Unused Warp Slots</strong> - <code class="docutils literal notranslate"><span class="pre">tpc__warps_inactive_sm_idle_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of inactive warps slots due to idle SMs to the the maximum number of warps per SM as a percentage.</p>
</li>
</ul>
<p>This is an indicator that the current workload on the SM is not sufficient to put work on all SMs. This can be due to:</p>
<ul>
<li><p>CPU starving the GPU</p></li>
<li><p>current work is too small to saturate the GPU</p></li>
<li><p>current work is trailing off but blocking next work</p></li>
<li><p><strong>SM Active</strong> - <code class="docutils literal notranslate"><span class="pre">sm__cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of cycles SMs had at least 1 warp in flight (allocated on SM) to the number of cycles as a percentage. A value of 0 indicates all SMs were idle (no warps in flight). A value of 50% can indicate some gradient between all SMs active 50% of the sample period or 50% of SMs active 100% of the sample period.</p>
</li>
<li><p><strong>SM Issue</strong> - <code class="docutils literal notranslate"><span class="pre">sm__inst_executed_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of cycles that SM sub-partitions (warp schedulers) issued an instruction to the number of cycles in the sample period as a percentage.</p>
</li>
<li><p><strong>Tensor Active</strong> - <code class="docutils literal notranslate"><span class="pre">sm__pipe_tensor_cycles_active_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of cycles the SM tensor pipes were active issuing tensor instructions to the number of cycles in the sample period as a percentage.</p>
<p>TU102/4/6: This metric is not available on TU10x for periodic sampling. Please see Tensor Active/FP16 Active.</p>
</li>
<li><p><strong>Tensor Active / FP16 Active</strong> - <code class="docutils literal notranslate"><span class="pre">sm__pipe_shared_cycles_active_realtime.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>TU102/4/6 only</p>
<p>The ratio of cycles the SM tensor pipes or FP16x2 pipes were active issuing tensor instructions to the number of cycles in the sample period as a percentage.</p>
</li>
<li><p><strong>DRAM Read Bandwidth</strong> - <code class="docutils literal notranslate"><span class="pre">dramc__read_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">dram__read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p></li>
<li><p><strong>VRAM Read Bandwidth</strong> - <code class="docutils literal notranslate"><span class="pre">FBPA.TriageA.dramc__read_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">FBSP.TriageSCG.dramc__read_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">FBSP.TriageAC.dramc__read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of cycles the DRAM interface was active reading data to the elapsed cycles in the same period as a percentage.</p>
</li>
<li><p><strong>DRAM Write Bandwidth</strong> - <code class="docutils literal notranslate"><span class="pre">dramc__write_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">dram__write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p></li>
<li><p><strong>VRAM Write Bandwidth</strong> - <code class="docutils literal notranslate"><span class="pre">FBPA.TriageA.dramc__write_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">FBSP.TriageSCG.dramc__write_throughput.avg.pct_of_peak_sustained_elapsed</span></code>, <code class="docutils literal notranslate"><span class="pre">FBSP.TriageAC.dramc__write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of cycles the DRAM interface was active writing data to the elapsed cycles in the same period as a percentage.</p>
</li>
<li><p><strong>NVLink bytes received</strong> - <code class="docutils literal notranslate"><span class="pre">nvlrx__bytes.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of bytes received on the NVLink interface to the maximum number of bytes receivable in the sample period as a percentage. This value includes protocol overhead.</p>
</li>
<li><p><strong>NVLink bytes transmitted</strong> - <code class="docutils literal notranslate"><span class="pre">nvltx__bytes.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of bytes transmitted on the NVLink interface to the maximum number of bytes transmittable in the sample period as a percentage. This value includes protocol overhead.</p>
</li>
<li><p><strong>PCIe Read Throughput</strong> - <code class="docutils literal notranslate"><span class="pre">pcie__read_bytes.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of bytes received on the PCIe interface to the maximum number of bytes receivable in the sample period as a percentage. The theoretical value is calculated based upon the PCIe generation and number of lanes. This value includes protocol overhead.</p>
</li>
<li><p><strong>PCIe Write Throughput</strong> - <code class="docutils literal notranslate"><span class="pre">pcie__write_bytes.avg.pct_of_peak_sustained_elapsed</span></code></p>
<p>The ratio of bytes transmitted on the PCIe interface to the maximum number of bytes receivable in the sample period as a percentage. The theoretical value is calculated based upon the PCIe generation and number of lanes. This value includes protocol overhead.</p>
</li>
<li><p><strong>PCIe Read Requests to BAR1</strong> - <code class="docutils literal notranslate"><span class="pre">pcie__rx_requests_aperture_bar1_op_read.sum</span></code></p></li>
<li><p><strong>PCIe Write Requests to BAR1</strong> - <code class="docutils literal notranslate"><span class="pre">pcie__rx_requests_aperture_bar1_op_write.sum</span></code></p>
<p>BAR1 is a PCI Express (PCIe) interface used to allow the CPU or other devices to directly access GPU memory. The GPU normally transfers memory with its copy engines, which would not show up as BAR1 activity. The GPU drivers on the CPU do a small amount of BAR1 accesses, but heavier traffic is typically coming from other technologies.</p>
<p>On Linux, technologies like GPU Direct, GPU Direct RDMA, and GPU Direct Storage transfer data across PCIe BAR1. In the case of GPU Direct RDMA, that would be an Ethernet or InfiniBand adapter directly writing to GPU memory.</p>
<p>On Windows, Direct3D12 resources can also be made accessible directly to the CPU via NVAPI functions to support small writes or reads from GPU buffers, in this case too many BAR1 accesses can indicate a performance issue, like it has been demonstrated in the Optimizing DX12 Resource Uploads to the GPU Using CPU-Visible VRAM technical blog post.</p>
</li>
</ul>
</section>
<section id="exporting-and-querying-data">
<h3>Exporting and Querying Data<a class="headerlink" href="#exporting-and-querying-data" title="Permalink to this heading"></a></h3>
<p>It is possible to access metric values for automated processing using the Nsight Systems CLI export capabilities.</p>
<p>An example that extracts values of “SM Active”:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="k">export</span><span class="w"> </span><span class="o">-</span><span class="n">t</span><span class="w"> </span><span class="n">sqlite</span><span class="w"> </span><span class="n">report</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"></span>
<span class="n">$</span><span class="w"> </span><span class="n">sqlite3</span><span class="w"> </span><span class="n">report</span><span class="p">.</span><span class="n">sqlite</span><span class="w"> </span><span class="s">"SELECT rawTimestamp, CAST(JSON_EXTRACT(data, '$.</span><span class="se">\"</span><span class="s">SM Active</span><span class="se">\"</span><span class="s">') as INTEGER) as value FROM GENERIC_EVENTS WHERE value != 0 LIMIT 10"</span><span class="w"></span>
<span class="mi">309277039</span><span class="o">|</span><span class="mi">80</span><span class="w"></span>
<span class="mi">309301295</span><span class="o">|</span><span class="mi">99</span><span class="w"></span>
<span class="mi">309325583</span><span class="o">|</span><span class="mi">99</span><span class="w"></span>
<span class="mi">309349776</span><span class="o">|</span><span class="mi">99</span><span class="w"></span>
<span class="mi">309373872</span><span class="o">|</span><span class="mi">60</span><span class="w"></span>
<span class="mi">309397872</span><span class="o">|</span><span class="mi">19</span><span class="w"></span>
<span class="mi">309421840</span><span class="o">|</span><span class="mi">100</span><span class="w"></span>
<span class="mi">309446000</span><span class="o">|</span><span class="mi">100</span><span class="w"></span>
<span class="mi">309470096</span><span class="o">|</span><span class="mi">100</span><span class="w"></span>
<span class="mi">309494161</span><span class="o">|</span><span class="mi">99</span><span class="w"></span>
</pre></div>
</div>
<p>An overview of data stored in each event (JSON):</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">sqlite3</span><span class="w"> </span><span class="n">report</span><span class="p">.</span><span class="n">sqlite</span><span class="w"> </span><span class="s">"SELECT data FROM GENERIC_EVENTS LIMIT 1"</span><span class="w"></span>
<span class="p">{</span><span class="w"></span>
<span class="s">"Unallocated Warps in Active SM"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Compute Warps In Flight"</span><span class="o">:</span><span class="w"> </span><span class="s">"52"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Pixel Warps In Flight"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Vertex\/Tess\/Geometry Warps In Flight"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Total SM Occupancy"</span><span class="o">:</span><span class="w"> </span><span class="s">"52"</span><span class="p">,</span><span class="w"></span>
<span class="s">"GR Active (GE\/CE)"</span><span class="o">:</span><span class="w"> </span><span class="s">"100"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Sync Compute In Flight"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Async Compute In Flight"</span><span class="o">:</span><span class="w"> </span><span class="s">"98"</span><span class="p">,</span><span class="w"></span>
<span class="s">"NVLink bytes received"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"NVLink bytes transmitted"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"PCIe Rx Throughput"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"PCIe Tx Throughput"</span><span class="o">:</span><span class="w"> </span><span class="s">"1"</span><span class="p">,</span><span class="w"></span>
<span class="s">"DRAM Read Throughput"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"DRAM Write Throughput"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"Tensor Active \/ FP16 Active"</span><span class="o">:</span><span class="w"> </span><span class="s">"0"</span><span class="p">,</span><span class="w"></span>
<span class="s">"SM Issue"</span><span class="o">:</span><span class="w"> </span><span class="s">"10"</span><span class="p">,</span><span class="w"></span>
<span class="s">"SM Active"</span><span class="o">:</span><span class="w"> </span><span class="s">"52"</span><span class="w"></span>
<span class="p">}</span><span class="w"></span>
</pre></div>
</div>
<p>Values are integer percentages (0..100)</p>
</section>
<section id="id2">
<h3>Limitations<a class="headerlink" href="#id2" title="Permalink to this heading"></a></h3>
<ul>
<li><p>If metric sets with NVLink are used but the links are not active, they may appear as fully utilized.</p></li>
<li><p>Only one tool that subscribes to these counters can be used at a time, therefore, Nsight Systems GPU Metrics feature cannot be used at the same time as the following tools:</p>
<ul>
<li><p>Nsight Graphics</p></li>
<li><p>Nsight Compute</p></li>
<li><p>DCGM (Data Center GPU Manager)</p>
<p>Use the following command:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">dcgmi</span> <span class="pre">profile</span> <span class="pre">--pause</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">dcgmi</span> <span class="pre">profile</span> <span class="pre">--resume</span></code></p></li>
</ul>
<p>Or API:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">dcgmProfPause</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">dcgmProfResume</span></code></p></li>
</ul>
</li>
<li><p>Non-NVIDIA products which use:</p>
<ul class="simple">
<li><p>CUPTI sampling used directly in the application. CUPTI trace is okay (although it will block Nsight Systems CUDA trace)</p></li>
<li><p>DCGM library</p></li>
</ul>
</li>
</ul>
</li>
<li><p>Nsight Systems limits the amount of memory that can be used to store GPU Metrics samples. Analysis with higher sampling rates or on GPUs with more SMs has a risk of exceeding this limit. This will lead to gaps on timeline filled with <code class="docutils literal notranslate"><span class="pre">Missing</span> <span class="pre">Data</span></code> ranges. Future releases will reduce the frequency of this happening.</p></li>
</ul>
</section>
</section>
<section id="soc-metrics">
<h2>SOC Metrics<a class="headerlink" href="#soc-metrics" title="Permalink to this heading"></a></h2>
<section id="id3">
<h3>Overview<a class="headerlink" href="#id3" title="Permalink to this heading"></a></h3>
<p>SOC Metrics feature is intended to identify performance limiters in applications running on NVIDIA SOCs and is similar to GPU Metrics.</p>
<p>Nsight Systems SOC Metrics is only available for Linux and QNX targets on aarch64. It requires NVIDIA Orin architecture or newer.</p>
</section>
<section id="id4">
<h3>Available metrics<a class="headerlink" href="#id4" title="Permalink to this heading"></a></h3>
<ul>
<li><dl class="simple">
<dt><strong>CPU Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_cpu_op_read.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>CPU Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_cpu_op_write.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the SOC memory controllers were actively processing read/write operations from the CPU to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>GPU Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_gpu_op_read.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>GPU Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_gpu_op_write.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the SOC memory controllers were actively processing read/write operations from the GPU to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>DBB Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_dbb_op_read.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>DBB Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_srcnode_dbb_op_write.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the SOC memory controllers were actively processing read/write operations from not-CPU/not-GPU to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>DRAM Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_op_read.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>DRAM Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">mcc__dram_throughput_op_write.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the SOC memory controllers were actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>DLA Active</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvdla__cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the DLA (Deep Learning Accelerator) was actively processing a command to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>DLA Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvdla__dbb2nvdla_read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>DLA Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvdla__nvdla2dbb_write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the DLA (Deep Learning Accelerator) was actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>NVENC Active</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvenc__cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the NVENC unit was actively processing a command to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>NVENC Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvenc__memif2nvenc_read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>NVENC Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">nvenc__nvenc2memif_write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the NVENC unit was actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>PVA VPU Active</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">pvavpu__vpu_cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the PVA (Programmable Vision Accelerator) VPU (Vector Processing Unit) was actively processing a command to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>PVA DMA Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">pva__dbb2pvadma_read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>PVA DMA Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">pva__pvadma2dbb_write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the PVA (Programmable Vision Accelerator) VPU (Vector Processing Unit) was actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>OFA Active</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">ofa_cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the OFA (Optical Flow Accelerator) was actively processing a command to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>OFA Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">ofa__memif2ofa_read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>OFA Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">ofa__ofa2memif_write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the OFA (Optical Flow Accelerator) was actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>VIC Active</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">vic_cycles_active.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the VIC (Video Image Compositor) was actively processing a command to the number of cycles in the same sample period as a percentage.</p>
</li>
<li><dl class="simple">
<dt><strong>VIC Read Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">vic__dbb2vic_read_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
<dt><strong>VIC Write Throughput</strong></dt><dd><p><code class="docutils literal notranslate"><span class="pre">vic__vic2dbb_write_throughput.avg.pct_of_peak_sustained_elapsed</span></code></p>
</dd>
</dl>
<p>The ratio of cycles the VIC (Video Image Compositor) was actively processing read/write operations to the number of cycles in the same sample period as a percentage.</p>
</li>
</ul>
</section>
<section id="launching-soc-metrics-from-the-cli">
<h3>Launching SOC Metrics from the CLI<a class="headerlink" href="#launching-soc-metrics-from-the-cli" title="Permalink to this heading"></a></h3>
<p>SOC Metrics feature is controlled with 3 CLI switches:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics=[true,</span> <span class="pre">false]</span></code> enables SOC Metrics sampling (default is false)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-set=[<alias>]</span></code> selects metric set to use (default is the 1st suitable from the list)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--soc-metrics-frequency=[100..200000]</span></code> selects sampling frequency in Hz (default is 10000)</p></li>
</ul>
<p>To profile with default options:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp"># Must have elevated permissions (see https:</span><span class="c1">//developer.nvidia.com/ERR_NVGPUCTRPERM) or be root</span>
<span class="n">$</span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">soc</span><span class="o">-</span><span class="n">metrics</span><span class="o">=</span><span class="nb">true</span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">my</span><span class="o">-</span><span class="n">app</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="launching-soc-metrics-from-the-gui">
<h3>Launching SOC Metrics from the GUI<a class="headerlink" href="#launching-soc-metrics-from-the-gui" title="Permalink to this heading"></a></h3>
<p>When launching analysis in Nsight Systems, select <strong>Collect SOC Metrics</strong>.</p>
<p>The settings are similar to <a class="reference external" href="topics/gpu-metrics-gui">GPU Metrics</a>.</p>
<p>For commands to launch SOC Metrics from the CLI with examples, see the <a class="reference external" href="topics/cli-profiling">CLI documentation</a>.</p>
</section>
</section>
<section id="cpu-profiling-using-linux-os-perf-subsystem">
<h2>CPU Profiling Using Linux OS Perf Subsystem<a class="headerlink" href="#cpu-profiling-using-linux-os-perf-subsystem" title="Permalink to this heading"></a></h2>
<p>Nsight Systems on Linux targets, utilizes the Linux OS’ perf subsystem to sample CPU Instruction Pointers (IPs) and backtraces, trace CPU context switches, and sample CPU and OS event counts. The Linux perf tool utilizes the same perf subsystem.</p>
<p>Nsight Systems, on L4T and potentially other ARM targets, may use a custom kernel module to collect the same data. The Nsight Systems CLI command <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--environment</span></code> indicates when the kernel module is used instead of the Linux OS’ perf subsystem.</p>
<p><strong>Features</strong></p>
<ul>
<li><p><strong>CPU Instruction Pointer / Backtrace Sampling</strong></p>
<p>Nsight Systems can sample CPU Instruction Pointers / backtraces periodically. The collection of a sample is triggered by a hardware event overflow - e.g. a sample is collected after every 1 million CPU reference cycles on a per thread basis. In the GUI, samples are shown on the individual thread timelines, in the Event Viewer, and in the Top Down, Bottom Up, or Flat views which provide histogram-like summaries of the data. IP / backtrace collections can be configured in process-tree or system-wide mode. In process-tree mode, Nsight Systems will sample the process, and any of its descendants, launched by the tool. In system-wide mode, Nsight Systems will sample all processes running on the system, including any processes launched by the tool.</p>
</li>
<li><p><strong>CPU Context Switch Tracing</strong></p>
<p>Nsight Systems can trace every time the OS schedules a thread on a logical CPU and every time the OS thread gets unscheduled from a logical CPU. The data is used to show CPU utilization and OS thread utilization within the Nsight Systems GUI. Context switch collections can be configured in process-tree or system-wide mode. In process-tree mode, Nsight Systems will trace the process, and any of its descendants, launched by Nsight Systems. In system-wide mode, Nsight Systems will trace all processes running on the system, including any processes launched by the Nsight Systems.</p>
</li>
<li><p><strong>CPU Event Sampling</strong></p>
<p>Nsight Systems can periodically sample CPU hardware event counts and OS event counts and show the event’s rate over time in the Nsight Systems GUI. Event sample collections can be configured in system-wide mode only. In system-wide mode, Nsight Systems will sample event counts of all CPUs and the OS event counts running on the system. Event counts are not directly associated with processes or threads.</p>
</li>
</ul>
<p><strong>System Requirements</strong></p>
<ul>
<li><p><strong>Paranoid Level</strong></p>
<p>The <a class="reference external" href="https://www.kernel.org/doc/Documentation/sysctl/kernel.txt">system’s paranoid level</a> must be 2 or lower.</p>
</li>
</ul>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Paranoid Level</p></th>
<th class="head"><p>CPU IP/backtrace Sampling process-tree mode</p></th>
<th class="head"><p>CPU IP/backtrace Sampling system-wide mode</p></th>
<th class="head"><p>CPU Context Switch Tracing process-tree mode</p></th>
<th class="head"><p>CPU Context Switch Tracing system-wide mode</p></th>
<th class="head"><p>Event Sampling system-wide mode</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>3 or greater</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
</tr>
<tr class="row-odd"><td><p>2</p></td>
<td><p>User mode IP/backtrace samples only</p></td>
<td><p>not available</p></td>
<td><p>available</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
</tr>
<tr class="row-even"><td><p>1</p></td>
<td><p>Kernel and user mode IP/backtrace samples</p></td>
<td><p>not available</p></td>
<td><p>available</p></td>
<td><p>not available</p></td>
<td><p>not available</p></td>
</tr>
<tr class="row-odd"><td><p>0, -1</p></td>
<td><p>Kernel and user mode IP/backtrace samples</p></td>
<td><p>Kernel and user mode IP/backtrace samples</p></td>
<td><p>available</p></td>
<td><p>available</p></td>
<td><p>hardware and OS events</p></td>
</tr>
</tbody>
</table>
<ul>
<li><p><strong>Kernel Version</strong></p>
<p>To support the CPU profiling features utilized by Nsight Systems, the kernel version must be greater than or equal to v4.3. RedHat has backported the required features to the v3.10.0-693 kernel. RedHat distros and their derivatives (e.g. CentOS) require a 3.10.0-693 or later kernel. Use the <code class="docutils literal notranslate"><span class="pre">uname</span> <span class="pre">-r</span></code> command to check the kernel’s version.</p>
</li>
<li><p><strong>perf_event_open syscall</strong></p>
<p>The perf_event_open syscall needs to be available. When running within a Docker container, the default seccomp settings will normally block the perf_event_open syscall. To workaround this issue, use the Docker <code class="docutils literal notranslate"><span class="pre">run</span> <span class="pre">--privileged</span></code> switch when launching the docker or modify the docker’s seccomp settings. Some VMs (virtual machines), e.g. AWS, may also block the perf_event_open syscall.</p>
</li>
<li><p><strong>Sampling Trigger</strong></p>
<p>In some rare case, a sampling trigger is not available. The sampling trigger is either a hardware or software event that causes a sample to be collected. Some VMs block hardware events from being accessed and therefore, prevent hardware events from being used as sampling triggers. In those cases, Nsight Systems will fall back to using a software trigger if possible.</p>
</li>
<li><p><strong>Checking Your Target System</strong></p>
<p>Use the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--environment</span></code> command to check if a system meets the Nsight Systems CPU profiling requirements. Example output from this command is shown below. Note that this command does not check for Linux capability overrides - i.e. if the user or executable files have CAP_SYS_ADMIN or CAP_PERFMON capability. Also, note that this command does not indicate if system-wide mode can be used.</p>
<img alt="environment status output" class="image" src="../_images/sampling-environment-check2022.png" />
</li>
</ul>
<p><strong>Configuring a CPU Profiling Collection</strong></p>
<p>When configuring Nsight Systems for CPU Profiling from the CLI, use some or all of the following options: <code class="docutils literal notranslate"><span class="pre">--sample</span></code>, <code class="docutils literal notranslate"><span class="pre">--cpuctxsw</span></code>, <code class="docutils literal notranslate"><span class="pre">--event-sample</span></code>, <code class="docutils literal notranslate"><span class="pre">--backtrace</span></code>, <code class="docutils literal notranslate"><span class="pre">--cpu-core-events</span></code>, <code class="docutils literal notranslate"><span class="pre">--event-sampling-frequency</span></code>, <code class="docutils literal notranslate"><span class="pre">--os-events</span></code>, <code class="docutils literal notranslate"><span class="pre">--samples-per-backtrace</span></code>, and <code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code>.</p>
<p>Details about these options, including examples can be found in the Profiling from the CLI section of the User Guide</p>
<p>When configuring from the GUI, the following options are available:</p>
<img alt="GUI configuration for CPU profiling" class="image" src="../_images/linuxperf-gui-config.png" />
<p>The configuration used during CPU profiling is documented in the Analysis Summary:</p>
<img alt="CPU profiling in analysis summary" class="image" src="../_images/lp-options.png" />
<p>As well as in the Diagnosics Summary:</p>
<img alt="CPU profiling in diagnostics summary" class="image" src="../_images/lp-diagnostics-page.png" />
<p><strong>Visualizing CPU Profiling Results</strong></p>
<p>Here are example screenshots visualizing CPU profiling results. For details about navigating the Timeline View and the backtraces, see the section on Timeline View in the Reading Your Report in the GUI section of the User Guide.</p>
<p>Example of CPU IP/Backtrace Data</p>
<img alt="Timeline showing CPU IP/backtrace information" class="image" src="../_images/cpu-ipbacktrace.png" />
<p>In the timeline, yellow-orange marks can be found under each thread’s timeline that indicate the moment an IP / backtrace sample was collected on that thread (e.g. see the yellow-orange marks in the Specific Samples box above). Hovering the cursor over a mark will cause a tooltip to display the backtrace for that sample.</p>
<p>Below the Timeline is a drop-down list with multiple options including Events View, Top-Down View, Bottom-Up View, and Flat View. All four of these views can be used to view CPU IP / back trace sampling data.</p>
<p>Example of Event Sampling</p>
<img alt="CPU Event Sampling in GUI" class="image" src="../_images/cpu-eventsampling.png" />
<p>Event sampling samples hardware or software event counts during a collection and then graphs those events as rates on the Timeline. The above screenshot shows 4 hardware events. Core and cache events are graphed under the associated CPU row (see the red box in the screenshot) while uncore and OS events are graphed in their own row (see the green box in the screenshot). Hovering the cursor over an event sampling row in the timeline shows the event’s rate at that moment.</p>
<p><strong>Common Issues</strong></p>
<ul>
<li><p><strong>Reducing Overhead Caused By Sampling</strong></p>
<p>There are several ways to reduce overhead caused by sampling.</p>
<ul class="simple">
<li><p>disable sampling (i.e. use the <code class="docutils literal notranslate"><span class="pre">--sampling=none</span></code> switch)</p></li>
<li><p>increase the sampling period (i.e. reduce the sampling rate) using the <code class="docutils literal notranslate"><span class="pre">--sampling-period</span></code> switch</p></li>
<li><p>stop collecting backtraces (i.e. use the <code class="docutils literal notranslate"><span class="pre">--backtrace=none</span></code> switch) or collect more efficient backtraces - if available, use the <code class="docutils literal notranslate"><span class="pre">--backtrace=lbr</span></code> switch.</p></li>
<li><p>reduce the number of backtraces collected per sample. See documentation for the <code class="docutils literal notranslate"><span class="pre">--samples-per-backtrace</span></code> switch.</p></li>
</ul>
</li>
<li><p><strong>Throttling</strong></p>
<p>The Linux operating system enforces a maximum time to handle sampling interrupts. This means that if collecting samples takes more than a specified amount of time, the OS will throttle (i.e slow down) the sampling rate to prevent the perf subsystem from causing too much overhead. When this occurs, sampling data may become irregular even though the thread is very busy.</p>
<img alt="Throttling see in GUI" class="image" src="../_images/throttling.png" />
<p>The above screenshot shows a case where CPU IP / backtrace sampling was throttled during a collection. Note the irregular intervals of sampling tickmarks on the thread timeline. The number of times a collection throttled is provided in the Nsight Systems GUI’s Diagnostics messages. If a collection throttles frequently (e.g. 1000s of times), increasing the sampling period should help reduce throttling.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>When throttling occurs, the OS sets a new (lower) maximum sampling rate in the procfs. This value must be reset before the sampling rate can be increased again. Use the following command to reset the OS’ max sampling rate <code class="docutils literal notranslate"><span class="pre">echo</span> <span class="pre">'100000'</span> <span class="pre">|</span> <span class="pre">sudo</span> <span class="pre">tee</span> <span class="pre">/proc/sys/kernel/perf_event_max_sample_rate</span></code></p>
</div>
</li>
<li><p><strong>Sample intervals are irregular</strong></p>
<p>My samples are not periodic - why? My samples are clumped up - why? There are gaps in between the samples - why? Likely reasons:</p>
<ul class="simple">
<li><p>Throttling, as described above</p></li>
<li><p>The paranoid level is set to 2. If the paranoid level is set to 2, anytime the workload makes a system call and spends time executing kernel mode code, samples will not be collected and there will be gaps in the sampling data.</p></li>
<li><p>The sampling trigger itself is not periodic. If the trigger event is not periodic, for example, the Instructions Retired. event, sample collection will primarily occur when cache misses are occurring.</p></li>
</ul>
</li>
<li><p><strong>No CPU profiling data is collected</strong></p>
<p>There are a few common issues that cause CPU profiling data to not be collected</p>
<ul class="simple">
<li><p>System requirements are not met. Check your system settings with the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--environment</span></code> command and see the System Requirements section above.</p></li>
<li><p>I profiled my workload in a Docker container but no sampling data was collected. By default, Docker containers prevent the perf_event_open syscall from being utilized. To override this behavior, launch the Docker with the <code class="docutils literal notranslate"><span class="pre">--privileged</span></code> switch or modify the Docker’s <a class="reference external" href="https://docs.nvidia.com/nsight-systems/index.html#nsight_systems/2018.3.0-x86/07-docker-profiling.htm">seccomp</a>settings.</p></li>
<li><p>I profiled my workload in a Docker container running Ubuntu 20+ running on top of a host system running CentOS with a kernel version < 3.10.0-693. The <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--environment</span></code> command indicated that CPU profiling was supported. The host OS kernel version determines if CPU profiling is allowed and a CentOS host with a version < 3.10.0-693 is too old. In this case, the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--environment</span></code> command is incorrect.</p></li>
</ul>
</li>
</ul>
</section>
<section id="nvidia-video-codec-sdk-trace">
<h2>NVIDIA Video Codec SDK Trace<a class="headerlink" href="#nvidia-video-codec-sdk-trace" title="Permalink to this heading"></a></h2>
<p>Nsight Systems for x86 Linux and Windows targets can trace calls from the NV Video Codec SDK. This software trace can be launched from the GUI or using the <code class="docutils literal notranslate"><span class="pre">--trace</span> <span class="pre">nvvideo</span></code> from the CLI</p>
<img alt="NV Video Codec SDK trace selection" class="image" src="../_images/select-video-codec.png" />
<p>On the timeline, calls on the CPU to the NV Encoder API and NV Decoder API will be shown.</p>
<img alt="NV Video Codec SDK trace in timeline" class="image" src="../_images/video-codec-trace.png" />
<section id="nv-encoder-api-functions-traced-by-default">
<h3>NV Encoder API Functions Traced by Default<a class="headerlink" href="#nv-encoder-api-functions-traced-by-default" title="Permalink to this heading"></a></h3>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">NvEncodeAPICreateInstance</span><span class="w"></span>
<span class="n">nvEncOpenEncodeSession</span><span class="w"></span>
<span class="n">nvEncGetEncodeGUIDCount</span><span class="w"></span>
<span class="n">nvEncGetEncodeGUIDs</span><span class="w"></span>
<span class="n">nvEncGetEncodeProfileGUIDCount</span><span class="w"></span>
<span class="n">nvEncGetEncodeProfileGUIDs</span><span class="w"></span>
<span class="n">nvEncGetInputFormatCount</span><span class="w"></span>
<span class="n">nvEncGetInputFormats</span><span class="w"></span>
<span class="n">nvEncGetEncodeCaps</span><span class="w"></span>
<span class="n">nvEncGetEncodePresetCount</span><span class="w"></span>
<span class="n">nvEncGetEncodePresetGUIDs</span><span class="w"></span>
<span class="n">nvEncGetEncodePresetConfig</span><span class="w"></span>
<span class="n">nvEncGetEncodePresetConfigEx</span><span class="w"></span>
<span class="n">nvEncInitializeEncoder</span><span class="w"></span>
<span class="n">nvEncCreateInputBuffer</span><span class="w"></span>
<span class="n">nvEncDestroyInputBuffer</span><span class="w"></span>
<span class="n">nvEncCreateBitstreamBuffer</span><span class="w"></span>
<span class="n">nvEncDestroyBitstreamBuffer</span><span class="w"></span>
<span class="n">nvEncEncodePicture</span><span class="w"></span>
<span class="n">nvEncLockBitstream</span><span class="w"></span>
<span class="n">nvEncUnlockBitstream</span><span class="w"></span>
<span class="n">nvEncLockInputBuffer</span><span class="w"></span>
<span class="n">nvEncUnlockInputBuffer</span><span class="w"></span>
<span class="n">nvEncGetEncodeStats</span><span class="w"></span>
<span class="n">nvEndGetSequenceParams</span><span class="w"></span>
<span class="n">nvEncRegisterAsyncEvent</span><span class="w"></span>
<span class="n">nvEncUnregisterAsyncEvent</span><span class="w"></span>
<span class="n">nvEncMapInputResource</span><span class="w"></span>
<span class="n">nvEncUnmapInputResource</span><span class="w"></span>
<span class="n">nvEncDestroyEncoder</span><span class="w"></span>
<span class="n">nvEncInvalidateRefFrames</span><span class="w"></span>
<span class="n">nvEncOpenEncodeSessionEx</span><span class="w"></span>
<span class="n">nvEncRegisterResource</span><span class="w"></span>
<span class="n">nvEncUnregisterResource</span><span class="w"></span>
<span class="n">nvEncReconfigureEncoder</span><span class="w"></span>
<span class="n">nvEncCreateMVBuffer</span><span class="w"></span>
<span class="n">nvEncDestroyMVBuffer</span><span class="w"></span>
<span class="n">nvEncRunMotionEstimationOnly</span><span class="w"></span>
<span class="n">nvEncGetLastErrorString</span><span class="w"></span>
<span class="n">nvEncSetIOCudaStreams</span><span class="w"></span>
<span class="n">nvEncGetSequenceParamEx</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="nv-decoder-api-functions-traced-by-default">
<h3>NV Decoder API Functions Traced by Default<a class="headerlink" href="#nv-decoder-api-functions-traced-by-default" title="Permalink to this heading"></a></h3>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">cuvidCreateVideoSource</span><span class="w"></span>
<span class="n">cuvidCreateVideoSourceW</span><span class="w"></span>
<span class="n">cuvidDestroyVideoSource</span><span class="w"></span>
<span class="n">cuvidSetVideoSourceState</span><span class="w"></span>
<span class="n">cudaVideoState</span><span class="w"></span>
<span class="n">cuvidGetSourceVideoFormat</span><span class="w"></span>
<span class="n">cuvidGetSourceAudioFormat</span><span class="w"></span>
<span class="n">cuvidCreateVideoParser</span><span class="w"></span>
<span class="n">cuvidParseVideoData</span><span class="w"></span>
<span class="n">cuvidDestroyVideoParser</span><span class="w"></span>
<span class="n">cuvidCreateDecoder</span><span class="w"></span>
<span class="n">cuvidDestroyDecoder</span><span class="w"></span>
<span class="n">cuvidDecodePicture</span><span class="w"></span>
<span class="n">cuvidGetDecodeStatus</span><span class="w"></span>
<span class="n">cuvidReconfigureDecoder</span><span class="w"></span>
<span class="n">cuvidMapVideoFrame</span><span class="w"></span>
<span class="n">cuvidUnmapVideoFrame</span><span class="w"></span>
<span class="n">cuvidMapVideoFrame64</span><span class="w"></span>
<span class="n">cuvidUnmapVideoFrame64</span><span class="w"></span>
<span class="n">cuvidCtxLockCreate</span><span class="w"></span>
<span class="n">cuvidCtxLockDestroy</span><span class="w"></span>
<span class="n">cuvidCtxLock</span><span class="w"></span>
<span class="n">cuvidCtxUnlock</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="nv-jpeg-api-functions-traced-by-default">
<h3>NV JPEG API Functions Traced by Default<a class="headerlink" href="#nv-jpeg-api-functions-traced-by-default" title="Permalink to this heading"></a></h3>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nvjpegBufferDeviceCreate</span><span class="w"></span>
<span class="n">nvjpegBufferDeviceDestroy</span><span class="w"></span>
<span class="n">nvjpegBufferDeviceRetrieve</span><span class="w"></span>
<span class="n">nvjpegBufferPinnedCreate</span><span class="w"></span>
<span class="n">nvjpegBufferPinnedDestroy</span><span class="w"></span>
<span class="n">nvjpegBufferPinnedRetrieve</span><span class="w"></span>
<span class="n">nvjpegCreate</span><span class="w"></span>
<span class="n">nvjpegCreateEx</span><span class="w"></span>
<span class="n">nvjpegCreateSimple</span><span class="w"></span>
<span class="n">nvjpegDecode</span><span class="w"></span>
<span class="n">nvjpegDecodeBatched</span><span class="w"></span>
<span class="n">nvjpegDecodeBatchedEx</span><span class="w"></span>
<span class="n">nvjpegDecodeBatchedInitialize</span><span class="w"></span>
<span class="n">nvjpegDecodeBatchedPreAllocate</span><span class="w"></span>
<span class="n">nvjpegDecodeBatchedSupported</span><span class="w"></span>
<span class="n">nvjpegDecodeBatchedSupportedEx</span><span class="w"></span>
<span class="n">nvjpegDecodeJpeg</span><span class="w"></span>
<span class="n">nvjpegDecodeJpegDevice</span><span class="w"></span>
<span class="n">nvjpegDecodeJpegHost</span><span class="w"></span>
<span class="n">nvjpegDecodeJpegTransferToDevice</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsCreate</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsDestroy</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsSetAllowCMYK</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsSetOutputFormat</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsSetROI</span><span class="w"></span>
<span class="n">nvjpegDecodeParamsSetScaleFactor</span><span class="w"></span>
<span class="n">nvjpegDecoderCreate</span><span class="w"></span>
<span class="n">nvjpegDecoderDestroy</span><span class="w"></span>
<span class="n">nvjpegDecoderJpegSupported</span><span class="w"></span>
<span class="n">nvjpegDecoderStateCreate</span><span class="w"></span>
<span class="n">nvjpegDestroy</span><span class="w"></span>
<span class="n">nvjpegEncodeGetBufferSize</span><span class="w"></span>
<span class="n">nvjpegEncodeImage</span><span class="w"></span>
<span class="n">nvjpegEncodeRetrieveBitstream</span><span class="w"></span>
<span class="n">nvjpegEncodeRetrieveBitstreamDevice</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsCopyHuffmanTables</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsCopyMetadata</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsCopyQuantizationTables</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsCreate</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsDestroy</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsSetEncoding</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsSetOptimizedHuffman</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsSetQuality</span><span class="w"></span>
<span class="n">nvjpegEncoderParamsSetSamplingFactors</span><span class="w"></span>
<span class="n">nvjpegEncoderStateCreate</span><span class="w"></span>
<span class="n">nvjpegEncoderStateDestroy</span><span class="w"></span>
<span class="n">nvjpegEncodeYUV</span><span class="p">,(</span><span class="n">nvjpegHandle_t</span><span class="w"> </span><span class="n">handle</span><span class="w"></span>
<span class="n">nvjpegGetCudartProperty</span><span class="w"></span>
<span class="n">nvjpegGetDeviceMemoryPadding</span><span class="w"></span>
<span class="n">nvjpegGetImageInfo</span><span class="w"></span>
<span class="n">nvjpegGetPinnedMemoryPadding</span><span class="w"></span>
<span class="n">nvjpegGetProperty</span><span class="w"></span>
<span class="n">nvjpegJpegStateCreate</span><span class="w"></span>
<span class="n">nvjpegJpegStateDestroy</span><span class="w"></span>
<span class="n">nvjpegJpegStreamCreate</span><span class="w"></span>
<span class="n">nvjpegJpegStreamDestroy</span><span class="w"></span>
<span class="n">nvjpegJpegStreamGetChromaSubsampling</span><span class="w"></span>
<span class="n">nvjpegJpegStreamGetComponentDimensions</span><span class="w"></span>
<span class="n">nvjpegJpegStreamGetComponentsNum</span><span class="w"></span>
<span class="n">nvjpegJpegStreamGetFrameDimensions</span><span class="w"></span>
<span class="n">nvjpegJpegStreamGetJpegEncoding</span><span class="w"></span>
<span class="n">nvjpegJpegStreamParse</span><span class="w"></span>
<span class="n">nvjpegJpegStreamParseHeader</span><span class="w"></span>
<span class="n">nvjpegSetDeviceMemoryPadding</span><span class="w"></span>
<span class="n">nvjpegSetPinnedMemoryPadding</span><span class="w"></span>
<span class="n">nvjpegStateAttachDeviceBuffer</span><span class="w"></span>
<span class="n">nvjpegStateAttachPinnedBuffer</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
<section id="network-communication-profiling">
<h2>Network Communication Profiling<a class="headerlink" href="#network-communication-profiling" title="Permalink to this heading"></a></h2>
<p>Nsight Systems can be used to profiles several popular network communication protocols. To enable this, please select the <strong>Communication profiling options</strong> dropdown.</p>
<img alt="Project settings screen" class="image" src="../_images/project-settings-all.png" />
<p>Then select the libraries you would like to trace:</p>
<img alt="Communication library selection screen" class="image" src="../_images/project-settings-communication.png" />
<section id="mpi-api-trace">
<h3>MPI API Trace<a class="headerlink" href="#mpi-api-trace" title="Permalink to this heading"></a></h3>
<p>For Linux x86_64, ARM and Power targets, Nsight Systems is capable of capturing information about the MPI APIs executed in the profiled process. It has built-in API trace support for Open MPI and MPICH based MPI implementations.</p>
<img alt="MPI trace selection" class="image" src="../_images/mpi-project.png" />
<p>Only a subset of the MPI API, including blocking and non-blocking point-to-point and collective communication, and file I/O operations, is traced. If you require more control over the list of traced APIs or if you are using a different MPI implementation, you can use the <a class="reference external" href="https://github.com/NVIDIA/cuda-profiler/tree/master/nvtx_pmpi_wrappers">NVTX wrappers for MPI</a>. If you set the environment variable <code class="docutils literal notranslate"><span class="pre">LD_PRELOAD</span></code> to the path of generated wrapper library, Nsight Systems will capture and report the MPI API trace information when NVTX tracing is enabled. Choose an NVTX domain name other than “MPI”, since it is filtered out by Nsight Systems when MPI tracing is not enabled.</p>
<img alt="MPI API trace" class="image" src="../_images/mpi-trace.png" />
<p><strong>MPI Communication Parameters</strong></p>
<p>Nsight Systems can get additional information about MPI communication parameters. Currently, the parameters are only visible in the mouseover tooltips or in the eventlog. This means that the data is only available via the GUI. Future versions of the tool will export this information into the SQLite data files for postrun analysis.</p>
<p>In order to fully interpret MPI communications, data for all ranks associated with a communication operation must be loaded into Nsight Systems.</p>
<p>Here is an example of MPI_COMM_WORLD data. This does not require any additional team data, since local rank is the same as global rank.</p>
<p>(Screenshot shows communication parameters for an MPI_Bcast call on rank 3)</p>
<img alt="MPI communication parameter trace" class="image" src="../_images/mpi-comparam1.png" />
<p>When not all processes that are involved in an MPI communication are loaded into Nsight Systems the following information is available.</p>
<ul class="simple">
<li><p>Right-hand screenshot shows a reused communicator handle (last number increased).</p></li>
<li><p>Encoding: MPI_COMM[*team size*]*global-group-root-rank*.*group-ID*</p></li>
</ul>
<img alt="MPI communication parameter trace" class="image" src="../_images/mpi-comparam2.png" />
<p>When all reports are loaded into Nsight Systems:</p>
<ul class="simple">
<li><p>World rank is shown in addition to group-local rank “(world rank X)”</p></li>
<li><p>Encoding: MPI_COMM[*team size*]{rank0, rank1, …}</p></li>
<li><p>At most 8 ranks are shown (the numbers represent world ranks, the position in the list is the group-local rank)</p></li>
</ul>
<img alt="MPI communication parameter trace" class="image" src="../_images/mpi-comparam3.png" />
<p><strong>MPI functions traced:</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">MPI_Init</span><span class="p">[</span><span class="n">_thread</span><span class="p">],</span><span class="w"> </span><span class="n">MPI_Finalize</span><span class="w"></span>
<span class="n">MPI_Send</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_</span><span class="p">{</span><span class="n">B</span><span class="p">,</span><span class="n">S</span><span class="p">,</span><span class="n">R</span><span class="p">}</span><span class="n">send</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Recv</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Mrecv</span><span class="w"></span>
<span class="n">MPI_Sendrecv</span><span class="p">[</span><span class="n">_replace</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Barrier</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Bcast</span><span class="w"></span>
<span class="n">MPI_Scatter</span><span class="p">[</span><span class="n">v</span><span class="p">],</span><span class="w"> </span><span class="n">MPI_Gather</span><span class="p">[</span><span class="n">v</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Allgather</span><span class="p">[</span><span class="n">v</span><span class="p">],</span><span class="w"> </span><span class="n">MPI_Alltoall</span><span class="p">[{</span><span class="n">v</span><span class="p">,</span><span class="n">w</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_Allreduce</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Reduce</span><span class="p">[</span><span class="n">_</span><span class="p">{</span><span class="n">scatter</span><span class="p">,</span><span class="n">scatter_block</span><span class="p">,</span><span class="n">local</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_Scan</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Exscan</span><span class="w"></span>
<span class="n">MPI_Isend</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_I</span><span class="p">{</span><span class="n">b</span><span class="p">,</span><span class="n">s</span><span class="p">,</span><span class="n">r</span><span class="p">}</span><span class="n">send</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_I</span><span class="p">[</span><span class="n">m</span><span class="p">]</span><span class="n">recv</span><span class="w"></span>
<span class="n">MPI_</span><span class="p">{</span><span class="n">Send</span><span class="p">,</span><span class="n">Bsend</span><span class="p">,</span><span class="n">Ssend</span><span class="p">,</span><span class="n">Rsend</span><span class="p">,</span><span class="n">Recv</span><span class="p">}</span><span class="n">_init</span><span class="w"></span>
<span class="n">MPI_Start</span><span class="p">[</span><span class="n">all</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Ibarrier</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Ibcast</span><span class="w"></span>
<span class="n">MPI_Iscatter</span><span class="p">[</span><span class="n">v</span><span class="p">],</span><span class="w"> </span><span class="n">MPI_Igather</span><span class="p">[</span><span class="n">v</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Iallgather</span><span class="p">[</span><span class="n">v</span><span class="p">],</span><span class="w"> </span><span class="n">MPI_Ialltoall</span><span class="p">[{</span><span class="n">v</span><span class="p">,</span><span class="n">w</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_Iallreduce</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Ireduce</span><span class="p">[{</span><span class="n">scatter</span><span class="p">,</span><span class="n">scatter_block</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_I</span><span class="p">[</span><span class="n">ex</span><span class="p">]</span><span class="n">scan</span><span class="w"></span>
<span class="n">MPI_Wait</span><span class="p">[{</span><span class="n">all</span><span class="p">,</span><span class="n">any</span><span class="p">,</span><span class="n">some</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_Put</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Rput</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Get</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Rget</span><span class="w"></span>
<span class="n">MPI_Accumulate</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Raccumulate</span><span class="w"></span>
<span class="n">MPI_Get_accumulate</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Rget_accumulate</span><span class="w"></span>
<span class="n">MPI_Fetch_and_op</span><span class="p">,</span><span class="w"> </span><span class="n">MPI_Compare_and_swap</span><span class="w"></span>
<span class="n">MPI_Win_allocate</span><span class="p">[</span><span class="n">_shared</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Win_create</span><span class="p">[</span><span class="n">_dynamic</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Win_</span><span class="p">{</span><span class="n">attach</span><span class="p">,</span><span class="w"> </span><span class="n">detach</span><span class="p">}</span><span class="w"></span>
<span class="n">MPI_Win_free</span><span class="w"></span>
<span class="n">MPI_Win_fence</span><span class="w"></span>
<span class="n">MPI_Win_</span><span class="p">{</span><span class="n">start</span><span class="p">,</span><span class="w"> </span><span class="n">complete</span><span class="p">,</span><span class="w"> </span><span class="n">post</span><span class="p">,</span><span class="w"> </span><span class="n">wait</span><span class="p">}</span><span class="w"></span>
<span class="n">MPI_Win_</span><span class="p">[</span><span class="n">un</span><span class="p">]</span><span class="n">lock</span><span class="p">[</span><span class="n">_all</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Win_flush</span><span class="p">[</span><span class="n">_local</span><span class="p">][</span><span class="n">_all</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Win_sync</span><span class="w"></span>
<span class="n">MPI_File_</span><span class="p">{</span><span class="n">open</span><span class="p">,</span><span class="n">close</span><span class="p">,</span><span class="k">delete</span><span class="p">,</span><span class="n">sync</span><span class="p">}</span><span class="w"></span>
<span class="n">MPI_File_</span><span class="p">{</span><span class="n">read</span><span class="p">,</span><span class="n">write</span><span class="p">}[</span><span class="n">_</span><span class="p">{</span><span class="n">all</span><span class="p">,</span><span class="n">all_begin</span><span class="p">,</span><span class="n">all_end</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_File_</span><span class="p">{</span><span class="n">read</span><span class="p">,</span><span class="n">write</span><span class="p">}</span><span class="n">_at</span><span class="p">[</span><span class="n">_</span><span class="p">{</span><span class="n">all</span><span class="p">,</span><span class="n">all_begin</span><span class="p">,</span><span class="n">all_end</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_File_</span><span class="p">{</span><span class="n">read</span><span class="p">,</span><span class="n">write</span><span class="p">}</span><span class="n">_shared</span><span class="w"></span>
<span class="n">MPI_File_</span><span class="p">{</span><span class="n">read</span><span class="p">,</span><span class="n">write</span><span class="p">}</span><span class="n">_ordered</span><span class="p">[</span><span class="n">_</span><span class="p">{</span><span class="n">begin</span><span class="p">,</span><span class="n">end</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_File_i</span><span class="p">{</span><span class="n">read</span><span class="p">,</span><span class="n">write</span><span class="p">}[</span><span class="n">_</span><span class="p">{</span><span class="n">all</span><span class="p">,</span><span class="n">at</span><span class="p">,</span><span class="n">at_all</span><span class="p">,</span><span class="n">shared</span><span class="p">}]</span><span class="w"></span>
<span class="n">MPI_File_set_</span><span class="p">{</span><span class="n">size</span><span class="p">,</span><span class="n">view</span><span class="p">,</span><span class="n">info</span><span class="p">}</span><span class="w"></span>
<span class="n">MPI_File_get_</span><span class="p">{</span><span class="n">size</span><span class="p">,</span><span class="n">view</span><span class="p">,</span><span class="n">info</span><span class="p">,</span><span class="n">group</span><span class="p">,</span><span class="n">amode</span><span class="p">}</span><span class="w"></span>
<span class="n">MPI_File_preallocate</span><span class="w"></span>
<span class="n">MPI_Pack</span><span class="p">[</span><span class="n">_external</span><span class="p">]</span><span class="w"></span>
<span class="n">MPI_Unpack</span><span class="p">[</span><span class="n">_external</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="openshmem-library-trace">
<h3>OpenSHMEM Library Trace<a class="headerlink" href="#openshmem-library-trace" title="Permalink to this heading"></a></h3>
<p>If OpenSHMEM library trace is selected Nsight Systems will trace the subset of OpenSHMEM API functions that are most likely be involved in performance bottlenecks. To keep overhead low Nsight Systems does not trace all functions.</p>
<p><strong>OpenSHMEM 1.5 Functions Not Traced</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">shmem_my_pe</span><span class="w"></span>
<span class="n">shmem_n_pes</span><span class="w"></span>
<span class="n">shmem_global_exit</span><span class="w"></span>
<span class="n">shmem_pe_accessible</span><span class="w"></span>
<span class="n">shmem_addr_accessible</span><span class="w"></span>
<span class="n">shmem_ctx_</span><span class="p">{</span><span class="n">create</span><span class="p">,</span><span class="n">destroy</span><span class="p">,</span><span class="n">get_team</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_global_exit</span><span class="w"></span>
<span class="n">shmem_info_get_</span><span class="p">{</span><span class="n">version</span><span class="p">,</span><span class="n">name</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_</span><span class="p">{</span><span class="n">my_pe</span><span class="p">,</span><span class="n">n_pes</span><span class="p">,</span><span class="n">pe_accessible</span><span class="p">,</span><span class="n">ptr</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_query_thread</span><span class="w"></span>
<span class="n">shmem_team_</span><span class="p">{</span><span class="n">create_ctx</span><span class="p">,</span><span class="n">destroy</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_team_get_config</span><span class="w"></span>
<span class="n">shmem_team_</span><span class="p">{</span><span class="n">my_pe</span><span class="p">,</span><span class="n">n_pes</span><span class="p">,</span><span class="n">translate_pe</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_team_split_</span><span class="p">{</span><span class="mi">2</span><span class="n">d</span><span class="p">,</span><span class="n">strided</span><span class="p">}</span><span class="w"></span>
<span class="n">shmem_test</span><span class="o">*</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="ucx-library-trace">
<h3>UCX Library Trace<a class="headerlink" href="#ucx-library-trace" title="Permalink to this heading"></a></h3>
<p>If UCX library trace is selected Nsight Systems will trace the subset of functions of the UCX protocol layer UCP that are most likely be involved in performance bottlenecks. To keep overhead low Nsight Systems does not trace all functions.</p>
<p><strong>UCX functions traced:</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ucp_am_send_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_am_recv_data_nbx</span><span class="w"></span>
<span class="n">ucp_am_data_release</span><span class="w"></span>
<span class="n">ucp_atomic_</span><span class="p">{</span><span class="n">add</span><span class="p">{</span><span class="mi">32</span><span class="p">,</span><span class="mi">64</span><span class="p">},</span><span class="n">cswap</span><span class="p">{</span><span class="mi">32</span><span class="p">,</span><span class="mi">64</span><span class="p">},</span><span class="n">fadd</span><span class="p">{</span><span class="mi">32</span><span class="p">,</span><span class="mi">64</span><span class="p">},</span><span class="n">swap</span><span class="p">{</span><span class="mi">32</span><span class="p">,</span><span class="mi">64</span><span class="p">}}</span><span class="w"></span>
<span class="n">ucp_atomic_</span><span class="p">{</span><span class="n">post</span><span class="p">,</span><span class="n">fetch_nb</span><span class="p">,</span><span class="n">op_nbx</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_cleanup</span><span class="w"></span>
<span class="n">ucp_config_</span><span class="p">{</span><span class="n">modify</span><span class="p">,</span><span class="n">read</span><span class="p">,</span><span class="n">release</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_disconnect_nb</span><span class="w"></span>
<span class="n">ucp_dt_</span><span class="p">{</span><span class="n">create_generic</span><span class="p">,</span><span class="n">destroy</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_ep_</span><span class="p">{</span><span class="n">create</span><span class="p">,</span><span class="n">destroy</span><span class="p">,</span><span class="n">modify_nb</span><span class="p">,</span><span class="n">close_nbx</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_ep_flush</span><span class="p">[{</span><span class="n">_nb</span><span class="p">,</span><span class="n">_nbx</span><span class="p">}]</span><span class="w"></span>
<span class="n">ucp_listener_</span><span class="p">{</span><span class="n">create</span><span class="p">,</span><span class="n">destroy</span><span class="p">,</span><span class="n">query</span><span class="p">,</span><span class="n">reject</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_mem_</span><span class="p">{</span><span class="n">advise</span><span class="p">,</span><span class="n">map</span><span class="p">,</span><span class="n">unmap</span><span class="p">,</span><span class="n">query</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_</span><span class="p">{</span><span class="n">put</span><span class="p">,</span><span class="n">get</span><span class="p">}[</span><span class="n">_nbi</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_</span><span class="p">{</span><span class="n">put</span><span class="p">,</span><span class="n">get</span><span class="p">}</span><span class="n">_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_request_</span><span class="p">{</span><span class="n">alloc</span><span class="p">,</span><span class="n">cancel</span><span class="p">,</span><span class="n">is_completed</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_rkey_</span><span class="p">{</span><span class="n">buffer_release</span><span class="p">,</span><span class="n">destroy</span><span class="p">,</span><span class="n">pack</span><span class="p">,</span><span class="n">ptr</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_stream_data_release</span><span class="w"></span>
<span class="n">ucp_stream_recv_data_nb</span><span class="w"></span>
<span class="n">ucp_stream_</span><span class="p">{</span><span class="n">send</span><span class="p">,</span><span class="n">recv</span><span class="p">}</span><span class="n">_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_stream_worker_poll</span><span class="w"></span>
<span class="n">ucp_tag_msg_recv_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_tag_</span><span class="p">{</span><span class="n">send</span><span class="p">,</span><span class="n">recv</span><span class="p">}</span><span class="n">_nbr</span><span class="w"></span>
<span class="n">ucp_tag_</span><span class="p">{</span><span class="n">send</span><span class="p">,</span><span class="n">recv</span><span class="p">}</span><span class="n">_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_tag_send_sync_nb</span><span class="p">[</span><span class="n">x</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_worker_</span><span class="p">{</span><span class="n">create</span><span class="p">,</span><span class="n">destroy</span><span class="p">,</span><span class="n">get_address</span><span class="p">,</span><span class="n">get_efd</span><span class="p">,</span><span class="n">arm</span><span class="p">,</span><span class="n">fence</span><span class="p">,</span><span class="n">wait</span><span class="p">,</span><span class="n">signal</span><span class="p">,</span><span class="n">wait_mem</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_worker_flush</span><span class="p">[{</span><span class="n">_nb</span><span class="p">,</span><span class="n">_nbx</span><span class="p">}]</span><span class="w"></span>
<span class="n">ucp_worker_set_am_</span><span class="p">{</span><span class="n">handler</span><span class="p">,</span><span class="n">recv_handler</span><span class="p">}</span><span class="w"></span>
</pre></div>
</div>
<p><strong>UCX Functions Not Traced:</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ucp_config_print</span><span class="w"></span>
<span class="n">ucp_conn_request_query</span><span class="w"></span>
<span class="n">ucp_context_</span><span class="p">{</span><span class="n">query</span><span class="p">,</span><span class="n">print_info</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_get_version</span><span class="p">[</span><span class="n">_string</span><span class="p">]</span><span class="w"></span>
<span class="n">ucp_ep_</span><span class="p">{</span><span class="n">close_nb</span><span class="p">,</span><span class="n">print_info</span><span class="p">,</span><span class="n">query</span><span class="p">,</span><span class="n">rkey_unpack</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_mem_print_info</span><span class="w"></span>
<span class="n">ucp_request_</span><span class="p">{</span><span class="n">check_status</span><span class="p">,</span><span class="n">free</span><span class="p">,</span><span class="n">query</span><span class="p">,</span><span class="n">release</span><span class="p">,</span><span class="n">test</span><span class="p">}</span><span class="w"></span>
<span class="n">ucp_stream_recv_request_test</span><span class="w"></span>
<span class="n">ucp_tag_probe_nb</span><span class="w"></span>
<span class="n">ucp_tag_recv_request_test</span><span class="w"></span>
<span class="n">ucp_worker_</span><span class="p">{</span><span class="n">address_query</span><span class="p">,</span><span class="n">print_info</span><span class="p">,</span><span class="n">progress</span><span class="p">,</span><span class="n">query</span><span class="p">,</span><span class="n">release_address</span><span class="p">}</span><span class="w"></span>
</pre></div>
</div>
<p>Additional API functions from other UCX layers may be added in a future version of the product.</p>
</section>
<section id="nvidia-nvshmem-and-nccl-trace">
<h3>NVIDIA NVSHMEM and NCCL Trace<a class="headerlink" href="#nvidia-nvshmem-and-nccl-trace" title="Permalink to this heading"></a></h3>
<p>The NVIDIA network communication libraries NVSHMEM and NCCL have been instrumented using NVTX annotations. To enable tracing these libraries in Nsight Systems, turn on NVTX tracing in the GUI or CLI. To enable the NVTX instrumentation of the NVSHMEM library, make sure that the environment variable <code class="docutils literal notranslate"><span class="pre">NVSHMEM_NVTX</span></code> is set properly, e.g. <code class="docutils literal notranslate"><span class="pre">NVSHMEM_NVTX=common</span></code>.</p>
</section>
<section id="nic-metric-sampling">
<h3>NIC Metric Sampling<a class="headerlink" href="#nic-metric-sampling" title="Permalink to this heading"></a></h3>
<p><strong>Overview</strong></p>
<p>NVIDIA ConnectX smart network interface cards (smart NICs) offer advanced hardware offloads and accelerations for network operations. Viewing smart NICs metrics, on Nsight Systems timeline, enables developers to better understand their application’s network usage. Developers can use this information to optimize the application’s performance.</p>
<p><strong>Limitations/Requirements</strong></p>
<ul class="simple">
<li><p>NIC metric sampling supports NVIDIA ConnectX boards starting with ConnectX 5</p></li>
<li><p>NIC metric sampling is supported on Linux x86_64 and ARM Server (SBSA) machines only, having minimum Linux kernel 4.12 and minimum MLNX_OFED 4.1. You can download the latest and archived versions of the MLX_OFED driver from the <a class="reference external" href="https://network.nvidia.com/products/infiniband-drivers/linux/mlnx_ofed/">MLNX_OFED Download Center</a>. If collecting NIC metrics within a container, make sure that the container has access to the driver on the host machine. To check manually if OFED is installed and get its version you can run:</p>
<ul>
<li><p><code class="docutils literal notranslate"><span class="pre">/usr/bin/ofed_info</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">cat</span> <span class="pre">/sys/module/"$(cat</span> <span class="pre">/proc/modules</span> <span class="pre">|</span> <span class="pre">grep</span> <span class="pre">-o</span> <span class="pre">-E</span> <span class="pre">"^mlx._core")"/version</span></code></p></li>
</ul>
</li>
</ul>
<p>To check if the target system meets the requirements for NIC metrics collection you can run <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">status</span> <span class="pre">--network</span></code>.</p>
<p><strong>Collecting NIC Metrics Using the Command Line</strong></p>
<p>To collect NIC performance metrics, using Nsight Systems CLI, add the <code class="docutils literal notranslate"><span class="pre">--nic-metrics</span></code> command line switch:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">nic</span><span class="o">-</span><span class="n">metrics</span><span class="o">=</span><span class="nb">true</span><span class="w"> </span><span class="n">my_app</span><span class="w"></span>
</pre></div>
</div>
<img alt="NIC metric sampling screenshot" class="image" src="../_images/nic-metrics-sampling.png" />
<p><strong>Available Metrics</strong></p>
<ul class="simple">
<li><p><strong>Bytes sent</strong> - Number of bytes sent through all NIC ports.</p></li>
<li><p><strong>Bytes received</strong> - Number of bytes received by all NIC ports.</p></li>
<li><p><strong>CNPs sent</strong> - Number of congestion notification packets sent by the NIC.</p></li>
<li><p><strong>CNPs received</strong> - Number of congestion notification packets received and handled by the NIC.</p></li>
<li><p><strong>Send waits</strong> - The number of ticks during which ports had data to transmit but no data was sent during the entire tick (either because of insufficient credits or because of lack of arbitration)</p></li>
</ul>
<p>Note: Each one of the mentioned metrics is shown only if it has non-zero value during profiling.</p>
<p><strong>Usage Examples</strong></p>
<ul class="simple">
<li><p>The <code class="docutils literal notranslate"><span class="pre">Bytes</span> <span class="pre">sent/sec</span></code> and the <code class="docutils literal notranslate"><span class="pre">Bytes</span> <span class="pre">received/sec</span></code> metrics enables identifying idle and busy NIC times.</p>
<ul>
<li><p>Developers may shift network operations from busy to idle times to reduce network congestion and latency.</p></li>
<li><p>Developers can use idle NIC times to send additional data without reducing application performance.</p></li>
</ul>
</li>
<li><p>CNPs (congestion notification packets) received/sent and Send waits metrics may explain network latencies. A developer seeing the time periods when the network was congested may rewrite his algorithm to avoid the observed congestions.</p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>RDMA over Converged Ethernet (RoCE) traffic is not logged into the Nsight Systems NIC metrics.</p>
</div>
</section>
<section id="infiniband-switch-metric-sampling">
<h3>InfiniBand Switch Metric Sampling<a class="headerlink" href="#infiniband-switch-metric-sampling" title="Permalink to this heading"></a></h3>
<p>NVIDIA Quantum InfiniBand switches offer high-bandwidth, low-latency communication. Viewing switch metrics, on Nsight Systems timeline, enables developers to better understand their application’s network usage. Developers can use this information to optimize the application’s performance.</p>
<p><strong>Limitations/Requirements</strong></p>
<p>IB switch metric sampling supports all NVIDIA Quantum switches. The user needs to have permission to query the InfiniBand switch metrics.</p>
<p>To check if the current user has permissions to query the InfiniBand switch metrics, check that the user have permission to access /dev/umad</p>
<p>To give user permissions to query InfiniBand switch metrics on RedHat systems, follow the directions at <a class="reference external" href="https://access.redhat.com/solutions/5929621">RedHat Solutions</a>.</p>
<p>To collect InfiniBand switch performance metric, using Nsight Systems CLI, add the <code class="docutils literal notranslate"><span class="pre">--ib-switch-metrics-device</span></code> command line switch, followed by a comma separated list of InfiniBand switch GUIDs. For example:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">--</span><span class="n">ib</span><span class="o">-</span><span class="k">switch</span><span class="o">-</span><span class="n">metrics</span><span class="o">-</span><span class="n">device</span><span class="o">=<</span><span class="n">IB</span><span class="w"> </span><span class="k">switch</span><span class="w"> </span><span class="n">GUID</span><span class="o">></span><span class="w"> </span><span class="n">my_app</span><span class="w"></span>
</pre></div>
</div>
<p>To get a list of InfiniBand switches, reachable by a given NIC, use:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="n">ibswitches</span><span class="w"> </span><span class="o">-</span><span class="n">C</span><span class="w"> </span><span class="o"><</span><span class="n">nic</span><span class="w"> </span><span class="n">name</span><span class="o">></span><span class="w"></span>
</pre></div>
</div>
<img alt="InfiniBand Switch performance metrics sampling screenshot" class="image" src="../_images/ib-switch-metrics-sampling.png" />
<p><strong>Available Metrics</strong></p>
<ul class="simple">
<li><p><strong>Bytes sent</strong> - Number of bytes sent through all switch ports</p></li>
<li><p><strong>Bytes received</strong> - Number of bytes received by all switch ports</p></li>
<li><p><strong>Send waits</strong> - The number of ticks during which switch ports, selected by PortSelect, had data to transmit but no data was sent during the entire tick (either because of insufficient credits or of lack of arbitration)</p></li>
<li><p><strong>Average sent packet size</strong> - Average sent InfiniBand packet size</p></li>
<li><p><strong>Average received packet size</strong> - Average received InfiniBand packet size</p></li>
</ul>
</section>
</section>
<section id="python-profiling">
<h2>Python Profiling<a class="headerlink" href="#python-profiling" title="Permalink to this heading"></a></h2>
<section id="python-backtrace-sampling">
<h3>Python Backtrace Sampling<a class="headerlink" href="#python-backtrace-sampling" title="Permalink to this heading"></a></h3>
<p>Nsight Systems for Arm server (SBSA) platforms, x86 Linux and Windows targets, is capable of periodically capturing Python backtrace information. This functionality is available when tracing Python interpreters of version 3.9 or later. Capturing python backtrace is done in periodic samples, in a selected frequency ranging from 1Hz - 2KHz with a default value of 1KHz. Note that this feature provides meaningful backtraces for Python processes, when profiling Python-only workflows, consider disabling the CPU sampling option to reduce overhead.</p>
<p>To enable Python backtrace sampling from Nsight Systems:</p>
<p><strong>CLI</strong> — Set <code class="docutils literal notranslate"><span class="pre">--python-sampling=true</span></code> and use the <code class="docutils literal notranslate"><span class="pre">--python-sampling-frequency</span></code> option to set the sampling rate.</p>
<p><strong>GUI</strong> — Select the <strong>Collect Python backtrace samples</strong> checkbox.</p>
<blockquote>
<div><img alt="Configure Python Backtrace" class="image" src="../_images/configure-python-backtrace-sampling.png" />
</div></blockquote>
<p>Example screenshot:</p>
<img alt="Python Backtrace sampling" class="image" src="../_images/python-backtrace-sampling.png" />
</section>
<section id="python-nvtx-annotations">
<h3>Python NVTX Annotations<a class="headerlink" href="#python-nvtx-annotations" title="Permalink to this heading"></a></h3>
<p>Nsight Systems for Arm server (SBSA) platforms, x86 Linux and Windows targets, is capable of using NVTX to annotate Python functions.</p>
<p>The Python source code does not require any changes. This feature requires CPython interpreter, release 3.8 or later.</p>
<p>The annotations are configured in a JSON file. An example file is located in Nsight Systems installation folder in <code class="docutils literal notranslate"><span class="pre"><target-platform-folder>/PythonNvtx/annotations.json</span></code>.</p>
<p><strong>Notes:</strong></p>
<ul class="simple">
<li><p>Annotating function from module <code class="docutils literal notranslate"><span class="pre">__main__</span></code> is not supported.</p></li>
</ul>
<p>To enable Python NVTX annotations from Nsight Systems:</p>
<p><strong>CLI</strong> — Set <code class="docutils literal notranslate"><span class="pre">--python-nvtx-annotations=<json_file></span></code>.</p>
<p><strong>GUI</strong> — Select the <strong>Python NVTX annotations</strong> checkbox and specify the JSON file.</p>
<blockquote>
<div><img alt="Configure Python NVTX Annotations" class="image" src="../_images/configure-python-nvtx-annotations.png" />
</div></blockquote>
<p>Example screenshot:</p>
<img alt="Python NVTX Annotations" class="image" src="../_images/python-nvtx-annotations.png" />
</section>
<section id="python-gil-tracing">
<h3>Python GIL Tracing<a class="headerlink" href="#python-gil-tracing" title="Permalink to this heading"></a></h3>
<p>Nsight Systems for Arm server (SBSA) platforms, x86 Linux and Windows targets, is capable of tracing when Python threads are waiting to hold and holding the GIL (Global Interpreter Lock).</p>
<p>The Python source code does not require any changes. This feature requires CPython interpreter, release 3.9 or later.</p>
<p><strong>CLI</strong> — Set <code class="docutils literal notranslate"><span class="pre">--trace=python-gil</span></code>.</p>
<p><strong>GUI</strong> — Select the <strong>Trace GIL</strong> checkbox under <strong>Python profiling options</strong>.</p>
<blockquote>
<div><img alt="Configure Python GIL Tracing" class="image" src="../_images/configure-python-gil-tracing.png" />
</div></blockquote>
<p>Example screenshot:</p>
<img alt="Python GIL Tracing" class="image" src="../_images/python-gil-tracing.png" />
</section>
</section>
<section id="reading-your-report-in-gui">
<h2>Reading Your Report in GUI<a class="headerlink" href="#reading-your-report-in-gui" title="Permalink to this heading"></a></h2>
<section id="generating-a-new-report">
<h3>Generating a New Report<a class="headerlink" href="#generating-a-new-report" title="Permalink to this heading"></a></h3>
<p>Users can generate a new report by stopping a profiling session. If a profiling session has been canceled, a report will not be generated, and all collected data will be discarded.</p>
<p>A new <code class="docutils literal notranslate"><span class="pre">.nsys-rep</span></code> file will be created and put into the same directory as the project file (<code class="docutils literal notranslate"><span class="pre">.qdproj</span></code>).</p>
</section>
<section id="opening-an-existing-report">
<h3>Opening an Existing Report<a class="headerlink" href="#opening-an-existing-report" title="Permalink to this heading"></a></h3>
<p>An existing <code class="docutils literal notranslate"><span class="pre">.nsys-rep</span></code> file can be opened using <strong>File > Open…</strong>.</p>
</section>
<section id="sharing-a-report-file">
<h3>Sharing a Report File<a class="headerlink" href="#sharing-a-report-file" title="Permalink to this heading"></a></h3>
<p>Report files (<code class="docutils literal notranslate"><span class="pre">.nsys-rep</span></code>) are self-contained and can be shared with other users of Nsight Systems. The only requirement is that the same or newer version of Nsight Systems is always used to open report files.</p>
<p>Project files (<code class="docutils literal notranslate"><span class="pre">.qdproj</span></code>) are currently not shareable, since they contain full paths to the report files.</p>
<p>To quickly navigate to the directory containing the report file, right click on it in the Project Explorer, and choose <strong>Show in folder…</strong> in the context menu.</p>
</section>
<section id="report-tab">
<h3>Report Tab<a class="headerlink" href="#report-tab" title="Permalink to this heading"></a></h3>
<p>While generating a new report or loading an existing one, a new tab will be created. The most important parts of the report tab are:</p>
<ul>
<li><p><strong>View selector</strong> — Allows switching between <em>Analysis Summary</em>, <em>Timeline View</em>, <em>Diagnostics Summary</em>, and <em>Symbol Resolution Logs</em> views.</p>
<blockquote>
<div><img alt="Page selector" class="image" src="../_images/report-view-selector.png" />
</div></blockquote>
</li>
<li><p><strong>Timeline</strong> — This is where all charts are displayed.</p></li>
<li><p><strong>Function table</strong> — Located below the timeline, it displays statistical information about functions in the target application in multiple ways.</p></li>
</ul>
<p>Additionally, the following controls are available:</p>
<ul class="simple">
<li><p><strong>Zoom slider</strong> — Allows you to vertically zoom the charts on the timeline.</p></li>
</ul>
</section>
<section id="analysis-summary-view">
<h3>Analysis Summary View<a class="headerlink" href="#analysis-summary-view" title="Permalink to this heading"></a></h3>
<p>This view shows a summary of the profiling session. In particular, it is useful to review the project configuration used to generate this report. Information from this view can be selected and copied using the mouse cursor.</p>
</section>
<section id="timeline-view">
<h3>Timeline View<a class="headerlink" href="#timeline-view" title="Permalink to this heading"></a></h3>
<p>The timeline view consists of two main controls: the timeline at the top, and a bottom pane that contains the events view and the function table. In some cases, when sampling of a process has not been enabled, the function table might be empty and hidden.</p>
<p>The bottom view selector sets the view that is displayed in the bottom pane.</p>
<blockquote>
<div><img alt="Bottom view selection" class="image" src="../_images/bottom-view-selector.png" />
</div></blockquote>
<section id="timeline">
<h4>Timeline<a class="headerlink" href="#timeline" title="Permalink to this heading"></a></h4>
<p>Timeline is a versatile control that contains a tree-like <strong>hierarchy</strong> on the left, and corresponding <em>charts</em> on the right.</p>
<p>Contents of the hierarchy depend on the project settings used to collect the report. For example, if a certain feature has not been enabled, corresponding rows will not be show on the timeline.</p>
<p>To generate a timeline screenshot without opening the full GUI, use the command</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="p">.</span><span class="n">exe</span><span class="w"> </span><span class="o">--</span><span class="n">screenshot</span><span class="w"> </span><span class="n">filename</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="events-view">
<h4>Events View<a class="headerlink" href="#events-view" title="Permalink to this heading"></a></h4>
<p>The Events View provides a tabular display of the trace events. The view contents can be searched and sorted.</p>
<p>Double-clicking an item in the Events View automatically focuses the Timeline View on the corresponding timeline item.</p>
<p>API calls, GPU executions, and debug markers that occurred within the boundaries of a debug marker are displayed nested to that debug marker. Multiple levels of nesting are supported.</p>
<p>Events view recognizes these types of debug markers:</p>
<ul class="simple">
<li><p>NVTX</p></li>
<li><p>Vulkan VK_EXT_debug_marker markers, VK_EXT_debug_utils labels</p></li>
<li><p>PIX events and markers</p></li>
<li><p>OpenGL KHR_debug markers</p></li>
</ul>
<img alt="Events View nested debug markers" class="image" src="../_images/events-view-nested-markers.png" />
<p>You can copy and paste from the events view by highlighting rows, using <strong>Shift</strong> or <strong>Ctrl</strong> to enable multi-select. Right clicking on the selection will give you a copy option.</p>
<img alt="Events View copy selection" class="image" src="../_images/event-copy.png" />
<p>Pasting into text gives you a tab separated view:</p>
<img alt="Events View paste into notepad" class="image" src="../_images/event-paste-notepad.png" />
<p>Pasting into spreadsheet properly copies into rows and columns:</p>
<img alt="Events View paste into spreadsheet" class="image" src="../_images/event-paste-spreadsheet.png" />
</section>
<section id="function-table-modes">
<h4>Function Table Modes<a class="headerlink" href="#function-table-modes" title="Permalink to this heading"></a></h4>
<blockquote>
<div><img alt="Function table modes" class="image" src="../_images/function-table-modes.png" />
</div></blockquote>
<p>The function table can work in three modes:</p>
<ul class="simple">
<li><p><strong>Top-Down View</strong> — In this mode, expanding top-level functions provides information about the <em>callee</em> functions. One of the top-level functions is typically the main function of your application, or another entry point defined by the runtime libraries.</p></li>
<li><p><strong>Bottom-Up View</strong> — This is a reverse of the Top-Down view. On the top level, there are functions directly hit by the sampling profiler. To explore all possible call chains leading to these functions, you need to expand the subtrees of the top-level functions.</p></li>
<li><p><strong>Flat View</strong> — This view enumerates all functions ever observed by the profiler, even if they have never been directly hit, but just appeared somewhere on the call stack. This view typically provides a high-level overview of which parts of the code are CPU-intensive.</p></li>
</ul>
<p>Each of the views helps understand particular performance issues of the application being profiled. For example:</p>
<ul class="simple">
<li><p>When trying to find specific bottleneck functions that can be optimized, the Bottom-Up view should be used. Typically, the top few functions should be examined. Expand them to understand in which contexts they are being used.</p></li>
<li><p>To navigate the call tree of the application and while generally searching for algorithms and parts of the code that consume unexpectedly large amount of CPU time, the Top-Down view should be used.</p></li>
<li><p>To quickly assess which parts of the application, or high level parts of an algorithm, consume significant amount of CPU time, use the Flat view.</p></li>
</ul>
<p>The Top-Down and Bottom-Up views have <em>Self</em> and <em>Total</em> columns, while the Flat view has a <em>Flat</em> column. It is important to understand the meaning of each of the columns:</p>
<ul class="simple">
<li><p>Top-Down view</p>
<ul>
<li><p><strong>Self</strong> column denotes the relative amount of time spent executing instructions of this particular function.</p></li>
<li><p><strong>Total</strong> column shows how much time has been spent executing this function, including all other functions called from this one. Total values of sibling rows sum up to the Total value of the parent row, or 100% for the top-level rows.</p></li>
</ul>
</li>
<li><p>Bottom-Up view</p>
<ul>
<li><p><strong>Self</strong> column for <em>top-level rows</em>, as in the Top-Down view, shows how much time has been spent directly in this function. Self times of all top-level rows add up to 100%.</p></li>
<li><p><strong>Self</strong> column for <em>children rows</em> breaks down the value of the parent row based on the various call chains leading to that function. Self times of sibling rows add up to the value of the parent row.</p></li>
</ul>
</li>
<li><p>Flat view</p>
<ul>
<li><p><strong>Flat</strong> column shows how much time this function has been anywhere on the call stack. Values in this column do not add up or have other significant relationships.</p></li>
</ul>
</li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If low-impact functions have been filtered out, values may not add up correctly to 100%, or to the value of the parent row. This filtering can be disabled.</p>
</div>
<p>Contents of the symbols table is tightly related to the timeline. Users can apply and modify filters on the timeline, and they will affect which information is displayed in the symbols table:</p>
<ul class="simple">
<li><p><strong>Per-thread filtering</strong> — Each thread that has sampling information associated with it has a checkbox next to it on the timeline. Only threads with selected checkboxes are represented in the symbols table.</p></li>
<li><p><strong>Time filtering</strong> — A time filter can be setup on the timeline by pressing the left mouse button, dragging over a region of interest on the timeline, and then choosing <strong>Filter by selection</strong> in the dropdown menu. In this case, only sampling information collected during the selected time range will be used to build the symbols table.</p></li>
</ul>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>If too little sampling data is being used to build the symbols table (for example, when the sampling rate is configured to be low, and a short period of time is used for time-based filtering), the numbers in the symbols table might not be representative or accurate in some cases.</p>
</div>
</section>
<section id="function-table-notes">
<h4>Function Table Notes<a class="headerlink" href="#function-table-notes" title="Permalink to this heading"></a></h4>
<p><strong>Last Branch Records vs Frame Pointers</strong></p>
<p>Two of the mechanisms available for collecting backtraces are Intel Last Branch Records (LBRs) and frame pointers. LBRs are used to trace every branch instruction via a limited set of hardware registers. They can be configured to generate backtraces but have finite depth based on the CPU’s microarchitecture. LBRs are effectively free to collect but may not be as deep as you need in order to fully understand how the workload arrived a specific Instruction Pointer (IP).</p>
<p>Frame pointers only work when a binary is compiled with the <code class="docutils literal notranslate"><span class="pre">-fno-omit-frame-pointer</span></code> compiler switch. To determine if frame pointers are enabled on an x86_64 binary running on Linux, dump a binary’s assembly code using the <code class="docutils literal notranslate"><span class="pre">objdump</span> <span class="pre">-d</span> <span class="pre">[binary_file]</span></code> command and look for this pattern at the beginning of all functions;</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>push %rbp
mov %rsp,%rbp
</pre></div>
</div>
<p>When frame pointers are available in a binary, full stack traces will be captured. Note that libraries that are frequently used by apps and ship with the operating system, such as libc, are generated in release mode and therefore do not include frame pointers. Frequently, when a backtrace includes an address from a system library, the backtrace will fail to resolve further as the frame pointer trail goes cold due to a missing frame pointer.</p>
<p>A simple application was developed to show the difference. The application calls function a(), which calls b(), which calls c(), etc. Function z() calls a heavy compute function called matrix_multiply(). Almost all of the IP samples are collected while matrix_multiple is executing. The next two screen shots show one of the main differences between frame pointers and LBRs.</p>
<blockquote>
<div><img alt="frame pointer backtrace" class="image" src="../_images/frame-pointers.png" />
</div></blockquote>
<blockquote>
<div><img alt="lbr backtrace" class="image" src="../_images/lbr-backtrace.png" />
</div></blockquote>
<p>Note that the frame pointer example, shows the full stack trace while the LBR example, only shows part of the stack due to the limited number of LBR registers in the CPU.</p>
<p><strong>Kernel Samples</strong></p>
<p>When an IP sample is captured while a kernel mode (i.e. operating system) function is executing, the sample will be shown with an address that starts with 0xffffffff and map to the [kernel.kallsyms] module.</p>
<blockquote>
<div><img alt="kernel mode sample backtrace" class="image" src="../_images/kernel-mode-sample.png" />
</div></blockquote>
<p><strong>[vdso]</strong></p>
<p>Samples may be collected while a CPU is executing functions in the Virtual Dynamic Shared Object. In this case, the sample will be resolved (i.e. mapped) to the [vdso] module. The vdso <a class="reference external" href="http://man7.org/linux/man-pages/man7/vdso.7.html">man page</a> provides the following description of the vdso:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>The “vDSO“ (virtual dynamic shared object) is a small shared library
that the kernel automatically maps into the address space of all
user-space applications. Applications usually do not need to concern
themselves with these details as the vDSO is most commonly called by
the C library. This way you can code in the normal way using
standard functions and the C library will take care of using any
functionality that is available via the vDSO.
Why does the vDSO exist at all? There are some system calls the
kernel provides that user-space code ends up using frequently, to the
point that such calls can dominate overall performance. This is due
both to the frequency of the call as well as the context-switch
overhead that results from exiting user space and entering the
kernel.
</pre></div>
</div>
<p><strong>[Unknown]</strong></p>
<p>When an address can not be resolved (i.e. mapped to a module), its address within the process’ address space will be shown and its module will be marked as [Unknown].</p>
</section>
<section id="filter-dialog">
<h4>Filter Dialog<a class="headerlink" href="#filter-dialog" title="Permalink to this heading"></a></h4>
<blockquote>
<div><img alt="Filter dialog" class="image" src="../_images/report-filter-dialog.png" />
</div></blockquote>
<ul class="simple">
<li><p><strong>Collapse unresolved lines</strong> is useful if some of the binary code does not have symbols. In this case, subtrees that consist of only unresolved symbols get collapsed in the Top-Down view, since they provide very little useful information.</p></li>
<li><p><strong>Hide functions with CPU usage below X%</strong> is useful for large applications, where the sampling profiler hits lots of function just a few times. To filter out the “long tail,” which is typically not important for CPU performance bottleneck analysis, this checkbox should be selected.</p></li>
</ul>
</section>
<section id="example-of-using-timeline-with-function-table">
<h4>Example of Using Timeline with Function Table<a class="headerlink" href="#example-of-using-timeline-with-function-table" title="Permalink to this heading"></a></h4>
<p>Here is an example walkthrough of using the timeline and function table with Instruction Pointer (IP)/backtrace Sampling Data</p>
<p><strong>Timeline</strong></p>
<p>When a collection result is opened in the Nsight Systems GUI, there are multiple ways to view the CPU profiling data - especially the CPU IP / backtrace data.</p>
<img alt="Timeline showing CPU IP/backtrace information" class="image" src="../_images/cpu-ipbacktrace.png" />
<p>In the timeline, yellow-orange marks can be found under each thread’s timeline that indicate the moment an IP / backtrace sample was collected on that thread (e.g. see the yellow-orange marks in the Specific Samples box above). Hovering the cursor over a mark will cause a tooltip to display the backtrace for that sample.</p>
<p>Below the Timeline is a drop-down list with multiple options including Events View, Top-Down View, Bottom-Up View, and Flat View. All four of these views can be used to view CPU IP / backtrace sampling data.</p>
<p>If the Bottom-Up View is selected, here is the sampling summary shown in the bottom half of the Timeline View screen. Notice that the summary includes the phrase “65,022 samples are used” indicating how many samples are summarized. By default, functions that were found in less less than 0.5% of the samples are not show. Use the <code class="docutils literal notranslate"><span class="pre">filter</span></code> button to modify that setting.</p>
<img alt="Timeline showing CPU IP/backtrace information" class="image" src="../_images/ip-backtrace-sampling-filter.png" />
<p>When sampling data is filtered, the Sampling Summary will summarize the selected samples. Samples can be filtered on an OS thread basis, on a time basis, or both. Above, deselecting a checkbox next to a thread removes its samples from the sampling summary. Dragging the cursor over the timeline and selecting “Filter and Zoom In” chooses the samples during the time selected, as seen below. The sample summary includes the phrase “0.35% (225 samples) of data is shown due to applied filters” indicating that only 225 samples are included in the summary results.</p>
<img alt="Timeline showing CPU IP/backtrace information, filtered" class="image" src="../_images/ip-backtrace-filtered.png" />
<p>Deselecting threads one at a time by deselecting their checkbox can be tedious. Click on the down arrow next to a thread and choose Show Only This Thread to deselect all threads except that thread.</p>
<img alt="How to deselect all threads except one" class="image" src="../_images/deselect-thread.png" />
<p>If Events View is selected in the Timeline View’s drop-down list, right click on a specific thread and choose Show in Events View. The samples collected while that thread executed will be shown in the Events View. Double clicking on a specific sample in the Events view causes the timeline to show when that sample was collected - see the green boxes below. The backtrace for that sample is also shown in the Events View.</p>
<img alt="events view" class="image" src="../_images/events-view-click.png" />
<p><strong>Backtraces</strong></p>
<p>To understand the code path used to get to a specific function shown in the sampling summary, right click on a function and select Expand.</p>
<img alt="expand backtrace" class="image" src="../_images/expand-backtrace.png" />
<p>The above shows what happens when a function’s backtraces are expanded. In this case, the PCQueuePop function was called from the CmiGetNonLocal function which was called by the CsdNextMessage function which was called by the CsdScheduleForever function. The [Max depth] string marks the end of the collected backtrace.</p>
<img alt="zoom in expand backtrace" class="image" src="../_images/zoomed-expand-backtrace.png" />
<p>Note that, by default, backtraces with less than 0.5% of the total backtraces are hidden. This behavior can make the percentage results hard to understand. If all backtraces are shown (i.e. the filter is disabled), the results look very different and the numbers add up as expected. To disable the filter, click on the Filter… button and uncheck the Hide functions with CPU usage below X% checkbox.</p>
<img alt="no function filter backtraces" class="image" src="../_images/no-filter-backtraces.png" />
<p>When the filter is disabled, the backtraces are recalculated. Note that you may need to right click on the function and select Expand again to get all of the backtraces to be shown.</p>
<img alt="reset backtraces" class="image" src="../_images/reset-backtraces.png" />
<p>When backtraces are collected, the whole sample (IP and backtrace) is handled as a single sample. If two samples have the exact same IP and backtrace, they are summed in the final results. If two samples have the same IP but a different backtrace, they will be shown as having the same leaf (i.e. IP) but a different backtrace. As mentioned earlier, when backtraces end, they are marked with the [Max depth] string (unless the backtrace can be traced back to its origin - e.g. __libc_start_main) or the backtrace breaks because an IP cannot be resolved.</p>
<p>Above, the leaf function is PCQueuePop. In this case, there are 11 different backtraces that lead to PCQueuPop - all of them end with [Max depth]. For example, the dominant path is PCQueuPop<-CmiGetNonLocal<-CsdNextmessage<-CsdScheduleForever<-[Max depth]. This path accounts for 5.67% of all samples as shown in line 5 (red numbers). The second most dominant path is PCQueuPop<-CmiGetNonLocal<-[Max depth] which accounts for 0.44% of all samples as shown in line 24 (red numbers). The path PCQueuPop<-CmiGetNonLocal<-CsdNextmessage<-CsdScheduleForever<-Sequencer::integrate(int)<-[Max depth] accounts for 0.03% of the samples as shown in line 7 (red numbers). Adding up percentages shown in the [Max depth] lines (lines 5, 7, 9, 13, 15, 16, 17, 19, 21, 23, and 24) generates 7.04% which equals the percentage of samples associated with the PCQueuePop function shown in line 0 (red numbers).</p>
</section>
</section>
<section id="diagnostics-summary-view">
<h3>Diagnostics Summary View<a class="headerlink" href="#diagnostics-summary-view" title="Permalink to this heading"></a></h3>
<p>This view shows important messages. Some of them were generated during the profiling session, while some were added while processing and analyzing data in the report. Messages can be one of the following types:</p>
<ul class="simple">
<li><p>Informational messages</p></li>
<li><p>Warnings</p></li>
<li><p>Errors</p></li>
</ul>
<p>To draw attention to important diagnostics messages, a summary line is displayed on the timeline view in the top right corner:</p>
<blockquote>
<div><img alt="Diagnostics messages" class="image" src="../_images/report-diagnostics-link.png" />
</div></blockquote>
<p>Information from this view can be selected and copied using the mouse cursor.</p>
</section>
<section id="symbol-resolution-logs-view">
<h3>Symbol Resolution Logs View<a class="headerlink" href="#symbol-resolution-logs-view" title="Permalink to this heading"></a></h3>
<p>This view shows all messages related to the process of resolving symbols. It might be useful to debug issues when some of the symbol names in the symbols table of the timeline view are unresolved.</p>
</section>
</section>
<section id="adding-report-to-the-timeline">
<h2>Adding Report to the Timeline<a class="headerlink" href="#adding-report-to-the-timeline" title="Permalink to this heading"></a></h2>
<p>Starting with 2021.3, Nsight Systems can load multiple report files into a single timeline. This is a BETA feature and will be improved in the future releases. Please let us know about your experience on the forums or through <strong>Help > Send Feedback…</strong> in the main menu.</p>
<p>To load multiple report files into a single timeline, first start by opening a report as usual — using <strong>File > Open…</strong> from the main menu, or double clicking on a report in the Project Explorer window. Then additional report files can be loaded into the same timeline using one of the methods:</p>
<ul class="simple">
<li><p><strong>File > Add Report (beta)…</strong> in the main menu, and select another report file that you want to open</p></li>
<li><p>Right click on the report in the project explorer window, and click <strong>Add Report (beta)</strong></p></li>
</ul>
<img alt="TODO" class="image" src="../_images/report-add-another.png" />
<section id="time-synchronization">
<h3>Time Synchronization<a class="headerlink" href="#time-synchronization" title="Permalink to this heading"></a></h3>
<p>When multiple reports are loaded into a single timeline, timestamps between them need to be adjusted, such that events that happened at the same time appear to be aligned.</p>
<p>Nsight Systems can automatically adjust timestamps based on <strong>UTC time</strong> recorded around the collection start time. This method is used by default when other more precise methods are not available. This time can be seen as <code class="docutils literal notranslate"><span class="pre">UTC</span> <span class="pre">time</span> <span class="pre">at</span> <span class="pre">t=0</span></code> in the <em>Analysis Summary</em> page of the report file. Refer to your OS documentation to learn how to sync the software clock using the Network Time Protocol (NTP). NTP-based time synchronization is not very precise, with the typical errors on the scale of one to tens of milliseconds.</p>
<p>Reports collected on the same physical machine can use synchronization based on <strong>Timestamp Counter (TSC) values</strong>. These are platform-specific counters, typically accessed in user space applications using the RDTSC instruction on x86_64 architecture, or by reading the CNTVCT register on Arm64. Their values converted to nanoseconds can be seen as <code class="docutils literal notranslate"><span class="pre">TSC</span> <span class="pre">value</span> <span class="pre">at</span> <span class="pre">t=0</span></code> in the <em>Analysis Summary</em> page of the report file. Reports synchronized using TSC values can be aligned with nanoseconds-level precision.</p>
<p>TSC-based time synchronization is activated automatically, when Nsight Systems detects that reports come from same target and that the same TSC value corresponds to very close UTC times. Targets are considered to be the same when either explicitly set environment variables <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> are the same for both reports or when target hostnames are the same and <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> is not set for either target. The difference between UTC and TSC time offsets must be below 1 second to choose TSC-based time synchronization.</p>
<p>To find out which synchronization method was used, navigate to the <em>Analysis Summary</em> tab of an added report and check the <code class="docutils literal notranslate"><span class="pre">Report</span> <span class="pre">alignment</span> <span class="pre">source</span></code> property of a target. Note, that the first report won’t have this parameter.</p>
<img alt="TODO" class="image" src="../_images/report-alignment-source-TSC.png" />
<img alt="TODO" class="image" src="../_images/report-alignment-source-UTC.png" />
<p>When loading multiple reports into a single timeline, it is always advisable to first check that time synchronization looks correct, by zooming into synchronization or communication events that are expected to be aligned.</p>
</section>
<section id="timeline-hierarchy">
<h3>Timeline Hierarchy<a class="headerlink" href="#timeline-hierarchy" title="Permalink to this heading"></a></h3>
<p>When reports are added to the same timeline Nsight Systems will automatically line them up by timestamps as described above. If you want Nsight Systems to also recognize matching process or hardware information, you will need to set environment variables <code class="docutils literal notranslate"><span class="pre">NSYS_SYSTEM_ID</span></code> and <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> as shown below at the time of report collection (such as when using “nsys profile …” command).</p>
<p>When loading a pair of given report files into the same timeline, they will be merged in one of the following configurations:</p>
<ul class="simple">
<li><p>Different hardware — is used when reports are coming from different physical machines, and no hardware resources are shared in these reports. This mode is used when neither <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> or <code class="docutils literal notranslate"><span class="pre">NSYS_SYSTEM_ID</span></code> is set and target hostnames are different or absent, and can be additionally signalled by specifying different <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> values.</p></li>
<li><p>Different systems, same hardware — is used when reports are collected on different virtual machines (VMs) or containers on the same physical machine. To activate this mode, specify the same value of <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> when collecting the reports.</p></li>
<li><p>Same system — is used when reports are collected within the same operating system (or container) environment. In this mode a process identifier (PID) 100 will refer to the same process in both reports. To manually activate this mode, specify the same value of <code class="docutils literal notranslate"><span class="pre">NSYS_SYSTEM_ID</span></code> when collecting the reports. This mode is automatically selected when target hostnames are the same and neither <code class="docutils literal notranslate"><span class="pre">NSYS_HW_ID</span></code> or <code class="docutils literal notranslate"><span class="pre">NSYS_SYSTEM_ID</span></code> is provided.</p></li>
</ul>
<p>The following diagrams demonstrate typical cases:</p>
<img alt="TODO" class="image" src="../_images/report-tiling-diagram.png" />
</section>
<section id="example-mpi">
<h3>Example: MPI<a class="headerlink" href="#example-mpi" title="Permalink to this heading"></a></h3>
<p>A typical scenario is when a computing job is run using one of the MPI implementations. Each instance of the app can be profiled separately, resulting in multiple report files. For example:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="cp"># Run MPI job without the profiler:</span>
<span class="n">mpirun</span><span class="w"> </span><span class="o"><</span><span class="n">mpirun</span><span class="o">-</span><span class="n">options</span><span class="o">></span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">myApp</span><span class="w"></span>
<span class="cp"># Run MPI job and profile each instance of the application:</span>
<span class="n">mpirun</span><span class="w"> </span><span class="o"><</span><span class="n">mpirun</span><span class="o">-</span><span class="n">options</span><span class="o">></span><span class="w"> </span><span class="n">nsys</span><span class="w"> </span><span class="n">profile</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="n">report</span><span class="o">-%</span><span class="n">p</span><span class="w"> </span><span class="o"><</span><span class="n">nsys</span><span class="o">-</span><span class="n">options</span><span class="o">></span><span class="p">.</span><span class="o">/</span><span class="n">myApp</span><span class="w"></span>
</pre></div>
</div>
<p>When each MPI rank runs on a different node, the command above works fine, since the default pairing mode (different hardware) will be used.</p>
<p>When all MPI ranks run the localhost only, use this command (value “A” was chosen arbitrarily, it can be any non-empty string):</p>
<p><code class="docutils literal notranslate"><span class="pre">NSYS_SYSTEM_ID=A</span> <span class="pre">mpirun</span> <span class="pre"><mpirun-options></span> <span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre">-o</span> <span class="pre">report-%p</span> <span class="pre"><</span> <span class="pre">nsys</span> <span class="pre">-options></span> <span class="pre">./myApp</span></code></p>
<p>For convenience, the MPI rank can be encoded into the report filename. For Open MPI, use the following command to create report files based on the global rank value:</p>
<p><code class="docutils literal notranslate"><span class="pre">mpirun</span> <span class="pre"><mpirun-options></span> <span class="pre">nsys</span> <span class="pre">profile</span> <span class="pre">-o</span> <span class="pre">report-%q{OMPI_COMM_WORLD_RANK}</span> <span class="pre"><</span> <span class="pre">nsys</span> <span class="pre">-options></span> <span class="pre">./myApp</span></code></p>
<p>MPICH-based implementations set the environment variable <code class="docutils literal notranslate"><span class="pre">PMI_RANK</span></code> and Slurm (<code class="docutils literal notranslate"><span class="pre">srun</span></code>) provides the global MPI rank in <code class="docutils literal notranslate"><span class="pre">SLURM_PROCID</span></code>.</p>
</section>
<section id="id5">
<h3>Limitations<a class="headerlink" href="#id5" title="Permalink to this heading"></a></h3>
<ul class="simple">
<li><p>Only report files collected with Nsight Systems version 2021.3 and newer are fully supported.</p></li>
<li><p>Sequential reports collected in a single CLI profiling session cannot be loaded into a single timeline yet.</p></li>
</ul>
</section>
</section>
<section id="post-collection-analysis">
<h2>Post-Collection Analysis<a class="headerlink" href="#post-collection-analysis" title="Permalink to this heading"></a></h2>
<p>Once you have profiled using Nsight Systems there are many options for analyzing the collected data as well as to output it in various formats. These options are available from the CLI or the GUI</p>
<section id="available-export-formats">
<h3>Available Export Formats<a class="headerlink" href="#available-export-formats" title="Permalink to this heading"></a></h3>
<section id="sqlite-schema-reference">
<h4>SQLite Schema Reference<a class="headerlink" href="#sqlite-schema-reference" title="Permalink to this heading"></a></h4>
<p>Nsight Systems has the ability to export SQLite database files from the .nsys-rep results file. From the CLI, use <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">export</span></code>. From the GUI, call <code class="docutils literal notranslate"><span class="pre">File->Export...</span></code>.</p>
<p><strong>Note:</strong> The .nsys-rep report format is the only data format for Nsight Systems that should be considered forward compatible. The SQLite schema can and will change in the future.</p>
<p>The schema for a concrete database can be obtained with the sqlite3 tool built-in command <code class="docutils literal notranslate"><span class="pre">.schema</span></code>. The sqlite3 tool can be located in the Target or Host directory of your Nsight Systems installation.</p>
<p><strong>Note:</strong> Currently tables are created lazily, and therefore not every table described in the documentation will be present in a particular database. This will change in a future version of the product. If you want a full schema of all possible tables, use <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">export</span> <span class="pre">--lazy=false</span></code> during export phase.</p>
<p>Currently, a table is created for each data type in the exported database. Since usage patterns for exported data may vary greatly and no default use cases have been established, no indexes or extra constraints are created. Instead, refer to the SQLite Examples section for a list of common recipes. This may change in a future version of the product.</p>
<p>To check the version of your exported SQLite file, check the value of <code class="docutils literal notranslate"><span class="pre">EXPORT_SCHEMA_VERSION</span></code> in the <code class="docutils literal notranslate"><span class="pre">EXPORT_META_DATA</span></code> table. The schema version is a common three-value major/minor/micro version number. The first value, or major value, indicates the overall format of the database, and is only changed if there is a major re-write or re-factor of the entire database format. It is assumed that if the major version changes, all scripts or queries will break. The middle, or minor, version is changed anytime there is a more localized, but potentially breaking change, such as renaming an existing column, or changing the type of an existing column. The last, or micro version is changed any time there are additions, such as a new table or column, that should not introduce any breaking change when used with well-written, best-practices queries.</p>
<p>This is the schema as of the 2023.2 release, schema version 3.1.7.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Consolidation of repetitive string values.</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID reference value.</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- String value.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ThreadNames</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Thread name</span>
<span class="w"> </span><span class="n">priority</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Priority of the thread.</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">filenameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- File name</span>
<span class="w"> </span><span class="n">contentId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Stream content</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_SYSTEM_ENV</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">globalVid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">devStateName</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device state name.</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Property name.</span>
<span class="w"> </span><span class="n">nameEnum</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Property enum value.</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Property value.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_NIC_INFO</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">globalId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device state globalId</span>
<span class="w"> </span><span class="n">stateName</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device state name</span>
<span class="w"> </span><span class="n">nicId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Network interface Id.</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Network interface name</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NET_DEVICE_ID(id)</span>
<span class="w"> </span><span class="n">vendorId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NET_VENDOR_ID(id)</span>
<span class="w"> </span><span class="n">linkLayer</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NET_LINK_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_SESSION_START_TIME</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">utcEpochNs</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- UTC Epoch timestamp at start of the capture (ns).</span>
<span class="w"> </span><span class="n">utcTime</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Start of the capture in UTC.</span>
<span class="w"> </span><span class="k">localTime</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Start of the capture in local time of target.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ANALYSIS_DETAILS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Details about the analysis session.</span>
<span class="w"> </span><span class="n">globalVid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">duration</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The total time span of the entire trace (ns).</span>
<span class="w"> </span><span class="n">startTime</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Trace start timestamp in nanoseconds.</span>
<span class="w"> </span><span class="n">stopTime</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Trace stop timestamp in nanoseconds.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_GPU</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device name.</span>
<span class="w"> </span><span class="n">busLocation</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- PCI bus location.</span>
<span class="w"> </span><span class="n">isDiscrete</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if discrete, false if integrated.</span>
<span class="w"> </span><span class="n">l2CacheSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Size of L2 cache (B).</span>
<span class="w"> </span><span class="n">totalMemory</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total amount of memory on the device (B).</span>
<span class="w"> </span><span class="n">memoryBandwidth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Amount of memory transferred (B).</span>
<span class="w"> </span><span class="n">clockRate</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Clock frequency (Hz).</span>
<span class="w"> </span><span class="n">smCount</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of multiprocessors on the device.</span>
<span class="w"> </span><span class="n">pwGpuId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- PerfWorks GPU ID.</span>
<span class="w"> </span><span class="n">uuid</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device UUID.</span>
<span class="w"> </span><span class="n">luid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device LUID.</span>
<span class="w"> </span><span class="n">chipName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Chip name.</span>
<span class="w"> </span><span class="n">cuDevice</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA device ID.</span>
<span class="w"> </span><span class="n">ctxswDevPath</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU context switch device node path.</span>
<span class="w"> </span><span class="n">ctrlDevPath</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU control device node path.</span>
<span class="w"> </span><span class="n">revision</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Revision number.</span>
<span class="w"> </span><span class="n">nodeMask</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device node mask.</span>
<span class="w"> </span><span class="n">constantMemory</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Memory available on device for __constant__ variables (B).</span>
<span class="w"> </span><span class="n">maxIPC</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum instructions per count.</span>
<span class="w"> </span><span class="n">maxRegistersPerBlock</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum number of 32-bit registers available per block.</span>
<span class="w"> </span><span class="n">maxShmemPerBlock</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum optin shared memory per block.</span>
<span class="w"> </span><span class="n">maxShmemPerBlockOptin</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum optin shared memory per block.</span>
<span class="w"> </span><span class="n">maxShmemPerSm</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum shared memory available per multiprocessor (B).</span>
<span class="w"> </span><span class="n">maxRegistersPerSm</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum number of 32-bit registers available per multiprocessor.</span>
<span class="w"> </span><span class="n">threadsPerWarp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Warp size in threads.</span>
<span class="w"> </span><span class="n">asyncEngines</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of asynchronous engines.</span>
<span class="w"> </span><span class="n">maxWarpsPerSm</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum number of warps per multiprocessor.</span>
<span class="w"> </span><span class="n">maxBlocksPerSm</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum number of blocks per multiprocessor.</span>
<span class="w"> </span><span class="n">maxThreadsPerBlock</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum number of threads per block.</span>
<span class="w"> </span><span class="n">maxBlockDimX</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum X-dimension of a block.</span>
<span class="w"> </span><span class="n">maxBlockDimY</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum Y-dimension of a block.</span>
<span class="w"> </span><span class="n">maxBlockDimZ</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum Z-dimension of a block.</span>
<span class="w"> </span><span class="n">maxGridDimX</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum X-dimension of a grid.</span>
<span class="w"> </span><span class="n">maxGridDimY</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum Y-dimension of a grid.</span>
<span class="w"> </span><span class="n">maxGridDimZ</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Maximum Z-dimension of a grid.</span>
<span class="w"> </span><span class="n">computeMajor</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Major compute capability version number.</span>
<span class="w"> </span><span class="n">computeMinor</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Minor compute capability version number.</span>
<span class="w"> </span><span class="n">smMajor</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Major multiprocessor version number.</span>
<span class="w"> </span><span class="n">smMinor</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Minor multiprocessor version number.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_XMC_SPEC</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">clientId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Client ID.</span>
<span class="w"> </span><span class="k">type</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Client type.</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Client name.</span>
<span class="w"> </span><span class="n">groupId</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Client group ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_PROCESS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">openGlVersion</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- OpenGL version.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Correlation ID of the kernel.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_NVTX_CUDA_DEVICE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA device name assigned using NVTX.</span>
<span class="w"> </span><span class="n">hwId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Hardware ID.</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_NVTX_CUDA_CONTEXT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA context name assigned using NVTX.</span>
<span class="w"> </span><span class="n">hwId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Hardware ID.</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_NVTX_CUDA_STREAM</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA stream name assigned using NVTX.</span>
<span class="w"> </span><span class="n">hwId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Hardware ID.</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_CUDA_NULL_STREAM</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">hwId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Hardware ID.</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_CUDA_STREAM</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">hwId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Hardware ID.</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">priority</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Priority of the stream.</span>
<span class="w"> </span><span class="n">flag</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUPTI_STREAM_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_WDDM_CONTEXTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">nodeOrdinal</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">friendlyName</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_PERF_COUNTER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">counterId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Counter ID value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Counter name</span>
<span class="w"> </span><span class="n">description</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Counter description</span>
<span class="w"> </span><span class="n">unit</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Counter measurement unit</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_NETWORK_METRICS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">metricsListId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Metric list ID</span>
<span class="w"> </span><span class="n">metricsIdx</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- List index of metric</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Name of metric</span>
<span class="w"> </span><span class="n">description</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Description of metric</span>
<span class="w"> </span><span class="n">unit</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Measurement unit of metric</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">EXPORT_META_DATA</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- information about nsys export process</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Name of export meta-data record</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Value of export meta-data record</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NSYS_EVENT_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Nsys event type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NSYS_EVENT_CLASS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Nsys event class labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_GPU_CTX_SWITCH</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- GPU context switch labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_MEMCPY_OPER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA memcpy operation labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_MEM_KIND</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA memory kind labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_MEMPOOL_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA mempool type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_MEMPOOL_OPER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA mempool operation labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_DEV_MEM_EVENT_OPER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA device mem event operation labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_KERNEL_LAUNCH_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA kernel launch type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_SHARED_MEM_LIMIT_CONFIG</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA shared memory limit config labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_UNIF_MEM_MIGRATION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA unified memory migration cause labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_UNIF_MEM_ACCESS_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA unified memory access type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUDA_FUNC_CACHE_CONFIG</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUDA function cache config labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUPTI_STREAM_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUPTI stream type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_CUPTI_SYNC_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- CUPTI synchronization type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_STACK_UNWIND_METHOD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Stack unwind method labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_SAMPLING_THREAD_STATE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Sampling thread state labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENGL_DEBUG_SOURCE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenGL debug source labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENGL_DEBUG_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenGL debug type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENGL_DEBUG_SEVERITY</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenGL debug severity labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_VULKAN_PIPELINE_CREATION_FLAGS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Vulkan pipeline creation feedback flag labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_D3D12_HEAP_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D12 heap type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_D3D12_PAGE_PROPERTY</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D12 CPU page property labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_D3D12_HEAP_FLAGS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D12 heap flag labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_D3D12_CMD_LIST_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D12 command list type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENACC_DEVICE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenACC device type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENACC_EVENT_KIND</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenACC event type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_EVENT_KIND</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP event kind labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_THREAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP thread labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_DISPATCH</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP dispatch labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_SYNC_REGION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP sync region labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_WORK</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP work labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_MUTEX</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP mutex labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_TASK_FLAG</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP task flags labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_OPENMP_TASK_STATUS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP task status labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_SLI_TRANSER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenMP task status labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_DXGI_FORMAT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- DXGI image format labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NVDRIVER_EVENT_ID</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NV-Driver event it labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_WDDM_PAGING_QUEUE_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- WDDM paging queue type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_WDDM_PACKET_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- WDDM packet type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_WDDM_ENGINE_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- WDDM engine type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_WDDM_INTERRUPT_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- WDDM DMA interrupt type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_WDDM_VIDMM_OP_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- WDDM VidMm operation type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NET_LINK_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NIC link layer labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NET_DEVICE_ID</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NIC PCIe device id labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_NET_VENDOR_ID</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NIC PCIe vendor id labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ENUM_ETW_MEMORY_TRANSFER_TYPE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- memory transfer type labels</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum numerical value</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Enum symbol name</span>
<span class="w"> </span><span class="n">label</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Enum human name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GENERIC_EVENT_SOURCES</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">sourceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="k">data</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- JSON encoded generic event source description.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GENERIC_EVENT_TYPES</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">sourceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES GENERIC_EVENT_SOURCES(sourceId)</span>
<span class="w"> </span><span class="k">data</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- JSON encoded generic event type description.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GENERIC_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Events for which the schema is known only at the run-time.</span>
<span class="w"> </span><span class="n">rawTimestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Raw event timestamp recorded during profiling.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event timestamp converted to the profiling session timeline.</span>
<span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES GENERIC_EVENT_TYPES(typeId)</span>
<span class="w"> </span><span class="k">data</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- JSON encoded event data.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ETW_PROVIDERS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Names and identifiers of ETW providers captured in the report.</span>
<span class="w"> </span><span class="n">providerId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Provider ID.</span>
<span class="w"> </span><span class="n">providerNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Provider name</span>
<span class="w"> </span><span class="n">guid</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- ETW Provider GUID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ETW_TASKS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Names and identifiers of ETW tasks captured in the report.</span>
<span class="w"> </span><span class="n">taskNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Task name</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The event task ID.</span>
<span class="w"> </span><span class="n">providerId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Provider ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ETW_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW events captured in the report.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">processId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Process ID.</span>
<span class="w"> </span><span class="n">threadId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Thread ID.</span>
<span class="w"> </span><span class="n">providerId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Provider ID.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The event task ID.</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event ID.</span>
<span class="w"> </span><span class="k">version</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The event version.</span>
<span class="w"> </span><span class="n">opcode</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The event opcode.</span>
<span class="w"> </span><span class="k">data</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- JSON encoded event data.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TARGET_INFO_GPU_METRICS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- GPU Metrics, metric names and ids.</span>
<span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES GENERIC_EVENT_TYPES(typeId)</span>
<span class="w"> </span><span class="n">sourceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES GENERIC_EVENT_SOURCES(sourceId)</span>
<span class="w"> </span><span class="n">typeName</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Name of event type.</span>
<span class="w"> </span><span class="n">metricId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Id of metric in event; not assumed to be stable.</span>
<span class="w"> </span><span class="n">metricName</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Definitive name of metric.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GPU_METRICS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- GPU Metrics, events and values.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event timestamp (ns).</span>
<span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_GPU_METRICS(typeId) and GENERIC_EVENT_TYPES(typeId)</span>
<span class="w"> </span><span class="n">metricId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_GPU_METRICS(metricId)</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Counter data value</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_COMMUNICATORS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Identification of MPI communication groups.</span>
<span class="w"> </span><span class="n">rank</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Active MPI rank</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Time of MPI communicator creation.</span>
<span class="w"> </span><span class="n">commHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI communicator handle</span>
<span class="w"> </span><span class="n">parentHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI communicator handle</span>
<span class="w"> </span><span class="n">localRank</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Local MPI rank in a communicator.</span>
<span class="w"> </span><span class="k">size</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI communicator size.</span>
<span class="w"> </span><span class="n">groupRoot</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Root rank (global) in MPI communicator.</span>
<span class="w"> </span><span class="n">groupRootUid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Group root's communicator ID.</span>
<span class="w"> </span><span class="n">members</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- MPI communicator members (index is global rank).</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_MEMCPY</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUPTI_ACTIVITY_KIND_RUNTIME(correlationId)</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">bytes</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of bytes transferred (B).</span>
<span class="w"> </span><span class="n">copyKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEMCPY_OPER(id)</span>
<span class="w"> </span><span class="n">deprecatedSrcId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Deprecated, use srcDeviceId instead.</span>
<span class="w"> </span><span class="n">srcKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEM_KIND(id)</span>
<span class="w"> </span><span class="n">dstKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEM_KIND(id)</span>
<span class="w"> </span><span class="n">srcDeviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Source device ID.</span>
<span class="w"> </span><span class="n">srcContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Source context ID.</span>
<span class="w"> </span><span class="n">dstDeviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Destination device ID.</span>
<span class="w"> </span><span class="n">dstContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Destination context ID.</span>
<span class="w"> </span><span class="n">migrationCause</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_UNIF_MEM_MIGRATION(id)</span>
<span class="w"> </span><span class="n">graphNodeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUDA_GRAPH_EVENTS(graphNodeId)</span>
<span class="w"> </span><span class="n">virtualAddress</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Virtual base address of the page/s being transferred.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_MEMSET</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUPTI_ACTIVITY_KIND_RUNTIME(correlationId)</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Value assigned to memory.</span>
<span class="w"> </span><span class="n">bytes</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of bytes set (B).</span>
<span class="w"> </span><span class="n">graphNodeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUDA_GRAPH_EVENTS(graphNodeId)</span>
<span class="w"> </span><span class="n">memKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEM_KIND(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUPTI_ACTIVITY_KIND_RUNTIME(correlationId)</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">demangledName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Kernel function name w/ templates</span>
<span class="w"> </span><span class="n">shortName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Base kernel function name</span>
<span class="w"> </span><span class="n">mangledName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Raw C++ mangled kernel function name</span>
<span class="w"> </span><span class="n">launchType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_KERNEL_LAUNCH_TYPE(id)</span>
<span class="w"> </span><span class="n">cacheConfig</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_FUNC_CACHE_CONFIG(id)</span>
<span class="w"> </span><span class="n">registersPerThread</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of registers required for each thread executing the kernel.</span>
<span class="w"> </span><span class="n">gridX</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- X-dimension grid size.</span>
<span class="w"> </span><span class="n">gridY</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Y-dimension grid size.</span>
<span class="w"> </span><span class="n">gridZ</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Z-dimension grid size.</span>
<span class="w"> </span><span class="n">blockX</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- X-dimension block size.</span>
<span class="w"> </span><span class="n">blockY</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Y-dimension block size.</span>
<span class="w"> </span><span class="n">blockZ</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Z-dimension block size.</span>
<span class="w"> </span><span class="n">staticSharedMemory</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Static shared memory allocated for the kernel (B).</span>
<span class="w"> </span><span class="n">dynamicSharedMemory</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Dynamic shared memory reserved for the kernel (B).</span>
<span class="w"> </span><span class="n">localMemoryPerThread</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Amount of local memory reserved for each thread (B).</span>
<span class="w"> </span><span class="n">localMemoryTotal</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total amount of local memory reserved for the kernel (B).</span>
<span class="w"> </span><span class="n">gridId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Unique grid ID of the kernel assigned at runtime.</span>
<span class="w"> </span><span class="n">sharedMemoryExecuted</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Shared memory size set by the driver.</span>
<span class="w"> </span><span class="n">graphNodeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUDA_GRAPH_EVENTS(graphNodeId)</span>
<span class="w"> </span><span class="n">sharedMemoryLimitConfig</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_SHARED_MEM_LIMIT_CONFIG(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_SYNCHRONIZATION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Correlation ID of the synchronization API to which this result is associated.</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">syncType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUPTI_SYNC_TYPE(id)</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Event ID for which the synchronization API is called.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_CUDA_EVENT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">streamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Stream ID.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Correlation ID of the event record API to which this result is associated.</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Event ID for which the event record API is called.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID used to identify events that this function call has triggered.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">returnValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Return value of the function call.</span>
<span class="w"> </span><span class="n">callchainId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES CUDA_CALLCHAINS(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDNN_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUBLAS_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_GRAPH_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">graphNodeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUDA_GRAPH_EVENTS(graphNodeId)</span>
<span class="w"> </span><span class="n">originalGraphNodeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Reference to the original graph node ID, if cloned node.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_UM_CPU_PAGE_FAULT_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">address</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Virtual address of the page that faulted.</span>
<span class="w"> </span><span class="n">originalFaultPc</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Program counter of the CPU instruction that caused the page fault.</span>
<span class="w"> </span><span class="n">CpuInstruction</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Module name</span>
<span class="w"> </span><span class="n">unresolvedFaultPc</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- True if the program counter was not resolved.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_UM_GPU_PAGE_FAULT_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">address</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Virtual address of the page that faulted.</span>
<span class="w"> </span><span class="n">numberOfPageFaults</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of page faults for the same page.</span>
<span class="w"> </span><span class="n">faultAccessType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_UNIF_MEM_ACCESS_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_GPU_MEMORY_USAGE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">address</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Virtual address of the allocation/deallocation.</span>
<span class="w"> </span><span class="n">pc</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Program counter of the allocation/deallocation.</span>
<span class="w"> </span><span class="n">bytes</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of bytes allocated/deallocated (B).</span>
<span class="w"> </span><span class="n">memKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEM_KIND(id)</span>
<span class="w"> </span><span class="n">memoryOperationType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_DEV_MEM_EVENT_OPER(id)</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Variable name, if available.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUPTI_ACTIVITY_KIND_RUNTIME(correlationId)</span>
<span class="w"> </span><span class="n">localMemoryPoolAddress</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Base address of the local memory pool used</span>
<span class="w"> </span><span class="n">localMemoryPoolReleaseThreshold</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Release threshold of the local memory pool used</span>
<span class="w"> </span><span class="n">localMemoryPoolSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Size of the local memory pool used</span>
<span class="w"> </span><span class="n">localMemoryPoolUtilizedSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Utilized size of the local memory pool used</span>
<span class="w"> </span><span class="n">importedMemoryPoolAddress</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Base address of the imported memory pool used</span>
<span class="w"> </span><span class="n">importedMemoryPoolProcessId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Process ID of the imported memory pool used</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_GPU_MEMORY_POOL_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">deviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Device ID.</span>
<span class="w"> </span><span class="n">address</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The base virtual address of the memory pool.</span>
<span class="w"> </span><span class="n">operationType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEMPOOL_OPER(id)</span>
<span class="w"> </span><span class="n">poolType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_CUDA_MEMPOOL_TYPE(id)</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES CUPTI_ACTIVITY_KIND_RUNTIME(correlationId)</span>
<span class="w"> </span><span class="n">minBytesToKeep</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Minimum number of bytes to keep of the memory pool.</span>
<span class="w"> </span><span class="n">localMemoryPoolReleaseThreshold</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Release threshold of the local memory pool used</span>
<span class="w"> </span><span class="n">localMemoryPoolSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Size of the local memory pool used</span>
<span class="w"> </span><span class="n">localMemoryPoolUtilizedSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Utilized size of the local memory pool used</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_CALLCHAINS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Part of PRIMARY KEY (id, stackDepth).</span>
<span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Module name</span>
<span class="w"> </span><span class="n">unresolved</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if the symbol was not resolved.</span>
<span class="w"> </span><span class="n">originalIP</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Instruction pointer value.</span>
<span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Zero-base index of the given function in call stack.</span>
<span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="w"> </span><span class="p">(</span><span class="n">id</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="p">)</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_RANKS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Mapping of global thread IDs (gtid) to MPI ranks</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">rank</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- MPI rank</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_P2P_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Registered NVTX domain</span>
<span class="w"> </span><span class="n">commHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI communicator handle</span>
<span class="w"> </span><span class="n">tag</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI message tag</span>
<span class="w"> </span><span class="n">remoteRank</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI remote rank (destination or source)</span>
<span class="w"> </span><span class="k">size</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI message size in bytes</span>
<span class="w"> </span><span class="n">requestHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- MPI request handle</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_COLLECTIVES_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Registered NVTX domain</span>
<span class="w"> </span><span class="n">commHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI communicator handle</span>
<span class="w"> </span><span class="n">rootRank</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- root rank in the collective</span>
<span class="w"> </span><span class="k">size</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI message size in bytes (send size for bidirectional ops)</span>
<span class="w"> </span><span class="n">recvSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- MPI receive size in bytes</span>
<span class="w"> </span><span class="n">requestHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- MPI request handle</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_START_WAIT_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- MPI_Wait* and MPI_Start*</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Registered NVTX domain</span>
<span class="w"> </span><span class="n">requestHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- MPI request handle</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MPI_OTHER_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- MPI events without additional parameters</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Registered NVTX domain</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_TYPE(id)</span>
<span class="w"> </span><span class="n">rangeId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Correlation ID returned from a nvtxRangeStart call.</span>
<span class="w"> </span><span class="n">category</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- User-controlled ID that can be used to group events.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="nb">text</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Optional text message for non registered strings.</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">endGlobalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Registered NVTX domain</span>
<span class="w"> </span><span class="n">domainId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- User-controlled ID that can be used to group events.</span>
<span class="w"> </span><span class="n">uint64Value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">int64Value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">doubleValue</span><span class="w"> </span><span class="nb">REAL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">uint32Value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">int32Value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">floatValue</span><span class="w"> </span><span class="nb">REAL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">jsonTextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="w"> </span><span class="n">jsonText</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- One of possible payload value union members.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENGL_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_TYPE(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">endGlobalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- First function name</span>
<span class="w"> </span><span class="n">endNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Last function name</span>
<span class="w"> </span><span class="n">returnValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Return value of the function call.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">display</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Display ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENGL_WORKLOAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_TYPE(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">endGlobalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- First function name</span>
<span class="w"> </span><span class="n">endNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Last function name</span>
<span class="w"> </span><span class="n">returnValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Return value of the function call.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">display</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Display ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">KHR_DEBUG_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_TYPE(id)</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Debug marker/group text</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="k">source</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENGL_DEBUG_SOURCE(id)</span>
<span class="w"> </span><span class="n">khrdType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENGL_DEBUG_TYPE(id)</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- KHR event ID.</span>
<span class="w"> </span><span class="n">severity</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENGL_DEBUG_SEVERITY(id)</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID used to correlate KHR CPU trace to GPU trace.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OSRT_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OS runtime libraries traced to gather information about low-level userspace APIs.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">returnValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Return value of the function call.</span>
<span class="w"> </span><span class="n">nestingLevel</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Zero-base index of the nesting level.</span>
<span class="w"> </span><span class="n">callchainId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES OSRT_CALLCHAINS(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Callchains attached to OSRT events, depending on selected profiling settings.</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Part of PRIMARY KEY (id, stackDepth).</span>
<span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Module name</span>
<span class="w"> </span><span class="n">kernelMode</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if kernel mode.</span>
<span class="w"> </span><span class="n">thumbCode</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if thumb code.</span>
<span class="w"> </span><span class="n">unresolved</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if the symbol was not resolved.</span>
<span class="w"> </span><span class="n">specialEntry</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if artifical entry added during processing callchain.</span>
<span class="w"> </span><span class="n">originalIP</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Instruction pointer value.</span>
<span class="w"> </span><span class="n">unwindMethod</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_STACK_UNWIND_METHOD(id)</span>
<span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Zero-base index of the given function in call stack.</span>
<span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="w"> </span><span class="p">(</span><span class="n">id</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="p">)</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">UnwindMethodType</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="nb">number</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">PROFILER_OVERHEAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">returnValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Return value of the function call.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SCHED_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Thread scheduling events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">cpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of CPU this thread was scheduled in or out.</span>
<span class="w"> </span><span class="n">isSchedIn</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- 0 if thread was scheduled out, non-zero otherwise.</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Thread sampling events.</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the composite event.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">cpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of CPU this thread was running on.</span>
<span class="w"> </span><span class="n">threadState</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_SAMPLING_THREAD_STATE(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">cpuCycles</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Value of Performance Monitoring Unit (PMU) counter.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Callchain entries obtained from composite events, used to construct function table views.</span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES COMPOSITE_EVENTS(id)</span>
<span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Module name</span>
<span class="w"> </span><span class="n">kernelMode</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if kernel mode.</span>
<span class="w"> </span><span class="n">thumbCode</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if thumb code.</span>
<span class="w"> </span><span class="n">unresolved</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if the symbol was not resolved.</span>
<span class="w"> </span><span class="n">specialEntry</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if artifical entry added during processing callchain.</span>
<span class="w"> </span><span class="n">originalIP</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Instruction pointer value.</span>
<span class="w"> </span><span class="n">unwindMethod</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_STACK_UNWIND_METHOD(id)</span>
<span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Zero-base index of the given function in call stack.</span>
<span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="w"> </span><span class="p">(</span><span class="n">id</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="p">)</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">PERF_EVENT_CPU_COUNTER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Sampled CPU Performance Counters.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">componentId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Perf Event Component ID (usually CPU Core ID)</span>
<span class="w"> </span><span class="n">counterId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_PERF_COUNTER(counterId)</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Counter data value</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SLI_QUERIES</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">occQueryIssued</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Occlusion query issued.</span>
<span class="w"> </span><span class="n">occQueryAsked</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Occlusion query asked.</span>
<span class="w"> </span><span class="n">eventQueryIssued</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event query issued.</span>
<span class="w"> </span><span class="n">eventQueryAsked</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event query asked.</span>
<span class="w"> </span><span class="n">numberOfTransferEvents</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of transfer events.</span>
<span class="w"> </span><span class="n">amountOfTransferredData</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Cumulative size of resource data that was transferred.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SLI_P2P</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">transferSkipped</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of transfers that were skipped.</span>
<span class="w"> </span><span class="n">srcGpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Source GPU ID.</span>
<span class="w"> </span><span class="n">dstGpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Destination GPU ID.</span>
<span class="w"> </span><span class="n">numSubResources</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of sub-resources to transfer.</span>
<span class="w"> </span><span class="n">resourceSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Size of resource.</span>
<span class="w"> </span><span class="n">subResourceIdx</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Sub-resource index.</span>
<span class="w"> </span><span class="n">smplWidth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Sub-resource surface width in samples.</span>
<span class="w"> </span><span class="n">smplHeight</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Sub-resource surface height in samples.</span>
<span class="w"> </span><span class="n">smplDepth</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Sub-resource surface depth in samples.</span>
<span class="w"> </span><span class="n">bytesPerElement</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of bytes per element.</span>
<span class="w"> </span><span class="n">dxgiFormat</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_DXGI_FORMAT(id)</span>
<span class="w"> </span><span class="n">logSurfaceNames</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Surface name.</span>
<span class="w"> </span><span class="n">transferInfo</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_SLI_TRANSER(id)</span>
<span class="w"> </span><span class="n">isEarlyPushManagedByNvApi</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if early push managed by NVAPI. False otherwise.</span>
<span class="w"> </span><span class="n">useAsyncP2pForResolve</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True if async Peer-to-Peer used for resolve. False otherwise.</span>
<span class="w"> </span><span class="n">transferFuncName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- "A - BE" for asynchronous transfer, "S - BE" for synchronous transfer.</span>
<span class="w"> </span><span class="n">regimeName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Name of the regime scope that includes the resource.</span>
<span class="w"> </span><span class="n">debugName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Debug name assigned to the resource by the application code.</span>
<span class="w"> </span><span class="n">bindType</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- Bind type.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SLI_STATS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">countComplexFrames</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Complex frames count.</span>
<span class="w"> </span><span class="n">countStats</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of frame statistics collected for the inactive-time histogram.</span>
<span class="w"> </span><span class="n">totalInactiveTime</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total inactive time (µs).</span>
<span class="w"> </span><span class="n">minPbSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Min push buffer size.</span>
<span class="w"> </span><span class="n">maxPbSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Max push buffer size.</span>
<span class="w"> </span><span class="n">totalPbSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Total push buffer size.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">DX12_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">shortContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the COM interface object address.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- PIX marker text</span>
<span class="w"> </span><span class="n">commandListType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_CMD_LIST_TYPE(id)</span>
<span class="w"> </span><span class="n">objectNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- D3D12 object name</span>
<span class="w"> </span><span class="n">longContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Long form of the COM interface object address.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">DX12_WORKLOAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">shortContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the COM interface object address.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- PIX marker text</span>
<span class="w"> </span><span class="n">commandListType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_CMD_LIST_TYPE(id)</span>
<span class="w"> </span><span class="n">objectNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- D3D12 object name</span>
<span class="w"> </span><span class="n">longContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Long form of the COM interface object address.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">DX12_MEMORY_OPERATION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">rangeStart</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Offset denoting the beginning of a memory range (B).</span>
<span class="w"> </span><span class="n">rangeEnd</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Offset denoting the end of a memory range (B).</span>
<span class="w"> </span><span class="n">subresourceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Subresource index.</span>
<span class="w"> </span><span class="n">heapType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_HEAP_TYPE(id)</span>
<span class="w"> </span><span class="n">heapFlags</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_HEAP_FLAGS(id)</span>
<span class="w"> </span><span class="n">cpuPageProperty</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_PAGE_PROPERTY(id)</span>
<span class="w"> </span><span class="n">nvApiFlags</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- NV specific flags. See docs for specifics.</span>
<span class="w"> </span><span class="n">traceEventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES DX12_API(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">VULKAN_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Short form of the interface object address.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">VULKAN_WORKLOAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the interface object address.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Vulkan CPU debug marker string</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">VULKAN_DEBUG_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the interface object address.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Vulkan CPU debug marker string</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">VULKAN_PIPELINE_CREATION_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the pipeline creation event.</span>
<span class="w"> </span><span class="n">duration</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event duration (ns).</span>
<span class="w"> </span><span class="n">flags</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_VULKAN_PIPELINE_CREATION_FLAGS(id)</span>
<span class="w"> </span><span class="n">traceEventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES VULKAN_API(id) -- ID of the attached vulkan API.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">VULKAN_PIPELINE_STAGE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">PRIMARY</span><span class="w"> </span><span class="k">KEY</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the pipeline stage event.</span>
<span class="w"> </span><span class="n">duration</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event duration (ns).</span>
<span class="w"> </span><span class="n">flags</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_VULKAN_PIPELINE_CREATION_FLAGS(id)</span>
<span class="w"> </span><span class="n">creationEventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES VULKAN_PIPELINE_CREATION_EVENTS(id) -- ID of the attached pipeline creation event.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GPU_CONTEXT_SWITCH_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="n">tag</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_GPU_CTX_SWITCH(id)</span>
<span class="w"> </span><span class="n">vmId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- VM ID.</span>
<span class="w"> </span><span class="n">seqNo</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Sequential event number.</span>
<span class="w"> </span><span class="n">contextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Context ID.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpuId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ETW_EVENTS_DEPRECATED_TABLE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">providerName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Tracing events provider name</span>
<span class="w"> </span><span class="n">taskName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event task name</span>
<span class="w"> </span><span class="n">description</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Decoded event value</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_THREAD</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">threadId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal thread sequence starting from 1.</span>
<span class="w"> </span><span class="n">threadType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_THREAD(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_PARALLEL</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal parallel region sequence starting from 1.</span>
<span class="w"> </span><span class="n">parentTaskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID for task that creates this parallel region.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_SYNC_REGION_WAIT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_SYNC_REGION(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_SYNC_REGION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_SYNC_REGION(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_TASK_FLAG(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MASTER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_REDUCTION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK_CREATE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parentTaskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parent task that is creating a new task.</span>
<span class="w"> </span><span class="n">newTaskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the new task that is being created.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK_SCHEDULE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">priorTaskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the task that is being switched out.</span>
<span class="w"> </span><span class="n">priorTaskStatus</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_TASK_STATUS(id)</span>
<span class="w"> </span><span class="n">nextTaskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that is being switched in.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_CANCEL</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that is being cancelled.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MUTEX_WAIT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_MUTEX(id)</span>
<span class="w"> </span><span class="n">waitId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID indicating the object being waited.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_CRITICAL_SECTION</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_MUTEX(id)</span>
<span class="w"> </span><span class="n">waitId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID indicating the object being held.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MUTEX_RELEASED</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_MUTEX(id)</span>
<span class="w"> </span><span class="n">waitId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID indicating the object being released.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_LOCK_INIT</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_MUTEX(id)</span>
<span class="w"> </span><span class="n">waitId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID indicating object being created/destroyed.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_LOCK_DESTROY</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_MUTEX(id)</span>
<span class="w"> </span><span class="n">waitId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID indicating object being created/destroyed.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_WORKSHARE</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_WORK(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="w"> </span><span class="k">count</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Measure of the quantity of work involved in the region.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_DISPATCH</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">kind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_DISPATCH(id)</span>
<span class="w"> </span><span class="n">parallelId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- ID of the parallel region that this event belongs to.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the task that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_FLUSH</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Currently unused.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENMP_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">threadId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- ID of the thread that this event belongs to.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">D3D11_PIX_DEBUG_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D11 debug marker events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">shortContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the COM interface object address.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- PIX marker text</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">D3D12_PIX_DEBUG_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- D3D12 debug marker events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">correlationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- First ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">endCorrelationId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Last ID matching an API call to GPU workloads.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="w"> </span><span class="n">shortContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Short form of the COM interface object address.</span>
<span class="w"> </span><span class="n">frameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Index of the graphics frame starting from 1.</span>
<span class="w"> </span><span class="n">color</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Encoded ARGB color value.</span>
<span class="w"> </span><span class="n">textId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- PIX marker text</span>
<span class="w"> </span><span class="n">commandListType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_D3D12_CMD_LIST_TYPE(id)</span>
<span class="w"> </span><span class="n">objectNameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- D3D12 object name</span>
<span class="w"> </span><span class="n">longContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Long form of the COM interface object address.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_EVICT_ALLOCATION_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW EvictAllocation events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">allocationHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Global allocation handle.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_PAGING_QUEUE_PACKET_START_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW PagingQueuePacketStart events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">dxgDevice</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of an IDXGIDevice.</span>
<span class="w"> </span><span class="n">dxgAdapter</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of an IDXGIAdapter.</span>
<span class="w"> </span><span class="n">pagingQueue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue.</span>
<span class="w"> </span><span class="n">pagingQueuePacket</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue packet.</span>
<span class="w"> </span><span class="n">sequenceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 0.</span>
<span class="w"> </span><span class="n">alloc</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Allocation handle.</span>
<span class="w"> </span><span class="n">vidMmOpType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_VIDMM_OP_TYPE(id)</span>
<span class="w"> </span><span class="n">pagingQueueType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PAGING_QUEUE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_PAGING_QUEUE_PACKET_STOP_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW PagingQueuePacketStop events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">pagingQueue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue.</span>
<span class="w"> </span><span class="n">pagingQueuePacket</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue packet.</span>
<span class="w"> </span><span class="n">sequenceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 0.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_PAGING_QUEUE_PACKET_INFO_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW PagingQueuePacketInfo events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">pagingQueue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue.</span>
<span class="w"> </span><span class="n">pagingQueuePacket</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the paging queue packet.</span>
<span class="w"> </span><span class="n">sequenceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 0.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_QUEUE_PACKET_START_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW QueuePacketStart events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">dmaBufferSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The dma buffer size.</span>
<span class="w"> </span><span class="n">dmaBuffer</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The reported address of dma buffer.</span>
<span class="w"> </span><span class="n">queuePacket</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The address of queue packet.</span>
<span class="w"> </span><span class="n">progressFenceValue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The fence value.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">submitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">allocationListSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The number of allocations referenced.</span>
<span class="w"> </span><span class="n">patchLocationListSize</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The number of patch locations.</span>
<span class="w"> </span><span class="n">present</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True or False if the packet is a present packet.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="w"> </span><span class="n">syncObject</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- The address of fence object.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_QUEUE_PACKET_STOP_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW QueuePacketStop events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">queuePacket</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The address of queue packet.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">submitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">preempted</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True or False if the packet is preempted.</span>
<span class="w"> </span><span class="n">timeouted</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True or False if the packet is timeouted.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_QUEUE_PACKET_INFO_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW QueuePacketInfo events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">submitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_DMA_PACKET_START_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW DmaPacketStart events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">queuePacketContext</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The queue packet context.</span>
<span class="w"> </span><span class="n">uliSubmissionId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The queue packet submission ID.</span>
<span class="w"> </span><span class="n">dmaBuffer</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The reported address of dma buffer.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">ulQueueSubmitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">quantumStatus</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The quantum Status.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_DMA_PACKET_STOP_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW DmaPacketStop events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">uliCompletionId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The queue packet completion ID.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">ulQueueSubmitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">preempted</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- True or False if the packet is preempted.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_DMA_PACKET_INFO_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW DmaPacketInfo events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">uliCompletionId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The queue packet completion ID.</span>
<span class="w"> </span><span class="n">faultedVirtualAddress</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The virtual address of faulted process.</span>
<span class="w"> </span><span class="n">faultedProcessHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The address of faulted process.</span>
<span class="w"> </span><span class="n">packetType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_PACKET_TYPE(id)</span>
<span class="w"> </span><span class="n">ulQueueSubmitSequence</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Internal sequence starting from 1.</span>
<span class="w"> </span><span class="n">interruptType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_INTERRUPT_TYPE(id)</span>
<span class="w"> </span><span class="n">quantumStatus</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The quantum Status.</span>
<span class="w"> </span><span class="n">pageFaultFlags</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The page fault flag ID.</span>
<span class="w"> </span><span class="n">engineType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_WDDM_ENGINE_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">WDDM_HW_QUEUE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW HwQueueStart events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">context</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The context ID of WDDM queue.</span>
<span class="w"> </span><span class="n">hwQueue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The address of HW queue.</span>
<span class="w"> </span><span class="n">parentDxgHwQueue</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- The address of parent Dxg HW queue.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NVVIDEO_ENCODER_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NV Video Encoder API traced to gather information about NVIDIA Video Codek SDK Encoder APIs.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NVVIDEO_DECODER_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NV Video Encoder API traced to gather information about NVIDIA Video Codek SDK Decoder APIs.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NVVIDEO_JPEG_API</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- NV Video Encoder API traced to gather information about NVIDIA Video Codek SDK JPEG APIs.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">eventClass</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_NSYS_EVENT_CLASS(id)</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GPU_MEMORY_BUDGET_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW VidMmProcessBudgetChange events.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">newBudget</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The new budget size in bytes.</span>
<span class="w"> </span><span class="n">segmentGroup</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- The segment group ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">GPU_MEMORY_USAGE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW VidMmProcessUsageChange events.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">newUsage</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The new usage size in bytes.</span>
<span class="w"> </span><span class="n">segmentGroup</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- The segment group ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">DEMOTED_BYTES_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW VidMmProcessDemotedCommitmentChange events.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">commitment</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- Total demoted bytes.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">TOTAL_BYTES_RESIDENT_IN_SEGMENT_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW TotalBytesResidentInSegment events.</span>
<span class="w"> </span><span class="k">timestamp</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">totalBytesResident</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total bytes resident in segment.</span>
<span class="w"> </span><span class="n">segmentGroup</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The segment group ID.</span>
<span class="w"> </span><span class="n">segmentId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- The segment ID.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">MEMORY_TRANSFER_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW Memory Transfer events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">taskId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The event task ID.</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event ID.</span>
<span class="w"> </span><span class="n">allocationGlobalHandle</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Address of the global allocation handle.</span>
<span class="w"> </span><span class="n">dmaBuffer</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The reported address of dma buffer.</span>
<span class="w"> </span><span class="k">size</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The size of the dma buffer in bytes.</span>
<span class="w"> </span><span class="k">offset</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The offset from the start of the reported dma buffer in bytes.</span>
<span class="w"> </span><span class="n">memoryTransferType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_ETW_MEMORY_TRANSFER_TYPE(id)</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NV_LOAD_BALANCE_MASTER_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW NV-wgf2um LoadBalanceMaster events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event ID.</span>
<span class="w"> </span><span class="n">task</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The task name.</span>
<span class="w"> </span><span class="n">frameCount</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The frame ID.</span>
<span class="w"> </span><span class="n">frameTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Frame duration.</span>
<span class="w"> </span><span class="n">averageFrameTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average of frame duration.</span>
<span class="w"> </span><span class="n">averageLatency</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average of latency.</span>
<span class="w"> </span><span class="n">minLatency</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The minimum latency.</span>
<span class="w"> </span><span class="n">averageQueuedFrames</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average number of queued frames.</span>
<span class="w"> </span><span class="n">totalActiveMs</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total active time in milliseconds.</span>
<span class="w"> </span><span class="n">totalIdleMs</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Total idle time in milliseconds.</span>
<span class="w"> </span><span class="n">idlePercent</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The percentage of idle time.</span>
<span class="w"> </span><span class="n">isGPUAlmostOneFrameAhead</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- True or False if GPU is almost one frame ahead.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NV_LOAD_BALANCE_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Raw ETW NV-wgf2um LoadBalance events.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">gpu</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- GPU index.</span>
<span class="w"> </span><span class="n">eventId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event ID.</span>
<span class="w"> </span><span class="n">task</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The task name.</span>
<span class="w"> </span><span class="n">averageFPS</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average frame per second.</span>
<span class="w"> </span><span class="n">queuedFrames</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The amount of queued frames.</span>
<span class="w"> </span><span class="n">averageQueuedFrames</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average number of queued frames.</span>
<span class="w"> </span><span class="n">currentCPUTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The current CPU time.</span>
<span class="w"> </span><span class="n">averageCPUTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average CPU time.</span>
<span class="w"> </span><span class="n">averageStallTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average of stall time.</span>
<span class="w"> </span><span class="n">averageCPUIdleTime</span><span class="w"> </span><span class="nb">REAL</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Average CPU idle time.</span>
<span class="w"> </span><span class="n">isGPUAlmostOneFrameAhead</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- True or False if GPU is almost one frame ahead.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">PROCESSES</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- Names and identifiers of processes captured in the report.</span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">pid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- The process ID.</span>
<span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="w"> </span><span class="c1">-- The process name.</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_OPENACC_DATA</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenACC data events collected using CUPTI.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">DeviceType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_DEVICE(id)</span>
<span class="w"> </span><span class="n">lineNo</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Line number of the directive or program construct.</span>
<span class="w"> </span><span class="n">cuDeviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA device ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA context ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuStreamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA stream ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">srcFile</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Source file name or path</span>
<span class="w"> </span><span class="n">funcName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function in which event occurred</span>
<span class="w"> </span><span class="n">bytes</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of bytes.</span>
<span class="w"> </span><span class="n">varName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Variable name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_OPENACC_LAUNCH</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenACC launch events collected using CUPTI.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">DeviceType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_DEVICE(id)</span>
<span class="w"> </span><span class="n">lineNo</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Line number of the directive or program construct.</span>
<span class="w"> </span><span class="n">cuDeviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA device ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA context ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuStreamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA stream ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">srcFile</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Source file name or path</span>
<span class="w"> </span><span class="n">funcName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function in which event occurred</span>
<span class="w"> </span><span class="n">numGangs</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of gangs created for this kernel launch.</span>
<span class="w"> </span><span class="n">numWorkers</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of workers created for this kernel launch.</span>
<span class="w"> </span><span class="n">vectorLength</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Number of vector lanes created for this kernel launch.</span>
<span class="w"> </span><span class="n">kernelName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Kernel name</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_OPENACC_OTHER</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="c1">-- OpenACC other events collected using CUPTI.</span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Event name</span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">eventKind</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_EVENT_KIND(id)</span>
<span class="w"> </span><span class="n">DeviceType</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES ENUM_OPENACC_DEVICE(id)</span>
<span class="w"> </span><span class="n">lineNo</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Line number of the directive or program construct.</span>
<span class="w"> </span><span class="n">cuDeviceId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA device ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuContextId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA context ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">cuStreamId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- CUDA stream ID. Valid only if deviceType is acc_device_nvidia.</span>
<span class="w"> </span><span class="n">srcFile</span><span class="w"> </span><span class="nb">INTEGER</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Source file name or path</span>
<span class="w"> </span><span class="n">funcName</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="c1">-- REFERENCES StringIds(id) -- Function in which event occurred</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NET_NIC_METRIC</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">metricsListId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_NETWORK_METRICS(metricsListId)</span>
<span class="w"> </span><span class="n">metricsIdx</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_NETWORK_METRICS(metricsIdx)</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Counter data value</span>
<span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">NET_IB_SWITCH_METRIC</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event start timestamp (ns).</span>
<span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Event end timestamp (ns).</span>
<span class="w"> </span><span class="n">globalId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- Serialized GlobalId.</span>
<span class="w"> </span><span class="n">metricsListId</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_NETWORK_METRICS(metricsListId)</span>
<span class="w"> </span><span class="n">metricsIdx</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="p">,</span><span class="w"> </span><span class="c1">-- REFERENCES TARGET_INFO_NETWORK_METRICS(metricsIdx)</span>
<span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="nb">INTEGER</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="c1">-- Counter data value</span>
<span class="p">);</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="sqlite-schema-event-values">
<h4>SQLite Schema Event Values<a class="headerlink" href="#sqlite-schema-event-values" title="Permalink to this heading"></a></h4>
<p>Here are the set values stored in enums in the Nsight Systems SQLite schema</p>
<p><strong>CUDA Event Class Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDA_RUNTIME</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDA_DRIVER</span><span class="w"></span>
<span class="mi">13</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDA_EGL_DRIVER</span><span class="w"></span>
<span class="mi">28</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDNN</span><span class="w"></span>
<span class="mi">29</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUBLAS</span><span class="w"></span>
<span class="mi">33</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDNN_START</span><span class="w"></span>
<span class="mi">34</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDNN_FINISH</span><span class="w"></span>
<span class="mi">35</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUBLAS_START</span><span class="w"></span>
<span class="mi">36</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUBLAS_FINISH</span><span class="w"></span>
<span class="mi">67</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDABACKTRACE</span><span class="w"></span>
<span class="mi">77</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_CUDA_GRAPH_NODE_CREATION</span><span class="w"></span>
</pre></div>
</div>
<p>See <a class="reference external" href="https://docs.nvidia.com/cupti/Cupti/index.html">CUPTI documentation</a> for detailed information on collected event and data types.</p>
<p><strong>NVTX Event Type Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">33</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxCategory</span><span class="w"></span>
<span class="mi">34</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxMark</span><span class="w"></span>
<span class="mi">39</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxThread</span><span class="w"></span>
<span class="mi">59</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxPushPopRange</span><span class="w"></span>
<span class="mi">60</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxStartEndRange</span><span class="w"></span>
<span class="mi">75</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxDomainCreate</span><span class="w"></span>
<span class="mi">76</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">NvtxDomainDestroy</span><span class="w"></span>
</pre></div>
</div>
<p>The difference between text and textId columns is that if an NVTX event message was passed via call to nvtxDomainRegisterString function, then the message will be available through textId field, otherwise the text field will contain the message if it was provided.</p>
<p><strong>OpenGL Events</strong></p>
<p>KHR event class values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">62</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">KhrDebugPushPopRange</span><span class="w"></span>
<span class="mi">63</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">KhrDebugGpuPushPopRange</span><span class="w"></span>
</pre></div>
</div>
<p>KHR source kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mh">0x8249</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SOURCE_THIRD_PARTY</span><span class="w"></span>
<span class="mh">0x824A</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SOURCE_APPLICATION</span><span class="w"></span>
</pre></div>
</div>
<p>KHR type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mh">0x824C</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_ERROR</span><span class="w"></span>
<span class="mh">0x824D</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_DEPRECATED_BEHAVIOR</span><span class="w"></span>
<span class="mh">0x824E</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_UNDEFINED_BEHAVIOR</span><span class="w"></span>
<span class="mh">0x824F</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_PORTABILITY</span><span class="w"></span>
<span class="mh">0x8250</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_PERFORMANCE</span><span class="w"></span>
<span class="mh">0x8251</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_OTHER</span><span class="w"></span>
<span class="mh">0x8268</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_MARKER</span><span class="w"></span>
<span class="mh">0x8269</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_PUSH_GROUP</span><span class="w"></span>
<span class="mh">0x826A</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_TYPE_POP_GROUP</span><span class="w"></span>
</pre></div>
</div>
<p>KHR severity values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mh">0x826B</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SEVERITY_NOTIFICATION</span><span class="w"></span>
<span class="mh">0x9146</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SEVERITY_HIGH</span><span class="w"></span>
<span class="mh">0x9147</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SEVERITY_MEDIUM</span><span class="w"></span>
<span class="mh">0x9148</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">GL_DEBUG_SEVERITY_LOW</span><span class="w"></span>
</pre></div>
</div>
<p><strong>OSRT Event Class Values</strong></p>
<p>OS runtime libraries can be traced to gather information about low-level userspace APIs. This traces the system call wrappers and thread synchronization interfaces exposed by the C runtime and POSIX Threads (pthread) libraries. This does not perform a complete runtime library API trace, but instead focuses on the functions that can take a long time to execute, or could potentially cause your thread be unscheduled from the CPU while waiting for an event to complete.</p>
<p>OSRT events may have callchains attached to them, depending on selected profiling settings. In such cases, one can use callchainId column to select relevant callchains from OSRT_CALLCHAINS table</p>
<p>OSRT event class values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">27</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OS_RUNTIME</span><span class="w"></span>
<span class="mi">31</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OS_RUNTIME_START</span><span class="w"></span>
<span class="mi">32</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OS_RUNTIME_FINISH</span><span class="w"></span>
</pre></div>
</div>
<p><strong>DX12 Event Class Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">41</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_API</span><span class="w"></span>
<span class="mi">42</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_WORKLOAD</span><span class="w"></span>
<span class="mi">43</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_START</span><span class="w"></span>
<span class="mi">44</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_FINISH</span><span class="w"></span>
<span class="mi">52</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_DISPLAY</span><span class="w"></span>
<span class="mi">59</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_CREATE_OBJECT</span><span class="w"></span>
</pre></div>
</div>
<p><strong>PIX Event Class Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">65</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX12_DEBUG_API</span><span class="w"></span>
<span class="mi">75</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_DX11_DEBUG_API</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Vulkan Event Class Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">53</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_API</span><span class="w"></span>
<span class="mi">54</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_WORKLOAD</span><span class="w"></span>
<span class="mi">55</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_START</span><span class="w"></span>
<span class="mi">56</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_FINISH</span><span class="w"></span>
<span class="mi">60</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_CREATE_OBJECT</span><span class="w"></span>
<span class="mi">66</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_VULKAN_DEBUG_API</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Vulkan Flags</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">VALID_BIT</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mh">0x00000001</span><span class="w"></span>
<span class="n">CACHE_HIT_BIT</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mh">0x00000002</span><span class="w"></span>
<span class="n">BASE_PIPELINE_ACCELERATION_BIT</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mh">0x00000004</span><span class="w"></span>
</pre></div>
</div>
<p><strong>SLI Event Class Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">62</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_SLI</span><span class="w"></span>
<span class="mi">63</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_SLI_START</span><span class="w"></span>
<span class="mi">64</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_SLI_FINISH</span><span class="w"></span>
</pre></div>
</div>
<p><strong>SLI Transfer Info Values</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_SKIPPED</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_EARLY_PUSH</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_PUSH_FAILED</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_2WAY_OR_PULL</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_PRESENT</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">P2P_DX12_INIT_PUSH_ON_WRITE</span><span class="w"></span>
</pre></div>
</div>
<p><strong>WDDM Event Values</strong></p>
<p>VIDMM operation type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">None</span><span class="w"></span>
<span class="mi">101</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">RestoreSegments</span><span class="w"></span>
<span class="mi">102</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">PurgeSegments</span><span class="w"></span>
<span class="mi">103</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">CleanupPrimary</span><span class="w"></span>
<span class="mi">104</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">AllocatePagingBufferResources</span><span class="w"></span>
<span class="mi">105</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">FreePagingBufferResources</span><span class="w"></span>
<span class="mi">106</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">ReportVidMmState</span><span class="w"></span>
<span class="mi">107</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">RunApertureCoherencyTest</span><span class="w"></span>
<span class="mi">108</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">RunUnmapToDummyPageTest</span><span class="w"></span>
<span class="mi">109</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DeferredCommand</span><span class="w"></span>
<span class="mi">110</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">SuspendMemorySegmentAccess</span><span class="w"></span>
<span class="mi">111</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">ResumeMemorySegmentAccess</span><span class="w"></span>
<span class="mi">112</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">EvictAndFlush</span><span class="w"></span>
<span class="mi">113</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">CommitVirtualAddressRange</span><span class="w"></span>
<span class="mi">114</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">UncommitVirtualAddressRange</span><span class="w"></span>
<span class="mi">115</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DestroyVirtualAddressAllocator</span><span class="w"></span>
<span class="mi">116</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">PageInDevice</span><span class="w"></span>
<span class="mi">117</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">MapContextAllocation</span><span class="w"></span>
<span class="mi">118</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">InitPagingProcessVaSpace</span><span class="w"></span>
<span class="mi">200</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">CloseAllocation</span><span class="w"></span>
<span class="mi">202</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">ComplexLock</span><span class="w"></span>
<span class="mi">203</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">PinAllocation</span><span class="w"></span>
<span class="mi">204</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">FlushPendingGpuAccess</span><span class="w"></span>
<span class="mi">205</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">UnpinAllocation</span><span class="w"></span>
<span class="mi">206</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">MakeResident</span><span class="w"></span>
<span class="mi">207</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Evict</span><span class="w"></span>
<span class="mi">208</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">LockInAperture</span><span class="w"></span>
<span class="mi">209</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">InitContextAllocation</span><span class="w"></span>
<span class="mi">210</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">ReclaimAllocation</span><span class="w"></span>
<span class="mi">211</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DiscardAllocation</span><span class="w"></span>
<span class="mi">212</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">SetAllocationPriority</span><span class="w"></span>
<span class="mi">1000</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">EvictSystemMemoryOfferList</span><span class="w"></span>
</pre></div>
</div>
<p>Paging queue type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">VIDMM_PAGING_QUEUE_TYPE_UMD</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">VIDMM_PAGING_QUEUE_TYPE_Default</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">VIDMM_PAGING_QUEUE_TYPE_Evict</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">VIDMM_PAGING_QUEUE_TYPE_Reclaim</span><span class="w"></span>
</pre></div>
</div>
<p>Packet type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_RENDER_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_DEFERRED_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_SYSTEM_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_MMIOFLIP_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_WAIT_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_SIGNAL_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_DEVICE_COMMAND_BUFFER</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGKETW_SOFTWARE_COMMAND_BUFFER</span><span class="w"></span>
</pre></div>
</div>
<p>Engine type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_OTHER</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_3D</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_VIDEO_DECODE</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_VIDEO_ENCODE</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_VIDEO_PROCESSING</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_SCENE_ASSEMBLY</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_COPY</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_OVERLAY</span><span class="w"></span>
<span class="mi">8</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">DXGK_ENGINE_TYPE_CRYPTO</span><span class="w"></span>
</pre></div>
</div>
<p>DMA interrupt type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">DXGK_INTERRUPT_DMA_COMPLETED</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">DXGK_INTERRUPT_DMA_PREEMPTED</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">DXGK_INTERRUPT_DMA_FAULTED</span><span class="w"></span>
<span class="mi">9</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">DXGK_INTERRUPT_DMA_PAGE_FAULTED</span><span class="w"></span>
</pre></div>
</div>
<p>Queue type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">Queue_Packet</span><span class="w"></span>
<span class="mi">1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">Dma_Packet</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">Paging_Queue_Packet</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Driver Events</strong></p>
<p>Load balance event type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">LoadBalanceEvent_GPU</span><span class="w"></span>
<span class="mi">8</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">LoadBalanceEvent_CPU</span><span class="w"></span>
<span class="mi">21</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">LoadBalanceMasterEvent_GPU</span><span class="w"></span>
<span class="mi">22</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">LoadBalanceMasterEvent_CPU</span><span class="w"></span>
</pre></div>
</div>
<p><strong>OpenMP Events</strong></p>
<p>OpenMP event class values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">78</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OPENMP</span><span class="w"></span>
<span class="mi">79</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OPENMP_START</span><span class="w"></span>
<span class="mi">80</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">TRACE_PROCESS_EVENT_OPENMP_FINISH</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP event kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">15</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK_CREATE</span><span class="w"></span>
<span class="mi">16</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK_SCHEDULE</span><span class="w"></span>
<span class="mi">17</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_CANCEL</span><span class="w"></span>
<span class="mi">20</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MUTEX_RELEASED</span><span class="w"></span>
<span class="mi">21</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_LOCK_INIT</span><span class="w"></span>
<span class="mi">22</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_LOCK_DESTROY</span><span class="w"></span>
<span class="mi">25</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_DISPATCH</span><span class="w"></span>
<span class="mi">26</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_FLUSH</span><span class="w"></span>
<span class="mi">27</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_THREAD</span><span class="w"></span>
<span class="mi">28</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_PARALLEL</span><span class="w"></span>
<span class="mi">29</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_SYNC_REGION_WAIT</span><span class="w"></span>
<span class="mi">30</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_SYNC_REGION</span><span class="w"></span>
<span class="mi">31</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_TASK</span><span class="w"></span>
<span class="mi">32</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MASTER</span><span class="w"></span>
<span class="mi">33</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_REDUCTION</span><span class="w"></span>
<span class="mi">34</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_MUTEX_WAIT</span><span class="w"></span>
<span class="mi">35</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_CRITICAL_SECTION</span><span class="w"></span>
<span class="mi">36</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OPENMP_EVENT_KIND_WORKSHARE</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP thread type values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OpenMP</span><span class="w"> </span><span class="n">Initial</span><span class="w"> </span><span class="n">Thread</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OpenMP</span><span class="w"> </span><span class="n">Worker</span><span class="w"> </span><span class="n">Thread</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">OpenMP</span><span class="w"> </span><span class="n">Internal</span><span class="w"> </span><span class="n">Thread</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Unknown</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP sync region kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Barrier</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Implicit</span><span class="w"> </span><span class="n">barrier</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Explicit</span><span class="w"> </span><span class="n">barrier</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Implementation</span><span class="o">-</span><span class="n">dependent</span><span class="w"> </span><span class="n">barrier</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Taskwait</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Taskgroup</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP task kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Initial</span><span class="w"> </span><span class="n">task</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Implicit</span><span class="w"> </span><span class="n">task</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Explicit</span><span class="w"> </span><span class="n">task</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP prior task status values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Task</span><span class="w"> </span><span class="n">completed</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Task</span><span class="w"> </span><span class="n">yielded</span><span class="w"> </span><span class="n">to</span><span class="w"> </span><span class="n">another</span><span class="w"> </span><span class="n">task</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Task</span><span class="w"> </span><span class="n">was</span><span class="w"> </span><span class="n">cancelled</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Task</span><span class="w"> </span><span class="n">was</span><span class="w"> </span><span class="n">switched</span><span class="w"> </span><span class="n">out</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">other</span><span class="w"> </span><span class="n">reasons</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP mutex kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Waiting</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">lock</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Testing</span><span class="w"> </span><span class="n">lock</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Waiting</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">nested</span><span class="w"> </span><span class="n">lock</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Tesing</span><span class="w"> </span><span class="n">nested</span><span class="w"> </span><span class="n">lock</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Waitng</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">entering</span><span class="w"> </span><span class="n">critical</span><span class="w"> </span><span class="n">section</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Waiting</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">entering</span><span class="w"> </span><span class="n">atomic</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Waiting</span><span class="w"> </span><span class="k">for</span><span class="w"> </span><span class="n">entering</span><span class="w"> </span><span class="n">ordered</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP critical section kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Critical</span><span class="w"> </span><span class="n">section</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Atomic</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Ordered</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP workshare kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Loop</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Sections</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">3</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Single</span><span class="w"> </span><span class="n">region</span><span class="w"> </span><span class="p">(</span><span class="n">executor</span><span class="p">)</span><span class="w"></span>
<span class="mi">4</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Single</span><span class="w"> </span><span class="n">region</span><span class="w"> </span><span class="p">(</span><span class="n">waiting</span><span class="p">)</span><span class="w"></span>
<span class="mi">5</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Workshare</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">6</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Distrubute</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
<span class="mi">7</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Taskloop</span><span class="w"> </span><span class="n">region</span><span class="w"></span>
</pre></div>
</div>
<p>OpenMP dispatch kind values</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="mi">1</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Iteration</span><span class="w"></span>
<span class="mi">2</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="n">Section</span><span class="w"></span>
</pre></div>
</div>
</section>
<section id="common-sqlite-examples">
<h4>Common SQLite Examples<a class="headerlink" href="#common-sqlite-examples" title="Permalink to this heading"></a></h4>
<p><strong>Common Helper Commands</strong></p>
<p>When utilizing sqlite3 command line tool, it’s helpful to have data printed as named columns, this can be done with:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>.mode column
.headers on
</pre></div>
</div>
<p>Default column width is determined by the data in the first row of results. If this doesn’t work out well, you can specify widths manually.</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>.width 10 20 50
</pre></div>
</div>
<p><strong>Obtaining Sample Report</strong></p>
<p>CLI interface of Nsight Systems was used to profile radixSortThrust CUDA sample, then the resulting .nsys-rep file was exported using the nsys export.</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys profile --trace=cuda,osrt radixSortThrust
nsys export --type sqlite report1.nsys-rep
</pre></div>
</div>
<p><strong>Serialized Process and Thread Identifiers</strong></p>
<p>Nsight Systems stores identifiers where events originated in serialized form. For events that have globalTid or globalPid fields exported, use the following code to extract numeric TID and PID.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">TID</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="k">TABLE_NAME</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Note: globalTid field includes both TID and PID values, while globalPid only containes the PID value.</p>
<p><strong>Correlate CUDA Kernel Launches With CUDA API Kernel Launches</strong></p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">kernelName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">kernelName</span><span class="w"> </span><span class="o">=</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"></span>
<span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">cuda_gpu</span><span class="w"></span>
<span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">cuda_gpu</span><span class="p">.</span><span class="n">shortName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="w"></span>
<span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="n">correlationId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">cuda_gpu</span><span class="p">.</span><span class="n">correlationId</span><span class="p">);</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="o">=</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
</pre></div>
</div>
<p>Select 10 longest CUDA API ranges that resulted in kernel execution.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">name</span><span class="p">,</span><span class="w"> </span><span class="n">kernelName</span><span class="p">,</span><span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">kernelName</span><span class="w"> </span><span class="k">IS</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">10</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>name kernelName start end
---------------------- ----------------------- ---------- ----------
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 658863435 658868490
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 609755015 609760075
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 632683286 632688349
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 606495356 606500439
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 603114486 603119586
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 802729785 802734906
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 593381170 593386294
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 658759955 658765090
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 681549917 681555059
cudaLaunchKernel_v7000 RadixSortScanBinsKernel 717812527 717817671
</pre></div>
</div>
<p><strong>Remove Ranges Overlapping With Overhead</strong></p>
<p>Use the this query to count CUDA API ranges overlapping with the overhead ones.</p>
<p>Replace “SELECT COUNT(*)” with “DELETE” to remove such ranges.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="k">COUNT</span><span class="p">(</span><span class="o">*</span><span class="p">)</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">rowid</span><span class="w"> </span><span class="k">IN</span><span class="w"></span>
<span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="n">cuda</span><span class="p">.</span><span class="n">rowid</span><span class="w"></span>
<span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">PROFILER_OVERHEAD</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">overhead</span><span class="w"></span>
<span class="w"> </span><span class="k">INNER</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">cuda</span><span class="w"> </span><span class="k">ON</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="n">cuda</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">BETWEEN</span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">and</span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">end</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="k">OR</span><span class="w"> </span><span class="p">(</span><span class="n">cuda</span><span class="p">.</span><span class="k">end</span><span class="w"> </span><span class="k">BETWEEN</span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">and</span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">end</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="k">OR</span><span class="w"> </span><span class="p">(</span><span class="n">cuda</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="o"><</span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">cuda</span><span class="p">.</span><span class="k">end</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="n">overhead</span><span class="p">.</span><span class="k">end</span><span class="p">)</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>COUNT(*)
----------
1095
</pre></div>
</div>
<p><strong>Find CUDA API Calls That Resulted in Original Graph Node Creation.</strong></p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">graph</span><span class="p">.</span><span class="n">graphNodeId</span><span class="p">,</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="n">graph</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">graphStart</span><span class="p">,</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="k">end</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="n">globalTid</span><span class="p">,</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="n">correlationId</span><span class="p">,</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="n">globalTid</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">where</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="n">nameId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">id</span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">name</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">api</span><span class="w"></span>
<span class="k">JOIN</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="n">graphNodeId</span><span class="p">,</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="k">from</span><span class="w"> </span><span class="n">CUDA_GRAPH_EVENTS</span><span class="w"></span>
<span class="w"> </span><span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">graphNodeId</span><span class="w"></span>
<span class="w"> </span><span class="k">HAVING</span><span class="w"> </span><span class="k">COUNT</span><span class="p">(</span><span class="n">originalGraphNodeId</span><span class="p">)</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mi">0</span><span class="w"></span>
<span class="w"> </span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">graph</span><span class="w"></span>
<span class="k">ON</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">graph</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="o"><</span><span class="w"> </span><span class="n">graph</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">api</span><span class="p">.</span><span class="k">end</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="n">graph</span><span class="p">.</span><span class="k">start</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">graphNodeId</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>graphNodeId start graphStart end globalTid correlationId globalTid name
----------- ---------- ---------- ---------- --------------- ------------- --------------- -----------------------------
1 584366518 584378040 584379102 281560221750233 109 281560221750233 cudaGraphAddMemcpyNode_v10000
2 584379402 584382428 584383139 281560221750233 110 281560221750233 cudaGraphAddMemsetNode_v10000
3 584390663 584395352 584396053 281560221750233 111 281560221750233 cudaGraphAddKernelNode_v10000
4 584396314 584397857 584398438 281560221750233 112 281560221750233 cudaGraphAddMemsetNode_v10000
5 584398759 584400311 584400812 281560221750233 113 281560221750233 cudaGraphAddKernelNode_v10000
6 584401083 584403047 584403527 281560221750233 114 281560221750233 cudaGraphAddMemcpyNode_v10000
7 584403928 584404920 584405491 281560221750233 115 281560221750233 cudaGraphAddHostNode_v10000
29 632107852 632117921 632121407 281560221750233 144 281560221750233 cudaMemcpyAsync_v3020
30 632122168 632125545 632127989 281560221750233 145 281560221750233 cudaMemsetAsync_v3020
31 632131546 632133339 632135584 281560221750233 147 281560221750233 cudaMemsetAsync_v3020
34 632162514 632167393 632169297 281560221750233 151 281560221750233 cudaMemcpyAsync_v3020
35 632170068 632173334 632175388 281560221750233 152 281560221750233 cudaLaunchHostFunc_v10000
</pre></div>
</div>
<p><strong>Backtraces for OSRT Ranges</strong></p>
<p>Adding text columns makes results of the query below more human-readable.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OSRT_API</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">OSRT_API</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">OSRT_API</span><span class="p">.</span><span class="n">nameId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">moduleName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">moduleName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
</pre></div>
</div>
<p>Print backtrace of the longest OSRT range</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">TID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="k">end</span><span class="p">,</span><span class="w"> </span><span class="n">name</span><span class="p">,</span><span class="w"> </span><span class="n">callchainId</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="p">,</span><span class="w"> </span><span class="n">symbolName</span><span class="p">,</span><span class="w"> </span><span class="n">moduleName</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">OSRT_API</span><span class="w"> </span><span class="k">LEFT</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">callchainId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">OSRT_CALLCHAINS</span><span class="p">.</span><span class="n">id</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="n">OSRT_API</span><span class="p">.</span><span class="n">rowid</span><span class="w"> </span><span class="k">IN</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">rowid</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">OSRT_API</span><span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="k">DESC</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">1</span><span class="p">)</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">10</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>PID TID start end name callchainId stackDepth symbolName moduleName
---------- ---------- ---------- ---------- ---------------------- ----------- ---------- ------------------------------ ----------------------------------------
19163 19176 360897690 860966851 pthread_cond_timedwait 88 0 pthread_cond_timedwait@GLIBC_2 /lib/x86_64-linux-gnu/libpthread-2.27.so
19163 19176 360897690 860966851 pthread_cond_timedwait 88 1 0x7fbc983b7227 /usr/lib/x86_64-linux-gnu/libcuda.so.418
19163 19176 360897690 860966851 pthread_cond_timedwait 88 2 0x7fbc9835d5c7 /usr/lib/x86_64-linux-gnu/libcuda.so.418
19163 19176 360897690 860966851 pthread_cond_timedwait 88 3 0x7fbc983b64a8 /usr/lib/x86_64-linux-gnu/libcuda.so.418
19163 19176 360897690 860966851 pthread_cond_timedwait 88 4 start_thread /lib/x86_64-linux-gnu/libpthread-2.27.so
19163 19176 360897690 860966851 pthread_cond_timedwait 88 5 __clone /lib/x86_64-linux-gnu/libc-2.27.so
</pre></div>
</div>
<p>Profiled processes output streams</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">filename</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">filename</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="p">.</span><span class="n">filenameId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">content</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">content</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="p">.</span><span class="n">contentId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
</pre></div>
</div>
<p>Select all collected stdout and stderr streams.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">select</span><span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"> </span><span class="n">filename</span><span class="p">,</span><span class="w"> </span><span class="n">content</span><span class="w"> </span><span class="k">from</span><span class="w"> </span><span class="n">ProcessStreams</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>PID filename content
---------- ------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------
19163 /tmp/nvidia/nsight_systems/streams/pid_19163_stdout.log /home/user_name/NVIDIA_CUDA-10.1_Samples/6_Advanced/radixSortThrust/radixSortThrust Starting...
GPU Device 0: "Quadro P2000" with compute capability 6.1
Sorting 1048576 32-bit unsigned int keys and values
radixSortThrust, Throughput = 401.0872 MElements/s, Time = 0.00261 s, Size = 1048576 elements
Test passed
19163 /tmp/nvidia/nsight_systems/streams/pid_19163_stderr.log
</pre></div>
</div>
<p><strong>Thread Summary</strong></p>
<p>Please note, that Nsight Systems applies additional logic during sampling events processing to work around lost events. This means that the results of the below query might differ slightly from the ones shown in “Analysis summary” tab.</p>
<p>Thread summary calculated using CPU cycles (when available).</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">TID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">ROUND</span><span class="p">(</span><span class="mi">100</span><span class="p">.</span><span class="mi">0</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="k">SUM</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="o">/</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="k">SUM</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="w"></span>
<span class="w"> </span><span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x100</span><span class="w"></span>
<span class="w"> </span><span class="p">),</span><span class="w"></span>
<span class="w"> </span><span class="mi">2</span><span class="w"></span>
<span class="w"> </span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">CPU_utilization</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="o">=</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">ThreadNames</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">ThreadNames</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="p">.</span><span class="n">globalTid</span><span class="w"></span>
<span class="w"> </span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">thread_name</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">globalTid</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">CPU_utilization</span><span class="w"> </span><span class="k">DESC</span><span class="w"></span>
<span class="k">LIMIT</span><span class="w"> </span><span class="mi">10</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>PID TID CPU_utilization thread_name
---------- ---------- --------------- ---------------
19163 19163 98.4 radixSortThrust
19163 19168 1.35 CUPTI worker th
19163 19166 0.25 [NS]
</pre></div>
</div>
<p>Thread running time may be calculated using scheduling data, when PMU counter data was not collected.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">CREATE</span><span class="w"> </span><span class="k">INDEX</span><span class="w"> </span><span class="n">sched_start</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">SCHED_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="k">start</span><span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CPU_USAGE</span><span class="w"> </span><span class="k">AS</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">globalTid</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">ThreadNames</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">ThreadNames</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="n">globalTid</span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">nameId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">sum</span><span class="p">(</span><span class="k">second</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="k">start</span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">total_duration</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">count</span><span class="p">()</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">ranges_count</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">SCHED_EVENTS</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="k">first</span><span class="w"></span>
<span class="k">LEFT</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">SCHED_EVENTS</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="k">second</span><span class="w"></span>
<span class="k">ON</span><span class="w"> </span><span class="k">second</span><span class="p">.</span><span class="n">rowid</span><span class="w"> </span><span class="o">=</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="n">rowid</span><span class="w"></span>
<span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">SCHED_EVENTS</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="n">globalTid</span><span class="w"></span>
<span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="k">ASC</span><span class="w"></span>
<span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">1</span><span class="w"></span>
<span class="w"> </span><span class="p">)</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="n">isSchedIn</span><span class="w"> </span><span class="o">!=</span><span class="w"> </span><span class="mi">0</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">first</span><span class="p">.</span><span class="n">globalTid</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">total_duration</span><span class="w"> </span><span class="k">DESC</span><span class="p">;</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">TID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">where</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">id</span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">thread_name</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">ROUND</span><span class="p">(</span><span class="mi">100</span><span class="p">.</span><span class="mi">0</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="n">total_duration</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="k">SUM</span><span class="p">(</span><span class="n">total_duration</span><span class="p">)</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">CPU_USAGE</span><span class="p">),</span><span class="w"> </span><span class="mi">2</span><span class="p">)</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">CPU_utilization</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">CPU_USAGE</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">CPU_utilization</span><span class="w"> </span><span class="k">DESC</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>PID TID thread_name CPU_utilization
---------- ---------- --------------- ---------------
19163 19163 radixSortThrust 93.74
19163 19169 radixSortThrust 3.22
19163 19168 CUPTI worker th 2.46
19163 19166 [NS] 0.44
19163 19172 radixSortThrust 0.07
19163 19167 [NS Comms] 0.05
19163 19176 radixSortThrust 0.02
19163 19170 radixSortThrust 0.0
</pre></div>
</div>
<p><strong>Function Table</strong></p>
<p>These examples demonstrate how to calculate Flat and BottomUp (for top level only) views statistics.</p>
<p>To set up:</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">moduleName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">moduleName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">module</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
</pre></div>
</div>
<p>To get flat view:</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">symbolName</span><span class="p">,</span><span class="w"> </span><span class="n">moduleName</span><span class="p">,</span><span class="w"> </span><span class="n">ROUND</span><span class="p">(</span><span class="mi">100</span><span class="p">.</span><span class="mi">0</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="k">sum</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="o">/</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="k">SUM</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="p">),</span><span class="w"> </span><span class="mi">2</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">flatTimePercentage</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"></span>
<span class="k">LEFT</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="p">.</span><span class="n">id</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="p">.</span><span class="n">id</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">symbol</span><span class="p">,</span><span class="w"> </span><span class="n">module</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">flatTimePercentage</span><span class="w"> </span><span class="k">DESC</span><span class="w"></span>
<span class="k">LIMIT</span><span class="w"> </span><span class="mi">5</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>To get BottomUp view (top level only):</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">symbolName</span><span class="p">,</span><span class="w"> </span><span class="n">moduleName</span><span class="p">,</span><span class="w"> </span><span class="n">ROUND</span><span class="p">(</span><span class="mi">100</span><span class="p">.</span><span class="mi">0</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="k">sum</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="o">/</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="k">SUM</span><span class="p">(</span><span class="n">cpuCycles</span><span class="p">)</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="p">),</span><span class="w"> </span><span class="mi">2</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">selfTimePercentage</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="w"></span>
<span class="k">LEFT</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">SAMPLING_CALLCHAINS</span><span class="p">.</span><span class="n">id</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">COMPOSITE_EVENTS</span><span class="p">.</span><span class="n">id</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="mi">0</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">symbol</span><span class="p">,</span><span class="w"> </span><span class="n">module</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">selfTimePercentage</span><span class="w"> </span><span class="k">DESC</span><span class="w"></span>
<span class="k">LIMIT</span><span class="w"> </span><span class="mi">5</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>symbolName moduleName flatTimePercentage
----------- ----------- ------------------
[Max depth] [Max depth] 99.92
thrust::zip /home/user_ 24.17
thrust::zip /home/user_ 24.17
thrust::det /home/user_ 24.17
thrust::det /home/user_ 24.17
symbolName moduleName selfTimePercentage
-------------- ------------------------------------------- ------------------
0x7fbc984982b6 /usr/lib/x86_64-linux-gnu/libcuda.so.418.39 5.29
0x7fbc982d0010 /usr/lib/x86_64-linux-gnu/libcuda.so.418.39 2.81
thrust::iterat /home/user_name/NVIDIA_CUDA-10.1_Samples/6_ 2.23
thrust::iterat /home/user_name/NVIDIA_CUDA-10.1_Samples/6_ 1.55
void thrust::i /home/user_name/NVIDIA_CUDA-10.1_Samples/6_ 1.55
</pre></div>
</div>
<p><strong>DX12 API Frame Duration Histogram</strong></p>
<p>The example demonstrates how to calculate DX12 CPU frames durartion and construct a histogram out of it.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">CREATE</span><span class="w"> </span><span class="k">INDEX</span><span class="w"> </span><span class="n">DX12_API_ENDTS</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">DX12_API</span><span class="w"> </span><span class="p">(</span><span class="k">end</span><span class="p">);</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="n">TEMP</span><span class="w"> </span><span class="k">VIEW</span><span class="w"> </span><span class="n">DX12_API_FPS</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">DX12_API</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="k">outer</span><span class="p">.</span><span class="k">end</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="ss">"IDXGISwapChain::Present"</span><span class="p">)</span><span class="w"></span>
<span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="k">ASC</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">1</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="k">end</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">DX12_API</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="k">outer</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">nameId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">id</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="ss">"IDXGISwapChain::Present"</span><span class="p">)</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">end</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Number of frames with a duration of [X, X + 1) milliseconds.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="k">CAST</span><span class="p">((</span><span class="k">end</span><span class="w"> </span><span class="o">-</span><span class="w"> </span><span class="k">start</span><span class="p">)</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">1000000</span><span class="p">.</span><span class="mi">0</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="nb">INT</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">duration_ms</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">count</span><span class="p">(</span><span class="o">*</span><span class="p">)</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">DX12_API_FPS</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="k">IS</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">duration_ms</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">duration_ms</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>duration_ms count(*)
----------- ----------
3 1
4 2
5 7
6 153
7 19
8 116
9 16
10 8
11 2
12 2
13 1
14 4
16 3
17 2
18 1
</pre></div>
</div>
<p><strong>GPU Context Switch Events Enumeration</strong></p>
<p>GPU context duration is between first BEGIN and a matching END event.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="p">(</span><span class="k">CASE</span><span class="w"> </span><span class="n">tag</span><span class="w"> </span><span class="k">WHEN</span><span class="w"> </span><span class="mi">8</span><span class="w"> </span><span class="k">THEN</span><span class="w"> </span><span class="ss">"BEGIN"</span><span class="w"> </span><span class="k">WHEN</span><span class="w"> </span><span class="mi">7</span><span class="w"> </span><span class="k">THEN</span><span class="w"> </span><span class="ss">"END"</span><span class="w"> </span><span class="k">END</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">tag</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">globalPid</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">vmId</span><span class="p">,</span><span class="w"> </span><span class="n">seqNo</span><span class="p">,</span><span class="w"> </span><span class="n">contextId</span><span class="p">,</span><span class="w"> </span><span class="k">timestamp</span><span class="p">,</span><span class="w"> </span><span class="n">gpuId</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">GPU_CONTEXT_SWITCH_EVENTS</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="n">tag</span><span class="w"> </span><span class="k">in</span><span class="w"> </span><span class="p">(</span><span class="mi">7</span><span class="p">,</span><span class="w"> </span><span class="mi">8</span><span class="p">)</span><span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">seqNo</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">10</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>tag PID vmId seqNo contextId timestamp gpuId
---------- ---------- ---------- ---------- ---------- ---------- ----------
BEGIN 23371 0 0 1048578 56759171 0
BEGIN 23371 0 1 1048578 56927765 0
BEGIN 23371 0 3 1048578 63799379 0
END 23371 0 4 1048578 63918806 0
BEGIN 19397 0 5 1048577 64014692 0
BEGIN 19397 0 6 1048577 64250369 0
BEGIN 19397 0 8 1048577 1918310004 0
END 19397 0 9 1048577 1918521098 0
BEGIN 19397 0 10 1048577 2024164744 0
BEGIN 19397 0 11 1048577 2024358650 0
</pre></div>
</div>
<p><strong>Resolve NVTX Category Name</strong></p>
<p>The example demonstrates how to resolve NVTX category name for NVTX marks and ranges.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">WITH</span><span class="w"></span>
<span class="w"> </span><span class="n">event</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="o">*</span><span class="w"></span>
<span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">eventType</span><span class="w"> </span><span class="k">IN</span><span class="w"> </span><span class="p">(</span><span class="mi">34</span><span class="p">,</span><span class="w"> </span><span class="mi">59</span><span class="p">,</span><span class="w"> </span><span class="mi">60</span><span class="p">)</span><span class="w"> </span><span class="c1">-- mark, push/pop, start/end</span>
<span class="w"> </span><span class="p">),</span><span class="w"></span>
<span class="w"> </span><span class="n">category</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="n">category</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">domainId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="nb">text</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">categoryName</span><span class="w"></span>
<span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">eventType</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="mi">33</span><span class="w"> </span><span class="c1">-- new category</span>
<span class="w"> </span><span class="p">)</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">end</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">globalTid</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">eventType</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">domainId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">category</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="n">categoryName</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="nb">text</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">event</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">category</span><span class="w"> </span><span class="k">USING</span><span class="w"> </span><span class="p">(</span><span class="n">category</span><span class="p">,</span><span class="w"> </span><span class="n">domainId</span><span class="p">)</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">start</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>start end globalTid eventType domainId category categoryName text
---------- ---------- --------------- ---------- ---------- ---------- ------------------------- ----------------
18281150 18311960 281534938484214 59 0 1 FirstCategoryUnderDefault Push Pop Range A
18288187 18306674 281534938484214 59 0 2 SecondCategoryUnderDefaul Push Pop Range B
18294247 281534938484214 34 0 1 FirstCategoryUnderDefault Mark A
18300034 281534938484214 34 0 2 SecondCategoryUnderDefaul Mark B
18345546 18372595 281534938484214 60 1 1 FirstCategoryUnderMyDomai Start End Range
18352924 18378342 281534938484214 60 1 2 SecondCategoryUnderMyDoma Start End Range
18359634 281534938484214 34 1 1 FirstCategoryUnderMyDomai Mark A
18365448 281534938484214 34 1 2 SecondCategoryUnderMyDoma Mark B
</pre></div>
</div>
<p><strong>Rename CUDA Kernels with NVTX</strong></p>
<p>The example demonstrates how to map innermost NVTX push-pop range to a matching CUDA kernel run.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">nvtxRange</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">CREATE</span><span class="w"> </span><span class="k">INDEX</span><span class="w"> </span><span class="n">nvtx_start</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="w"> </span><span class="p">(</span><span class="k">start</span><span class="p">);</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">nvtxRange</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="nb">text</span><span class="w"></span>
<span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">ON</span><span class="w"></span>
<span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="n">eventType</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="mi">59</span><span class="w"> </span><span class="k">AND</span><span class="w"></span>
<span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="n">globalTid</span><span class="w"> </span><span class="k">AND</span><span class="w"></span>
<span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="o"><=</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">AND</span><span class="w"></span>
<span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="k">end</span><span class="w"> </span><span class="o">>=</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="k">end</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"></span>
<span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="p">.</span><span class="n">correlationId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="n">correlationId</span><span class="w"></span>
<span class="w"> </span><span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">NVTX_EVENTS</span><span class="p">.</span><span class="k">start</span><span class="w"> </span><span class="k">DESC</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">1</span><span class="w"></span>
<span class="p">);</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="k">end</span><span class="p">,</span><span class="w"> </span><span class="n">globalPid</span><span class="p">,</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">value</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="n">shortName</span><span class="p">,</span><span class="w"> </span><span class="n">nvtxRange</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_KERNEL</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">shortName</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">id</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">6</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>start end globalPid shortName nvtxRange
---------- ---------- ----------------- ------------- ----------
526545376 526676256 72057700439031808 MatrixMulCUDA
526899648 527030368 72057700439031808 MatrixMulCUDA Add
527031648 527162272 72057700439031808 MatrixMulCUDA Add
527163584 527294176 72057700439031808 MatrixMulCUDA My Kernel
527296160 527426592 72057700439031808 MatrixMulCUDA My Range
527428096 527558656 72057700439031808 MatrixMulCUDA
</pre></div>
</div>
<p><strong>Select CUDA Calls With Backtraces</strong></p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">name</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="p">.</span><span class="n">nameId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">ALTER</span><span class="w"> </span><span class="k">TABLE</span><span class="w"> </span><span class="n">CUDA_CALLCHAINS</span><span class="w"> </span><span class="k">ADD</span><span class="w"> </span><span class="k">COLUMN</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="nb">TEXT</span><span class="p">;</span><span class="w"></span>
<span class="k">UPDATE</span><span class="w"> </span><span class="n">CUDA_CALLCHAINS</span><span class="w"> </span><span class="k">SET</span><span class="w"> </span><span class="n">symbolName</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="k">SELECT</span><span class="w"> </span><span class="n">value</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">StringIds</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">symbol</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">StringIds</span><span class="p">.</span><span class="n">id</span><span class="p">);</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"> </span><span class="n">globalTid</span><span class="w"> </span><span class="o">%</span><span class="w"> </span><span class="mi">0</span><span class="n">x1000000</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">TID</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="k">start</span><span class="p">,</span><span class="w"> </span><span class="k">end</span><span class="p">,</span><span class="w"> </span><span class="n">name</span><span class="p">,</span><span class="w"> </span><span class="n">callchainId</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="p">,</span><span class="w"> </span><span class="n">symbolName</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">CUDA_CALLCHAINS</span><span class="w"> </span><span class="k">JOIN</span><span class="w"> </span><span class="n">CUPTI_ACTIVITY_KIND_RUNTIME</span><span class="w"> </span><span class="k">ON</span><span class="w"> </span><span class="n">callchainId</span><span class="w"> </span><span class="o">==</span><span class="w"> </span><span class="n">CUDA_CALLCHAINS</span><span class="p">.</span><span class="n">id</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">callchainId</span><span class="p">,</span><span class="w"> </span><span class="n">stackDepth</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">11</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>TID start end name callchainId stackDepth symbolName
---------- ---------- ---------- ------------- ----------- ---------- --------------
11928 168976467 169077826 cuMemAlloc_v2 1 0 0x7f13c44f02ab
11928 168976467 169077826 cuMemAlloc_v2 1 1 0x7f13c44f0b8f
11928 168976467 169077826 cuMemAlloc_v2 1 2 0x7f13c44f3719
11928 168976467 169077826 cuMemAlloc_v2 1 3 cuMemAlloc_v2
11928 168976467 169077826 cuMemAlloc_v2 1 4 cudart::driver
11928 168976467 169077826 cuMemAlloc_v2 1 5 cudart::cudaAp
11928 168976467 169077826 cuMemAlloc_v2 1 6 cudaMalloc
11928 168976467 169077826 cuMemAlloc_v2 1 7 cudaError cuda
11928 168976467 169077826 cuMemAlloc_v2 1 8 main
11928 168976467 169077826 cuMemAlloc_v2 1 9 __libc_start_m
11928 168976467 169077826 cuMemAlloc_v2 1 10 _start
</pre></div>
</div>
<p><strong>SLI Peer-to-Peer Query</strong></p>
<p>The example demonstrates how to query SLI Peer-to-Peer events with resource size greater than value and within a time range sorted by resource size descending.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="o">*</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">SLI_P2P</span><span class="w"></span>
<span class="k">WHERE</span><span class="w"> </span><span class="n">resourceSize</span><span class="w"> </span><span class="o"><</span><span class="w"> </span><span class="mi">98304</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="k">start</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="mi">1568063100</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="k">end</span><span class="w"> </span><span class="o"><</span><span class="w"> </span><span class="mi">1579468901</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">resourceSize</span><span class="w"> </span><span class="k">DESC</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>start end eventClass globalTid gpu frameId transferSkipped srcGpu dstGpu numSubResources resourceSize subResourceIdx smplWidth smplHeight smplDepth bytesPerElement dxgiFormat logSurfaceNames transferInfo isEarlyPushManagedByNvApi useAsyncP2pForResolve transferFuncName regimeName debugName bindType
---------- ---------- ---------- ----------------- ---------- ---------- --------------- ---------- ---------- --------------- ------------ -------------- ---------- ---------- ---------- --------------- ---------- --------------- ------------ ------------------------- --------------------- ---------------- ---------- ---------- ----------
1570351100 1570351101 62 72057698056667136 0 771 0 256 512 1 1048576 0 256 256 1 16 2 3 0 0
1570379300 1570379301 62 72057698056667136 0 771 0 256 512 1 1048576 0 64 64 64 4 31 3 0 0
1572316400 1572316401 62 72057698056667136 0 773 0 256 512 1 1048576 0 256 256 1 16 2 3 0 0
1572345400 1572345401 62 72057698056667136 0 773 0 256 512 1 1048576 0 64 64 64 4 31 3 0 0
1574734300 1574734301 62 72057698056667136 0 775 0 256 512 1 1048576 0 256 256 1 16 2 3 0 0
1574767200 1574767201 62 72057698056667136 0 775 0 256 512 1 1048576 0 64 64 64 4 31 3 0 0
</pre></div>
</div>
<p><strong>Generic Events</strong></p>
<p>Syscall usage histogram by PID:</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">json_extract</span><span class="p">(</span><span class="k">data</span><span class="p">,</span><span class="w"> </span><span class="s1">'$.common_pid'</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">PID</span><span class="p">,</span><span class="w"> </span><span class="k">count</span><span class="p">(</span><span class="o">*</span><span class="p">)</span><span class="w"> </span><span class="k">AS</span><span class="w"> </span><span class="n">total</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">GENERIC_EVENTS</span><span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">PID</span><span class="w"> </span><span class="k">IS</span><span class="w"> </span><span class="k">NOT</span><span class="w"> </span><span class="k">NULL</span><span class="w"> </span><span class="k">AND</span><span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="w"></span>
<span class="w"> </span><span class="k">SELECT</span><span class="w"> </span><span class="n">typeId</span><span class="w"> </span><span class="k">FROM</span><span class="w"> </span><span class="n">GENERIC_EVENT_TYPES</span><span class="w"></span>
<span class="w"> </span><span class="k">WHERE</span><span class="w"> </span><span class="n">json_extract</span><span class="p">(</span><span class="k">data</span><span class="p">,</span><span class="w"> </span><span class="s1">'$.Name'</span><span class="p">)</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="ss">"raw_syscalls:sys_enter"</span><span class="p">)</span><span class="w"></span>
<span class="k">GROUP</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">PID</span><span class="w"></span>
<span class="k">ORDER</span><span class="w"> </span><span class="k">BY</span><span class="w"> </span><span class="n">total</span><span class="w"> </span><span class="k">DESC</span><span class="w"></span>
<span class="k">LIMIT</span><span class="w"> </span><span class="mi">10</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>PID total
---------- ----------
5551 32811
9680 3988
4328 1477
9564 1246
4376 1204
4377 1167
4357 656
4355 655
4356 640
4354 633
</pre></div>
</div>
<p><strong>Fetching Generic Events in JSON Format</strong></p>
<p>Text and JSON export modes don’t include generic events. Use the below queries (without LIMIT clause) to extract JSON lines representation of generic events, types and sources.</p>
<div class="highlight-sql notranslate"><div class="highlight"><pre><span></span><span class="k">SELECT</span><span class="w"> </span><span class="n">json_insert</span><span class="p">(</span><span class="s1">'{}'</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.sourceId'</span><span class="p">,</span><span class="w"> </span><span class="n">sourceId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.data'</span><span class="p">,</span><span class="w"> </span><span class="n">json</span><span class="p">(</span><span class="k">data</span><span class="p">)</span><span class="w"></span>
<span class="p">)</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">GENERIC_EVENT_SOURCES</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">2</span><span class="p">;</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"> </span><span class="n">json_insert</span><span class="p">(</span><span class="s1">'{}'</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.typeId'</span><span class="p">,</span><span class="w"> </span><span class="n">typeId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.sourceId'</span><span class="p">,</span><span class="w"> </span><span class="n">sourceId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.data'</span><span class="p">,</span><span class="w"> </span><span class="n">json</span><span class="p">(</span><span class="k">data</span><span class="p">)</span><span class="w"></span>
<span class="p">)</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">GENERIC_EVENT_TYPES</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">2</span><span class="p">;</span><span class="w"></span>
<span class="k">SELECT</span><span class="w"> </span><span class="n">json_insert</span><span class="p">(</span><span class="s1">'{}'</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.rawTimestamp'</span><span class="p">,</span><span class="w"> </span><span class="n">rawTimestamp</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.timestamp'</span><span class="p">,</span><span class="w"> </span><span class="k">timestamp</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.typeId'</span><span class="p">,</span><span class="w"> </span><span class="n">typeId</span><span class="p">,</span><span class="w"></span>
<span class="w"> </span><span class="s1">'$.data'</span><span class="p">,</span><span class="w"> </span><span class="n">json</span><span class="p">(</span><span class="k">data</span><span class="p">)</span><span class="w"></span>
<span class="p">)</span><span class="w"></span>
<span class="k">FROM</span><span class="w"> </span><span class="n">GENERIC_EVENTS</span><span class="w"> </span><span class="k">LIMIT</span><span class="w"> </span><span class="mi">2</span><span class="p">;</span><span class="w"></span>
</pre></div>
</div>
<p>Results:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>json_insert('{}',
'$.sourceId', sourceId,
'$.data', json(data)
)
---------------------------------------------------------------------------------------------------------------
{"sourceId":72057602627862528,"data":{"Name":"FTrace","TimeSource":"ClockMonotonicRaw","SourceGroup":"FTrace"}}
json_insert('{}',
'$.typeId', typeId,
'$.sourceId', sourceId,
'$.data', json(data)
)
--------------------------------------------------------------------------------------------------------------------
{"typeId":72057602627862547,"sourceId":72057602627862528,"data":{"Name":"raw_syscalls:sys_enter","Format":"\"NR %ld (%lx, %lx, %lx, %lx, %lx, %lx)\", REC->id, REC->args[0], REC->args[1], REC->args[2], REC->args[3], REC->args[4], REC->args[5]","Fields":[{"Name":"common_pid","Prefix":"int","Suffix":""},{"Name":"id","Prefix":"long","S
{"typeId":72057602627862670,"sourceId":72057602627862528,"data":{"Name":"irq:irq_handler_entry","Format":"\"irq=%d name=%s\", REC->irq, __get_str(name)","Fields":[{"Name":"common_pid","Prefix":"int","Suffix":""},{"Name":"irq","Prefix":"int","Suffix":""},{"Name":"name","Prefix":"__data_loc char[]","Suffix":""},{"Name":"common_type",
json_insert('{}',
'$.rawTimestamp', rawTimestamp,
'$.timestamp', timestamp,
'$.typeId', typeId,
'$.data', json(data)
)
--------------------------------------------------------------------------------------------------------------------
{"rawTimestamp":1183694330725221,"timestamp":6236683,"typeId":72057602627862670,"data":{"common_pid":"0","irq":"66","name":"327696","common_type":"142","common_flags":"9","common_preempt_count":"0"}}
{"rawTimestamp":1183694333695687,"timestamp":9207149,"typeId":72057602627862670,"data":{"common_pid":"0","irq":"66","name":"327696","common_type":"142","common_flags":"9","common_preempt_count":"0"}}
</pre></div>
</div>
</section>
<section id="arrow-format-description">
<h4>Arrow Format Description<a class="headerlink" href="#arrow-format-description" title="Permalink to this heading"></a></h4>
<p>The Arrow type exported file uses the IPC stream format to store the data in a file. The tables can be read by opening the file as an arrow stream. For example one can use the <code class="docutils literal notranslate"><span class="pre">open_stream</span></code> function from the arrow python package. For more information on the interfaces that can be used to read an IPC stream file, please refer to the Apache Arrow documentation [<a class="reference external" href="https://arrow.apache.org/docs/python/api/ipc.html">1</a>, <a class="reference external" href="https://arrow.apache.org/docs/python/ipc.html">2</a>].</p>
<p>The name of each table is included in the schema metadata. Thus, while reading each table, the user can extract the table title from the metadata. The table name metadata field has the key <code class="docutils literal notranslate"><span class="pre">table_name</span></code>. The titles of all the available tables can be found in section <a class="reference external" href="index.html#exporter-sqlite-schema">SQLite Schema Reference</a>.</p>
</section>
<section id="json-and-text-format-description">
<h4>JSON and Text Format Description<a class="headerlink" href="#json-and-text-format-description" title="Permalink to this heading"></a></h4>
<p>JSON and TXT export formats are generated by serializing buffered messages, each on a new line. First, all collected events are processed. Then strings are serialized, followed by stdout, stderr streams if any, followed by thread names.</p>
<p>Output layout:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>{Event #1}
{Event #2}
...
{Event #N}
{Strings}
{Streams}
{Threads}
</pre></div>
</div>
<p>For easier grepping of JSON output, the <code class="docutils literal notranslate"><span class="pre">--separate-strings</span></code> switch may be used to force manual splitting of strings, streams and thread names data.</p>
<p>Example line split: <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">export</span> <span class="pre">--export-json</span> <span class="pre">--separate-strings</span> <span class="pre">sample.nsys-rep</span> <span class="pre">--</span> <span class="pre">-</span></code></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>{"type":"String","id":"3720","value":"Process 14944 was launched by the profiler"}
{"type":"String","id":"3721","value":"Profiling has started."}
{"type":"String","id":"3722","value":"Profiler attached to the process."}
{"type":"String","id":"3723","value":"Profiling has stopped."}
{"type":"ThreadName","globalTid":"72057844756653436","nameId":"14","priority":"10"}
{"type":"ThreadName","globalTid":"72057844756657940","nameId":"15","priority":"10"}
{"type":"ThreadName","globalTid":"72057844756654400","nameId":"24","priority":"10"}
</pre></div>
</div>
<p>Compare with: <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">export</span> <span class="pre">--export-json</span> <span class="pre">sample.nsys-rep</span> <span class="pre">--</span> <span class="pre">-</span></code></p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>{"data":["[Unknown]","[Unknown kernel module]","[Max depth]","[Broken backtraces]",
"[Called from Java]","QnxKernelTrace","mm_","task_submit","class_id","syncpt_id",
"syncpt_thresh","pid","tid","FTrace","[NSys]","[NSys Comms]", "..." ,"Process
14944 was launched by the profiler","Profiling has started.","Profiler attached
to the process.","Profiling has stopped."]}
{"data":[{"nameIdx":"14","priority":"10","globalTid":"72057844756653436"},
{"nameIdx":"15","priority":"10","globalTid":"72057844756657940"},{"nameIdx":"24",
"priority":"10","globalTid":"72057844756654400"}]}
</pre></div>
</div>
<p>Note, that only last few lines are shown here for clarity and that carriage returns and indents were added to avoid wrapping documentation.</p>
</section>
</section>
<section id="statistical-analysis">
<h3>Statistical Analysis<a class="headerlink" href="#statistical-analysis" title="Permalink to this heading"></a></h3>
<section id="statistical-reports-shipped-with-product-name">
<h4>Statistical Reports Shipped With Nsight Systems<a class="headerlink" href="#statistical-reports-shipped-with-product-name" title="Permalink to this heading"></a></h4>
<p>The Nsight Systems development team created and maintains a set of report scripts for some of the commonly requested statistical reports. These scripts will be updated to adapt to any changes in SQLite schema or internal data structures.</p>
<p>These scripts are located in the Nsight Systems package in the Target-<architecture>/reports directory. The following standard reports are available:</p>
<p><strong>Note:</strong> The ability to display mangled names is a recent addition to the report file format, and requires that the profile data be captured with a recent version of Nsys. Re-exporting an existing report file is not sufficient. If the raw, mangled kernel name data is not available, the default demangled names will be used.</p>
<p><strong>Note:</strong> All time values given in nanoseconds by default. If you wish to output the results using a different time unit, use the <code class="docutils literal notranslate"><span class="pre">--timeunit</span></code> option when running the recipe.</p>
</section>
<section id="report-formatters-shipped-with-product-name">
<h4>Report Formatters Shipped With Nsight Systems<a class="headerlink" href="#report-formatters-shipped-with-product-name" title="Permalink to this heading"></a></h4>
<p>The following formats are available in Nsight Systems</p>
</section>
</section>
<section id="expert-systems-analysis">
<h3>Expert Systems Analysis<a class="headerlink" href="#expert-systems-analysis" title="Permalink to this heading"></a></h3>
<p>The Nsight Systems expert system is a feature aimed at automatic detection of performance optimization opportunities in an application’s profile. It uses a set of predefined rules to determine if the application has known bad patterns.</p>
<section id="using-expert-system-from-the-cli">
<h4>Using Expert System from the CLI<a class="headerlink" href="#using-expert-system-from-the-cli" title="Permalink to this heading"></a></h4>
<p>usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="w"> </span><span class="p">[</span><span class="n">global</span><span class="o">-</span><span class="n">options</span><span class="p">]</span><span class="w"> </span><span class="n">analyze</span><span class="w"> </span><span class="p">[</span><span class="n">options</span><span class="p">]</span><span class="w"></span>
<span class="w"> </span><span class="p">[</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="o">-</span><span class="n">or</span><span class="o">-</span><span class="n">sqlite</span><span class="o">-</span><span class="n">file</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>If a .nsys-rep file is given as the input file and there is no .sqlite file with the same name in the same directory, it will be generated.</p>
<p><strong>Note:</strong> The Expert System view in the GUI will give you the equivalent command line.</p>
</section>
<section id="using-expert-system-from-the-gui">
<h4>Using Expert System from the GUI<a class="headerlink" href="#using-expert-system-from-the-gui" title="Permalink to this heading"></a></h4>
<p>The Expert System View can be found in the same drop-down as the Events View. If there is no .sqlite file with the same name as the .nsys-rep file in the same directory, it will be generated.</p>
<p>The Expert System View has the following components:</p>
<ol class="arabic simple">
<li><p>Drop-down to select the rule to be run</p></li>
<li><p>Rule description and advice summary</p></li>
<li><p>CLI command that will give the same result</p></li>
<li><p>Table containing results of running the rule</p></li>
<li><p>Settings button that allows users to specify the rule’s arguments</p></li>
</ol>
<img alt="Expert systems information as shown in the GUI" class="image" src="../_images/expert-systems-gui.png" />
<p>A context menu is available to correlate the table entry with the timeline. The options are the same as the Events View:</p>
<ul class="simple">
<li><p>Zoom to Selected on Timeline (ctrl+double-click)</p></li>
</ul>
<p>The highlighting is not supported for rules that do not return an event but rather an arbitrary time range (e.g. GPU utilization rules).</p>
<p>The CLI and GUI share the same rule scripts and messages. There might be some formatting differences between the output table in GUI and CLI.</p>
</section>
<section id="expert-system-rules">
<h4>Expert System Rules<a class="headerlink" href="#expert-system-rules" title="Permalink to this heading"></a></h4>
<p>Rules are scripts that run on the SQLite DB output from Nsight Systems to find common improvable usage patterns.</p>
<p>Each rule has an advice summary with explanation of the problem found and suggestions to address it. Only the top 50 results are displayed by default.</p>
<p>There are currently six rules in the expert system. They are described below. Additional rules will be made available in a future version of Nsight Systems.</p>
</section>
</section>
<section id="multi-report-analysis">
<h3>Multi-Report Analysis<a class="headerlink" href="#multi-report-analysis" title="Permalink to this heading"></a></h3>
<p><strong>PREVIEW FEATURE</strong></p>
<p>Nsight Systems Multi-Report Analysis is new functionality that is being added to the Nsight Systems tool to better support complex statistical analysis across multiple result files. Possible use cases for this functionality include:</p>
<ul class="simple">
<li><p>Multi-Node Analysis - When you run Nsight Systems across a cluster, it typically generates one result file per rank on the cluster. While you can load multiple result files into the GUI for visualization, this analysis system allows you to run statistical analysis across all of the result files.</p></li>
<li><p>Multi-Pass Analysis - Some features in Nsight Systems cannot be run together due to overhead or hardware considerations. For example, there are frequently more CPU performance counters available than your CPU has registers. Using this analysis, you could run multiple runs with different sets of counters and then analyze the results together.</p></li>
<li><p>Multi-Run Analysis - Sometimes you want to compare two runs that were not taken at the same time together. Perhaps you ran the tool on two different hardware configurations and want to see what changed. Perhaps you are doing regression testing or performance improvement analysis and want to check your status. Comparing those result files statistically can show patterns.</p></li>
</ul>
<p><strong>Analysis Steps</strong></p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Prior to using multi-report analysis, please make sure that you have installed all required dependencies. See <strong>Installing Multi-Report Analysis System</strong> in the <strong>Installation Guide</strong> for more information.</p>
</div>
<ol class="arabic simple">
<li><p>Generate the reports - Generate the reports as you always have, in fact, you can use reports that you have generated previously.</p></li>
<li><p>Set up - Choose the recipe (See Available Recipes, below), give it any required parameters, and run.</p></li>
<li><p>Launch Analysis - Nsight Systems will run the analysis, using your local system or Dask, as you have selected.</p></li>
<li><p>Output - the output is an .nsys-analysis file, which can then be opened within the Nsight Systems GUI.</p></li>
<li><p>View the data - depending on your recipe, you can have any number of visualizations, from simple tabular information to Jupyter notebooks which can be opened inside the GUI.</p></li>
</ol>
<section id="available-recipes">
<h4>Available Recipes<a class="headerlink" href="#available-recipes" title="Permalink to this heading"></a></h4>
<p>All recipes are run using the new “recipe” CLI command switch.</p>
<p>usage:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsys recipe [args] <recipe-name> [recipe args]
</pre></div>
</div>
<p>Nsight Systems provides several initial analysis recipes, mostly based around making our existing statistics and expert systems rules run multi-report.</p>
<p>These recipes can be found at <code class="docutils literal notranslate"><span class="pre"><target-linux-x64>/python/packages/nsys_recipe</span></code>. They are written in Python, and you can edit them if you would like. However, be advised that as this is a preview release, it is likely that the APIs will change between now and the final release. Additional recipes will be added before the product release and on an ongoing basis.</p>
<p><strong>Statistics and Expert Systems Recipes</strong></p>
<p>The following stats and expert systems options from Nsight Systems are available as recipes. For more information about them please use <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">recipe</span> <span class="pre">[recipe</span> <span class="pre">name]</span> <span class="pre">--help</span></code> or see Nsight Systems Report Scripts in this documentation</p>
<ul class="simple">
<li><p>cuda_api_sum - CUDA API Summary</p></li>
<li><p>cuda_api_sync - CUDA Synchronization APIs</p></li>
<li><p>cuda_gpu_kern_sum - CUDA GPU Kernel Summary</p></li>
<li><p>cuda_gpu_mem_size_sum - CUDA GPU MemOps Summary (by Size)</p></li>
<li><p>cuda_gpu_mem_time_sum - CUDA GPU MemOps Summary (by Time)</p></li>
<li><p>cuda_memcpy_async - CUDA Async Memcpy with Pageable Memory</p></li>
<li><p>cuda_memcpy_sync - CUDA Synchronous Memcpy</p></li>
<li><p>cuda_memset_sync - CUDA Synchronous Memset</p></li>
<li><p>dx12_mem_ops - DX12 Memory Operations</p></li>
<li><p>gpu_gaps - GPU Gaps</p></li>
<li><p>gpu_time_util - GPU Time Utilization</p></li>
<li><p>nvtx_gpu_proj_trace - NVTX GPU Trace</p></li>
<li><p>nvtx_sum - NVTX Range Summary</p></li>
<li><p>osrt_sum - OS Runtime Summary</p></li>
</ul>
<p>Please note that all recipes are in the form of python scripts. You may alter the given recipes or write your own to meet your needs. Refer to <strong>Tutorial: Create a User-Defined Recipe</strong> for an example of how to do this</p>
<p><strong>Heatmap Recipes</strong></p>
<ul class="simple">
<li><p>cuda_gpu_time_util_map - CUDA GPU Kernel Time Utilization Heatmap</p></li>
<li><p>gpu_metric_util_map - GPU Metric Utilization Heatmap</p></li>
</ul>
<p>Both recipes generate a Jupyter notebook with code cells ready to plot the heatmap chart:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>cuda_gpu_time_util_map -- CUDA GPU Kernel Time Utilization Heatmap
$ nsys recipe cuda_gpu_time_util_map --help
usage: cuda_gpu_time_util_map.py [-h] [--output OUTPUT] [--force-overwrite]
[--start time] [--end time]
[--nvtx range[@domain]] [--rows limit]
[--bins BINS] --dir DIR
[--mode {none,concurrent,dask-futures}]
This recipe calculates the percentage of GPU utilization based on the presence
of CUDA kernels. Note that the utilization refers to the "time" utilization
and not the "resource" utilization. If multiple kernels run concurrently,
their utilization will be added up and may exceed 100%.
options:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--start time Start time used for filtering in nanoseconds
--end time End time used for filtering in nanoseconds
--nvtx range[@domain]
NVTX range and domain used for filtering
--rows limit Maximum number of rows per input file
--bins BINS Number of bins
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>gpu_metric_util_map -- GPU Metric Utilization Heatmap
$ nsys recipe gpu_metric_util_map --help
usage: gpu_metric_util_map.py [-h] [--output OUTPUT] [--force-overwrite]
[--start time] [--end time]
[--nvtx range[@domain]] [--rows limit]
[--bins BINS] --dir DIR
[--mode {none,concurrent,dask-futures}]
This recipe calculates the percentage of SM Active, SM Issue, and Tensor
Active metrics.
options:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--start time Start time used for filtering in nanoseconds
--end time End time used for filtering in nanoseconds
--nvtx range[@domain]
NVTX range and domain used for filtering
--rows limit Maximum number of rows per input file
--bins BINS Number of bins
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<p><strong>Pacing Recipes</strong></p>
<ul class="simple">
<li><p>cuda_gpu_kern_pace - CUDA GPU Kernel Pacing</p></li>
<li><p>nvtx_pace - NVTX Pacing</p></li>
</ul>
<p>Both recipes generate a Jupyter notebook with code cells ready to plot various graphs showing the progress of the target operation/range for each rank:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>cuda_gpu_kern_pace -- CUDA GPU Kernel Pacing
$ nsys recipe cuda_gpu_kern_pace --help
usage: cuda_gpu_kern_pace.py [-h] [--output OUTPUT] [--force-overwrite] --name
NAME --dir DIR
[--mode {none,concurrent,dask-futures}]
This recipe investigates the progress and consistency of an iteration based
application.
optional arguments:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--name NAME Name of the kernel used as delineator between
iterations
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nvtx_pace -- NVTX Pacing
$ nsys recipe nvtx_pace --help
usage: nvtx_pace.py [-h] [--output OUTPUT] [--force-overwrite] [--gpu] --name
NAME --dir DIR [--mode {none,concurrent,dask-futures}]
This recipe investigates the progress and consistency of an iteration based
application.
optional arguments:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--gpu GPU projection
--name NAME Name of the NVTX range used as delineator between
iterations
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<p><strong>Additional Statistics Recipes</strong></p>
<ul class="simple">
<li><p>mpi_sum - MPI Summary</p></li>
<li><p>nccl_sum - NCCL Summary</p></li>
<li><p>nvtx_gpu_proj_sum - NVTX GPU Projection Summary</p></li>
</ul>
<p>All recipes generate a Jupyter notebook with code cells ready to plot various statistical graphs:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>mpi_sum -- MPI Summary
$ nsys recipe mpi_sum --help
usage: mpi_sum.py [-h] [--output OUTPUT] [--force-overwrite] --dir DIR
[--mode {none,concurrent,dask-futures}]
This recipe provides a summary of MPI functions and their execution times.
optional arguments:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nccl_sum -- NCCL Summary
$ nsys recipe nccl_sum --help
usage: nccl_sum.py [-h] [--output OUTPUT] [--force-overwrite] [--gpu] --dir
DIR [--mode {none,concurrent,dask-futures}]
This recipe provides a summary of NCCL functions and their execution times.
optional arguments:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--gpu GPU projection
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nvtx_gpu_proj_sum -- NVTX GPU Projection Summary
$ nsys recipe nvtx_gpu_proj_sum --help
usage: nvtx_gpu_proj_sum.py [-h] [--output OUTPUT] [--force-overwrite] [--gpu]
--dir DIR [--mode {none,concurrent,dask-futures}]
This recipe provides a summary of NVTX time ranges projected from the CPU onto
the GPU, and their execution times.
optional arguments:
-h, --help show this help message and exit
Context:
--mode {none,concurrent,dask-futures}
Mode to run tasks
Recipe:
--output OUTPUT Output directory name
--force-overwrite Overwrite existing directory
--gpu GPU projection
--dir DIR Directory of nsys-rep files
</pre></div>
</div>
</section>
<section id="opening-in-jupyter-notebook">
<h4>Opening in Jupyter Notebook<a class="headerlink" href="#opening-in-jupyter-notebook" title="Permalink to this heading"></a></h4>
<p>Running the recipe command creates a new analysis file (.nsys-analysis). Open the Nsight Systems GUI and select <code class="docutils literal notranslate"><span class="pre">File->Open</span></code>, and pick your file.</p>
<img alt="Analysis in Nsys GUI" class="image" src="../_images/recipe-jupyter1.png" />
<p>Open the folder icon and click on the notebook icon to open the Jupyter notebook.</p>
<img alt="Select .nsys-analysis file" class="image" src="../_images/recipe-jupyter2.png" />
<p>Run the Jupyter notebook:</p>
<img alt="Run notebook" class="image" src="../_images/recipe-jupyter3.png" />
<p>And the output appears on-screen. In this case a heat map of activity running a Jacobi solver.</p>
<img alt="Heatmap" class="image" src="../_images/recipe-jupyter4.png" />
</section>
<section id="configuring-dask">
<h4>Configuring Dask<a class="headerlink" href="#configuring-dask" title="Permalink to this heading"></a></h4>
<p>The multi-report analysis system does not offer options to configure the Dask environment. However, you could achieve this by modifying the recipe script directly or using one of the following from Dask’s configuration system:</p>
<ul>
<li><p>YAML files: Dask by default searches for all YAML files in <code class="docutils literal notranslate"><span class="pre">~/.config/dask/</span></code> or <code class="docutils literal notranslate"><span class="pre">/etc/dask/</span></code>. This search path can be changed using the environment variable <code class="docutils literal notranslate"><span class="pre">DASK_ROOT_CONFIG</span></code> or <code class="docutils literal notranslate"><span class="pre">DASK_CONFIG</span></code>. See Dask documentation for the complete list of locations and the lookup order. Example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>$ cat example.yaml
'Distributed':
'scheduler':
'allowed-failures': 5
</pre></div>
</div>
</li>
<li><p>Environment variables: Dask searches for all environment variables that start with <code class="docutils literal notranslate"><span class="pre">DASK_</span></code>, then transforms keys by converting to lower-case and changing double-underscores to nested structures. See Dask documentation for the complete list of variables. Example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>DASK_DISTRIBUTED__SCHEDULER__ALLOWED_FAILURES=5
</pre></div>
</div>
</li>
</ul>
<p><strong>Dask Client</strong></p>
<p>With no configuration set, the dask-futures mode option initializes the Dask Client with the default arguments, which results in creating a LocalCluster in the background. The following are the YAML/environment variables that could be set to change the default behavior:</p>
<ul class="simple">
<li><p>distributed.comm.timeouts.connect / DASK_DISTRIBUTED__COMM__TIMEOUTS__CONNECT</p></li>
<li><p>client-name / DASK_CLIENT_NAME</p></li>
<li><p>scheduler-address / DASK_SCHEDULER_ADDRESS</p></li>
<li><p>distributed.client.heartbeat / DASK_DISTRIBUTED__CLIENT__HEARTBEAT</p></li>
<li><p>distributed.client.scheduler-info-interval / DASK_DISTRIBUTED__CLIENT__SCHEDULER_INFO_INTERVAL</p></li>
<li><p>distributed.client.preload / DASK_DISTRIBUTED__CLIENT__PRELOAD</p></li>
<li><p>distributed.client.preload-argv / DASK_DISTRIBUTED__CLIENT__PRELOAD_ARGV</p></li>
</ul>
<p><strong>Recipe’s environment variables</strong></p>
<p>Recipe has its own list of environment variables to make the configuration more complete and flexible. These environment variables are either missing from Dask’s configuration system or specific to the recipe system:</p>
<ul class="simple">
<li><p>NSYS_DASK_SCHEDULER_FILE: Path to a file with scheduler information. It will be used to initialize the Dask Client.</p></li>
<li><p>NSYS_DIR: Path to the directory of Nsight Systems containing the target and host directories. The nsys executable and the recipe dependencies will be searched in this directory instead of the one deduced from the currently running recipe file path.</p></li>
</ul>
</section>
<section id="tutorial-create-a-user-defined-recipe">
<h4>Tutorial: Create a User-Defined Recipe<a class="headerlink" href="#tutorial-create-a-user-defined-recipe" title="Permalink to this heading"></a></h4>
<p>The Nsight Systems recipe system is designed to be extensible and we hope that
many users will use it to create their own recipes. This short tutorial will
highlight the steps needed to create a recipe that is a customized version of
one of the recipes that is included in the Nsight Systems recipe package.</p>
<p><strong>Step 1: Create the recipe directory and script</strong></p>
<p>Make a new directory in the
<code class="docutils literal notranslate"><span class="pre"><install-dir>/target-linux-x64/python/packages/nsys_recipe</span></code> folder based on
the name of your new recipe. For this example, we will call our new recipe
new_metric_util_map. We will copy the existing gpu_metric_util_map.py script
and create a new script called
new_metric_util_map.py in the new_metric_util_map directory. We will also
copy the heatmap.ipynb file into the new_metric_util_map directory. Type these
steps in a Linux terminal window:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>> cd <install-dir>/target-linux-x64/python/packages/nsys_recipe
> mkdir new_metric_util_map
> cp gpu_metric_util_map/metadata.json new_metric_util_map/metadata.json
> cp gpu_metric_util_map/heatmap.ipynb new_metric_util_map/heatmap.ipynb
</pre></div>
</div>
<p>Replace the module name in metadata.json with new_metric_util_map
and update the display name and description to your preference. Also, rename
the class name GpuMetricUtilMap in gpu_metric_util_map.py to
NewMetricUtilMap. We will discuss the detailed functionality of the new
recipe code in the subsequent steps.</p>
<p><strong>Step 2: Modify the mapper function</strong></p>
<p>Many recipes are structured as a map-reduce algorithm. The mapper function is
called for every .nsys-rep file in the report directory. The mapper function
performs a series of calculations on the events in each Nsight Systems report
and produces an intermediate data set. The intermediate results are then
combined by the reduce function to produce the final results. The mapper
function can be called in parallel, either on multiple cores of a single node
(using the concurrent python module), or multiple ranks of a multi-node recipe
analysis (using the Dask distributed module).</p>
<p>When we create a new recipe, we need to create a class that derives from the
Recipe base class. For our example, that class will be called NewMetricUtilMap
(which we had renamed in step 1).</p>
<p>The mapper function is called mapper_func(). It will first convert the .nsys-rep
file into an SQLite database, if the SQLite file does not already exist. It then
reads all the necessary tables from the SQLite file into Pandas Dataframes
needed by the recipe. GPU Metric data is stored using a database schema table
called GENERIC_EVENTS. For extra flexibility, GENERIC_EVENTS represents the
data as a JSON object, which is stored as a string. The NewMetricUtilMap class
extracts fields from the JSON object and accumulates them over the histogram
bins of the heat map.</p>
<p>The original script retrieved three GPU metrics: SM Active, SM Issue, and Tensor
Active. In our new version of the script, we will extract a fourth metric,
Unallocated Warps in Active SMs.</p>
<ol class="arabic">
<li><p>Find this line (approximately line 65):</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>metric_cols = ['SM Active', 'SM Issue', 'Tensor Active']
</pre></div>
</div>
</li>
<li><p>Add the Unallocated Warps in Active SMs metric:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>metric_cols = ['SM Active', 'SM Issue', 'Tensor Active', 'Unallocated Warps in Active SMs']
</pre></div>
</div>
</li>
</ol>
<p><strong>Step 3: Modify the reduce function</strong></p>
<p>Our new mapper function will extract four GPU metrics and return them as a
Pandas DataFrame. The reduce function receives a list of DataFrames, one for
each .nsys-rep file in the analysis, and combines them into a single DataFrame
using the Pandas concat function. Since the reducer function is generic in our
case, no modifications are needed. However, if you would like to add any
additional post-processing, you can do so in this function.</p>
<p><strong>Step 4: Add a plot to the Jupyter notebook</strong></p>
<p>Our new recipe class will create a Parquet output file with all the data
produced by the reducer function, using the to_parquet() function. It will also
create a Jupyter notebook file using the create_notebook() function.</p>
<p>In this step, we will change the create_notebook() function to produce a plot
for our fourth metric. To do this, we need to change these two lines (located
in the second cell of new_metric_util_map/heatmap.ipynb):</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>metrics = ('SM Active', 'SM Issue', 'Tensor Active')
</pre></div>
</div>
<p>To this:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>metrics = ('SM Active', 'SM Issue', 'Tensor Active', 'Unallocated Warps in Active SMs')
</pre></div>
</div>
<p>That completes all the modifications for our NewMetricUtilMap class.</p>
<p><strong>Step 5: Run the new recipe</strong></p>
<p>If the new recipe is located in the default recipe directory nsys_recipe/recipes,
we can directly run it using the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">recipe</span></code> command like this:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>> nsys recipe new_metric_util_map --input <directory of reports>
</pre></div>
</div>
<p>It is also possible to have a recipe located outside of this directory. In this
case, you need to set the environment variable NSYS_RECIPE_PATH to the directory
containing the recipe when running the <code class="docutils literal notranslate"><span class="pre">nsys</span> <span class="pre">recipe</span></code> command.</p>
<p>When successful, the recipe should produce a new recipe result directory called
new_metric_util_map-1.</p>
<p>If we open the Jupyter notebook in that recipe and execute the code, we should
see our new heatmap along with the three plots produced by the original version
of the recipe. Here is an example:</p>
<img alt="Output from tutorial recipe" class="image" src="../_images/recipe-tutorial.png" />
</section>
</section>
</section>
<section id="import-nvtxt">
<h2>Import NVTXT<a class="headerlink" href="#import-nvtxt" title="Permalink to this heading"></a></h2>
<p><strong>ImportNvtxt</strong> is an utility which allows conversion of a <a class="reference external" href="https://docs.nvidia.com/gameworks/index.html#gameworkslibrary/nvtx/analysis_nvtxt_file_extension.htm">NVTXT</a> file to a Nsight Systems report file (*.nsys-rep) or to merge it with an existing report file.</p>
<p><strong>Note</strong>: NvtxtImport supports custom <strong>TimeBase</strong> values. Only these values are supported:</p>
<ul class="simple">
<li><p><strong>Manual</strong> — timestamps are set using absolute values.</p></li>
<li><p><strong>Relative</strong> — timestamps are set using relative values with regards to report file which is being merged with nvtxt file.</p></li>
<li><p><strong>ClockMonotonicRaw</strong> — timestamps values in nvtxt file are considered to be gathered on the same target as the report file which is to be merged with nvtxt using <code class="docutils literal notranslate"><span class="pre">clock_gettime(CLOCK_MONOTONIC_RAW,</span> <span class="pre">...)</span></code> call.</p></li>
<li><p><strong>CNTVCT</strong> — timestamps values in nvtxt file are considered to be gathered on the same target as the report file which is to be merged with nvtxt using CNTVCT values.</p></li>
</ul>
<p>You can get usage info via help message:</p>
<p>Print help message:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">-</span><span class="n">h</span><span class="w"> </span><span class="p">[</span><span class="w"> </span><span class="o">--</span><span class="n">help</span><span class="w"> </span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>Show information about report file:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">info</span><span class="w"> </span><span class="o">-</span><span class="n">i</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">input</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"></span>
</pre></div>
</div>
<p>Create report file from existing nvtxt file:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">create</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">output</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">-</span><span class="n">m</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">mode</span><span class="p">]</span><span class="w"> </span><span class="n">mode_name</span><span class="w"> </span><span class="n">mode_args</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">target</span><span class="w"> </span><span class="o"><</span><span class="n">Hw</span><span class="o">:</span><span class="n">Vm</span><span class="o">></span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">update_report_time</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>Merge nvtxt file to existing report file:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">merge</span><span class="w"> </span><span class="o">-</span><span class="n">i</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">input</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">output</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">-</span><span class="n">m</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">mode</span><span class="p">]</span><span class="w"> </span><span class="n">mode_name</span><span class="w"> </span><span class="n">mode_args</span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">target</span><span class="w"> </span><span class="o"><</span><span class="n">Hw</span><span class="o">:</span><span class="n">Vm</span><span class="o">></span><span class="p">]</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">update_report_time</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>Modes description:</p>
<ul>
<li><p>lerp - Insert with linear interpolation</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lerp</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">ns_b</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">nvtxt_b</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>lin - insert with linear equation</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lin</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">freq</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
</li>
</ul>
<p>Modes’ parameters:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">ns_a</span></code> - a nanoseconds value</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">ns_b</span></code> - a nanoseconds value (greater than <code class="docutils literal notranslate"><span class="pre">ns_a</span></code>)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> - an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_a</span></code> nanoseconds</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_b</span></code> - an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_b</span></code> nanoseconds</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">freq</span></code> - the nvtxt file’s timer frequency</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--target</span> <span class="pre"><Hw:Vm></span></code> - specify target id, e.g. <code class="docutils literal notranslate"><span class="pre">--target</span> <span class="pre">0:1</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">--update_report_time</span></code> - prolong report’s profiling session time while merging if needed. Without this option all events outside the profiling session time window will be skipped during merging.</p></li>
</ul>
<section id="commands">
<h3>Commands<a class="headerlink" href="#commands" title="Permalink to this heading"></a></h3>
<p><strong>Info</strong></p>
<p>To find out report’s start and end time use <strong>info</strong> command.</p>
<p>Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">info</span><span class="w"> </span><span class="o">-</span><span class="n">i</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">input</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"></span>
</pre></div>
</div>
<p>Example:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="n">info</span><span class="w"> </span><span class="n">Report</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"></span>
<span class="n">Analysis</span><span class="w"> </span><span class="n">start</span><span class="w"> </span><span class="p">(</span><span class="n">ns</span><span class="p">)</span><span class="w"> </span><span class="mi">83501026500000</span><span class="w"></span>
<span class="n">Analysis</span><span class="w"> </span><span class="n">end</span><span class="w"> </span><span class="p">(</span><span class="n">ns</span><span class="p">)</span><span class="w"> </span><span class="mi">83506375000000</span><span class="w"></span>
</pre></div>
</div>
<p><strong>Create</strong></p>
<p>You can create a report file using existing NVTXT with <strong>create</strong> command.</p>
<p>Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">create</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">output</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">-</span><span class="n">m</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">mode</span><span class="p">]</span><span class="w"> </span><span class="n">mode_name</span><span class="w"> </span><span class="n">mode_args</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>Available modes are:</p>
<ul class="simple">
<li><p><strong>lerp</strong> — insert with linear interpolation.</p></li>
<li><p><strong>lin</strong> — insert with linear equation.</p></li>
</ul>
<p>Usage for <strong>lerp</strong> mode is:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lerp</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">ns_b</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">nvtxt_b</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>with:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">ns_a</span></code> — a nanoseconds value.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">ns_b</span></code> — a nanoseconds value (greater than <code class="docutils literal notranslate"><span class="pre">ns_a</span></code>).</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_a</span></code> nanoseconds.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_b</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_b</span></code> nanoseconds.</p></li>
</ul>
<p>If <code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> and <code class="docutils literal notranslate"><span class="pre">nvtxt_b</span></code> are not specified, they are respectively set to nvtxt file’s minimum and maximum time value.</p>
<p>Usage for <strong>lin</strong> mode is:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lin</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">freq</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>with:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">ns_a</span></code> — a nanoseconds value.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">freq</span></code> — the nvtxt file’s timer frequency.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_a</span></code> nanoseconds.</p></li>
</ul>
<p>If <code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> is not specified, it is set to nvtxt file’s minimum time value.</p>
<p><strong>Examples:</strong></p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">create</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="n">Sample</span><span class="p">.</span><span class="n">nvtxt</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="n">Report</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"></span>
</pre></div>
</div>
<p>The output will be a new generated report file which can be opened and viewed by Nsight Systems.</p>
<p><strong>Merge</strong></p>
<p>To merge NVTXT file with an existing report file use <strong>merge</strong> command.</p>
<p>Usage:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">merge</span><span class="w"> </span><span class="o">-</span><span class="n">i</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">input</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">output</span><span class="p">]</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">-</span><span class="n">m</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">mode</span><span class="p">]</span><span class="w"> </span><span class="n">mode_name</span><span class="w"> </span><span class="n">mode_args</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>Available modes are:</p>
<ul class="simple">
<li><p><strong>lerp</strong> — insert with linear interpolation.</p></li>
<li><p><strong>lin</strong> — insert with linear equation.</p></li>
</ul>
<p>Usage for <strong>lerp</strong> mode is:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lerp</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">ns_b</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">nvtxt_b</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>with:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">ns_a</span></code> — a nanoseconds value.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">ns_b</span></code> — a nanoseconds value (greater than <code class="docutils literal notranslate"><span class="pre">ns_a</span></code>).</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_a</span></code> nanoseconds.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_b</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_b</span></code> nanoseconds.</p></li>
</ul>
<p>If <code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> and <code class="docutils literal notranslate"><span class="pre">nvtxt_b</span></code> are not specified, they are respectively set to nvtxt file’s minimum and maximum time value.</p>
<p>Usage for <strong>lin</strong> mode is:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">--</span><span class="n">mode</span><span class="w"> </span><span class="n">lin</span><span class="w"> </span><span class="o">--</span><span class="n">ns_a</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="o">--</span><span class="n">freq</span><span class="w"> </span><span class="n">arg</span><span class="w"> </span><span class="p">[</span><span class="o">--</span><span class="n">nvtxt_a</span><span class="w"> </span><span class="n">arg</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
<p>with:</p>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">ns_a</span></code> — a nanoseconds value.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">freq</span></code> — the nvtxt file’s timer frequency.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> — an nvtxt file’s time unit value corresponding to <code class="docutils literal notranslate"><span class="pre">ns_a</span></code> nanoseconds.</p></li>
</ul>
<p>If <code class="docutils literal notranslate"><span class="pre">nvtxt_a</span></code> is not specified, it is set to nvtxt file’s minimum time value.</p>
<p>Time values in <code class="docutils literal notranslate"><span class="pre"><filename.nvtxt></span></code> are assumed to be nanoseconds if no mode specified.</p>
<p>Example</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">ImportNvtxt</span><span class="w"> </span><span class="o">--</span><span class="n">cmd</span><span class="w"> </span><span class="n">merge</span><span class="w"> </span><span class="o">-</span><span class="n">i</span><span class="w"> </span><span class="n">Report</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"> </span><span class="o">-</span><span class="n">n</span><span class="w"> </span><span class="n">Sample</span><span class="p">.</span><span class="n">nvtxt</span><span class="w"> </span><span class="o">-</span><span class="n">o</span><span class="w"> </span><span class="n">NewReport</span><span class="p">.</span><span class="n">nsys</span><span class="o">-</span><span class="n">rep</span><span class="w"></span>
</pre></div>
</div>
</section>
</section>
<section id="visual-studio-integration">
<h2>Visual Studio Integration<a class="headerlink" href="#visual-studio-integration" title="Permalink to this heading"></a></h2>
<p>NVIDIA Nsight Integration is a Visual Studio extension that allows you to access the power of Nsight Systems from within Visual Studio.</p>
<p>When Nsight Systems is installed along with NVIDIA Nsight Integration, Nsight Systems activities will appear under the NVIDIA Nsight menu in the Visual Studio menu bar. These activities launch Nsight Systems with the current project settings and executable.</p>
<img alt="Install extension to Microsoft Visual Studio" class="image" src="../_images/visual_studio_extension_menu_trace_command.png" />
<p>Selecting the “Trace” command will launch Nsight Systems, create a new Nsight Systems project and apply settings from the current Visual Studio project:</p>
<ul class="simple">
<li><p>Target application path</p></li>
<li><p>Command line parameters</p></li>
<li><p>Working folder</p></li>
</ul>
<p>If the “Trace” command has already been used with this Visual Studio project then Nsight Systems will load the respective Nsight Systems project and any previously captured trace sessions will be available for review using the Nsight Systems project explorer tree.</p>
<p>For more information about using Nsight Systems from within Visual Studio, please visit</p>
<ul class="simple">
<li><p><a class="reference external" href="https:/developer.nvidia.com/nsight-tools-visual-studio-integration">NVIDIA Nsight Integration Overview</a></p></li>
<li><p><a class="reference external" href="https:/docs.nvidia.com/nsight-vs-integration/index.html">NVIDIA Nsight Integration User Guide</a></p></li>
</ul>
</section>
<section id="troubleshooting">
<h2>Troubleshooting<a class="headerlink" href="#troubleshooting" title="Permalink to this heading"></a></h2>
<section id="general-troubleshooting">
<h3>General Troubleshooting<a class="headerlink" href="#general-troubleshooting" title="Permalink to this heading"></a></h3>
<p><strong>Profiling</strong></p>
<p>If the profiler behaves unexpectedly during the profiling session, or the profiling session fails to start, try the following steps:</p>
<ul class="simple">
<li><p>Close the host application.</p></li>
<li><p>Restart the target device.</p></li>
<li><p>Start the host application and connect to the target device.</p></li>
</ul>
<p>Nsight Systems uses a settings file (<code class="docutils literal notranslate"><span class="pre">NVIDIA</span> <span class="pre">Nsight</span> <span class="pre">Systems.ini</span></code>) on the host to store information about loaded projects, report files, window layout configuration, etc. Location of the settings file is described in the <strong>Help → About</strong> dialog. Deleting the settings file will restore Nsight Systems to a fresh state, but all projects and reports will disappear from the Project Explorer.</p>
<p><strong>Environment Variables</strong></p>
<p>By default, Nsight Systems writes temporary files to <code class="docutils literal notranslate"><span class="pre">/tmp</span></code> directory. If you are using a system that does not allow writing to <code class="docutils literal notranslate"><span class="pre">/tmp</span></code> or where the <code class="docutils literal notranslate"><span class="pre">/tmp</span></code> directory has limited storage you can use the TMPDIR environment variable to set a different location. An example:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>TMPDIR=/testdata ./bin/nsys profile -t cuda matrixMul
</pre></div>
</div>
<p>Environment variable control support for Windows target trace is not available, but there is a quick workaround:</p>
<ul class="simple">
<li><p>Create a batch file that sets the env vars and launches your application.</p></li>
<li><p>Set Nsight Systems to launch the batch file as its target, i.e. set the project settings target path to the path of batch file.</p></li>
<li><p>Start the trace. Nsight Systems will launch the batch file in a new cmd instance and trace any child process it launches. In fact, it will trace the whole process tree whose root is the cmd running your batch file.</p></li>
</ul>
<p><strong>WebGL Testing</strong></p>
<p>Nsight Systems cannot profile using the default Chrome launch command. To profile WebGL please follow the following command structure:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>“C:\Program Files (x86)\Google\Chrome\Application\chrome.exe”
--inprocess-gpu --no-sandbox --disable-gpu-watchdog --use-angle=gl
https://webglsamples.org/aquarium/aquarium.html
</pre></div>
</div>
<p><strong>Common Issues with QNX Targets</strong></p>
<ul class="simple">
<li><p>Make sure that <code class="docutils literal notranslate"><span class="pre">tracelogger</span></code> utility is available and can be run on the target.</p></li>
<li><p>Make sure that <code class="docutils literal notranslate"><span class="pre">/tmp</span></code> directory is accessible and supports sub-directories.</p></li>
<li><p>When switching between Nsight Systems versions, processes related to the previous version, including profiled applications forked by the daemon, must be killed before the new version is used. If you experience issues after switching between Nsight Systems versions, try rebooting the target.</p></li>
</ul>
</section>
<section id="cli-troubleshooting">
<h3>CLI Troubleshooting<a class="headerlink" href="#cli-troubleshooting" title="Permalink to this heading"></a></h3>
<p>If you have collected a report file using the CLI and the report will not open in the GUI, check to see that your GUI version is the same or greater than the CLI version you used. If it is not, download a new version of the Nsight Systems GUI and you will be able to load and visualize your report.</p>
<p>This situation occurs most frequently when you update Nsight Systems using a CLI only package, such as the package available from the NVIDIA HPC SDK.</p>
</section>
<section id="launch-processes-in-stopped-state">
<h3>Launch Processes in Stopped State<a class="headerlink" href="#launch-processes-in-stopped-state" title="Permalink to this heading"></a></h3>
<p>In many cases, it is important to profile an application from the very beginning of its execution. When launching processes, Nsight Systems takes care of it by making sure that the profiling session is fully initialized before making the <code class="docutils literal notranslate"><span class="pre">exec()</span></code> system call on Linux.</p>
<p>If the process launch capabilities of Nsight Systems are not sufficient, the application should be launched manually, and the profiler should be configured to attach to the already launched process. One approach would be to call <code class="docutils literal notranslate"><span class="pre">sleep()</span></code> somewhere early in the application code, which would provide time for the user to attach to the process in Nsight Systems Embedded Platforms Edition, but there are two other more convenient mechanisms that can be used on Linux, without the need to recompile the application. (Note that the rest of this section is only applicable to Linux-based target devices.)</p>
<p>Both mechanisms ensure that between the time the process is created (and therefore its PID is known) and the time any of the application’s code is called, the process is stopped and waits for a signal to be delivered before continuing.</p>
<section id="ld-preload">
<h4>LD_PRELOAD<a class="headerlink" href="#ld-preload" title="Permalink to this heading"></a></h4>
<p>The first mechanism uses <code class="docutils literal notranslate"><span class="pre">LD_PRELOAD</span></code> environment variable. It only works with dynamically linked binaries, since static binaries do not invoke the runtime linker, and therefore are not affected by the <code class="docutils literal notranslate"><span class="pre">LD_PRELOAD</span></code> environment variable.</p>
<ul>
<li><p>For ARMv7 binaries, preload</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="n">libLauncher32</span><span class="p">.</span><span class="n">so</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Otherwise if running from host, preload</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="n">libLauncher64</span><span class="p">.</span><span class="n">so</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Otherwise if running from CLI, preload</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">installation_directory</span><span class="p">]</span><span class="o">/</span><span class="n">libLauncher64</span><span class="p">.</span><span class="n">so</span><span class="w"></span>
</pre></div>
</div>
</li>
</ul>
<p>The most common way to do that is to specify the environment variable as part of the process launch command, for example:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">LD_PRELOAD</span><span class="o">=/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="n">libLauncher64</span><span class="p">.</span><span class="n">so</span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">my</span><span class="o">-</span><span class="n">aarch64</span><span class="o">-</span><span class="n">binary</span><span class="w"> </span><span class="o">--</span><span class="n">arguments</span><span class="w"></span>
</pre></div>
</div>
<p>When loaded, this library will send itself a <code class="docutils literal notranslate"><span class="pre">SIGSTOP</span></code> signal, which is equivalent to typing <code class="docutils literal notranslate"><span class="pre">Ctrl+Z</span></code> in the terminal. The process is now a background job, and you can use standard commands like jobs, <code class="docutils literal notranslate"><span class="pre">fg</span></code> and <code class="docutils literal notranslate"><span class="pre">bg</span></code> to control them. Use <code class="docutils literal notranslate"><span class="pre">jobs</span> <span class="pre">-l</span></code> to see the PID of the launched process.</p>
<p>When attaching to a stopped process, Nsight Systems will send <code class="docutils literal notranslate"><span class="pre">SIGCONT</span></code> signal, which is equivalent to using the <code class="docutils literal notranslate"><span class="pre">bg</span></code> command.</p>
</section>
<section id="launcher">
<h4>Launcher<a class="headerlink" href="#launcher" title="Permalink to this heading"></a></h4>
<p>The second mechanism can be used with any binary. Use <code class="docutils literal notranslate"><span class="pre">[installation_directory]/launcher</span></code> to launch your application, for example:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="n">launcher</span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">my</span><span class="o">-</span><span class="n">binary</span><span class="w"> </span><span class="o">--</span><span class="n">arguments</span><span class="w"></span>
</pre></div>
</div>
<p>The process will be launched, daemonized, and wait for <code class="docutils literal notranslate"><span class="pre">SIGUSR1</span></code> signal. After attaching to the process with Nsight Systems, the user needs to manually resume execution of the process from command line:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">pkill</span><span class="w"> </span><span class="o">-</span><span class="n">USR1</span><span class="w"> </span><span class="n">launcher</span><span class="w"></span>
</pre></div>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Note that <code class="docutils literal notranslate"><span class="pre">pkill</span></code> will send the signal to any process with the matching name. If that is not desirable, use <code class="docutils literal notranslate"><span class="pre">kill</span></code> to send it to a specific process. The standard output and error streams are redirected to <code class="docutils literal notranslate"><span class="pre">/tmp/stdout_<PID>.txt</span></code> and <code class="docutils literal notranslate"><span class="pre">/tmp/stderr_<PID>.txt</span></code>.</p>
</div>
<p>The launcher mechanism is more complex and less automated than the LD_PRELOAD option, but gives more control to the user.</p>
</section>
</section>
<section id="gui-troubleshooting">
<h3>GUI Troubleshooting<a class="headerlink" href="#gui-troubleshooting" title="Permalink to this heading"></a></h3>
<p>If opening the Nsight Systems Linux GUI fails with one of the following errors, you may be missing some required libraries:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">This</span><span class="w"> </span><span class="n">application</span><span class="w"> </span><span class="n">failed</span><span class="w"> </span><span class="n">to</span><span class="w"> </span><span class="n">start</span><span class="w"> </span><span class="n">because</span><span class="w"> </span><span class="n">it</span><span class="w"> </span><span class="n">could</span><span class="w"> </span><span class="n">not</span><span class="w"> </span><span class="n">find</span><span class="w"> </span><span class="n">or</span><span class="w"> </span><span class="n">load</span><span class="w"> </span><span class="n">the</span><span class="w"> </span><span class="n">Qt</span><span class="w"> </span><span class="n">platform</span><span class="w"> </span><span class="n">plugin</span><span class="w"> </span><span class="s">"xcb"</span><span class="w"> </span><span class="n">in</span><span class="w"> </span><span class="s">""</span><span class="p">.</span><span class="w"> </span><span class="n">Available</span><span class="w"> </span><span class="n">platform</span><span class="w"> </span><span class="n">plugins</span><span class="w"> </span><span class="n">are</span><span class="o">:</span><span class="w"> </span><span class="n">xcb</span><span class="p">.</span><span class="w"> </span><span class="n">Reinstalling</span><span class="w"> </span><span class="n">the</span><span class="w"> </span><span class="n">application</span><span class="w"> </span><span class="n">may</span><span class="w"> </span><span class="n">fix</span><span class="w"> </span><span class="k">this</span><span class="w"> </span><span class="n">problem</span><span class="p">.</span><span class="w"></span>
</pre></div>
</div>
<p>or</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">error</span><span class="w"> </span><span class="k">while</span><span class="w"> </span><span class="n">loading</span><span class="w"> </span><span class="n">shared</span><span class="w"> </span><span class="n">libraries</span><span class="o">:</span><span class="w"> </span><span class="p">[</span><span class="n">library_name</span><span class="p">]</span><span class="o">:</span><span class="w"> </span><span class="n">cannot</span><span class="w"> </span><span class="n">open</span><span class="w"> </span><span class="n">shared</span><span class="w"> </span><span class="n">object</span><span class="w"> </span><span class="n">file</span><span class="o">:</span><span class="w"> </span><span class="n">No</span><span class="w"> </span><span class="n">such</span><span class="w"> </span><span class="n">file</span><span class="w"> </span><span class="n">or</span><span class="w"> </span><span class="n">directory</span><span class="w"></span>
</pre></div>
</div>
<section id="ubuntu-18-04-20-04-22-04-and-centos-7-8-9-with-root-privileges">
<h4>Ubuntu 18.04/20.04/22.04 and CentOS 7/8/9 with root privileges<a class="headerlink" href="#ubuntu-18-04-20-04-22-04-and-centos-7-8-9-with-root-privileges" title="Permalink to this heading"></a></h4>
<ul>
<li><p>Launch the following command, which will install all the required libraries in system directories:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">installation_path</span><span class="p">]</span><span class="o">/</span><span class="n">host</span><span class="o">-</span><span class="n">linux</span><span class="o">-</span><span class="p">[</span><span class="n">arch</span><span class="p">]</span><span class="o">/</span><span class="n">Scripts</span><span class="o">/</span><span class="n">DependenciesInstaller</span><span class="o">/</span><span class="n">install</span><span class="o">-</span><span class="n">dependencies</span><span class="p">.</span><span class="n">sh</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Launch the Linux GUI as usual.</p></li>
</ul>
</section>
<section id="ubuntu-18-04-20-04-22-04-and-centos-7-8-9-without-root-privileges">
<h4>Ubuntu 18.04/20.04/22.04 and CentOS 7/8/9 without root privileges<a class="headerlink" href="#ubuntu-18-04-20-04-22-04-and-centos-7-8-9-without-root-privileges" title="Permalink to this heading"></a></h4>
<ul>
<li><p>Choose the directory where dependencies will be installed (<code class="docutils literal notranslate"><span class="pre">dependencies_path</span></code>). This directory should be writeable for the current user.</p></li>
<li><p>Launch the following command (if it has already been run, move to the next step), which will install all the required libraries in <code class="docutils literal notranslate"><span class="pre">[dependencies_path]</span></code>:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="p">[</span><span class="n">installation_path</span><span class="p">]</span><span class="o">/</span><span class="n">host</span><span class="o">-</span><span class="n">linux</span><span class="o">-</span><span class="p">[</span><span class="n">arch</span><span class="p">]</span><span class="o">/</span><span class="n">Scripts</span><span class="o">/</span><span class="n">DependenciesInstaller</span><span class="o">/</span><span class="n">install</span><span class="o">-</span><span class="n">dependencies</span><span class="o">-</span><span class="n">without</span><span class="o">-</span><span class="n">root</span><span class="p">.</span><span class="n">sh</span><span class="w"> </span><span class="p">[</span><span class="n">dependencies_path</span><span class="p">]</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Further, use the following command to launch the Linux GUI:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">source</span><span class="w"> </span><span class="p">[</span><span class="n">installation_path</span><span class="p">]</span><span class="o">/</span><span class="n">host</span><span class="o">-</span><span class="n">linux</span><span class="o">-</span><span class="p">[</span><span class="n">arch</span><span class="p">]</span><span class="o">/</span><span class="n">Scripts</span><span class="o">/</span><span class="n">DependenciesInstaller</span><span class="o">/</span><span class="n">setup</span><span class="o">-</span><span class="n">dependencies</span><span class="o">-</span><span class="n">environment</span><span class="p">.</span><span class="n">sh</span><span class="w"> </span><span class="p">[</span><span class="n">dependencies_path</span><span class="p">]</span><span class="w"> </span><span class="o">&&</span><span class="w"> </span><span class="p">[</span><span class="n">installation_path</span><span class="p">]</span><span class="o">/</span><span class="n">host</span><span class="o">-</span><span class="n">linux</span><span class="o">-</span><span class="n">x64</span><span class="o">/</span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="w"></span>
</pre></div>
</div>
</li>
</ul>
</section>
<section id="other-platforms-or-if-the-previous-steps-did-not-help">
<h4>Other platforms, or if the previous steps did not help<a class="headerlink" href="#other-platforms-or-if-the-previous-steps-did-not-help" title="Permalink to this heading"></a></h4>
<p>Launch Nsight Systems using the following command line to determine which libraries are missing and install them.</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">QT_DEBUG_PLUGINS</span><span class="o">=</span><span class="mi">1</span><span class="w"> </span><span class="p">.</span><span class="o">/</span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="w"></span>
</pre></div>
</div>
<p>If the workload does not run when launched via Nsight Systems or the timeline is empty, check the stderr.log and stdout.log (click on drop-down menu showing <strong>Timeline View</strong> and click on <strong>Files</strong>) to see the errors encountered by the app.</p>
<blockquote>
<div><img alt="Stderr Log" class="image" src="../_images/stderr-log.png" />
</div></blockquote>
</section>
</section>
<section id="symbol-resolution">
<h3>Symbol Resolution<a class="headerlink" href="#symbol-resolution" title="Permalink to this heading"></a></h3>
<p>If stack trace information is missing symbols and you have a symbol file, you can manually re-resolve using the ResolveSymbols utility. This can be done by right-clicking the report file in the Project Explorer window and selecting “Resolve Symbols…”.</p>
<p>Alternatively, you can find the utility as a separate executable in the <code class="docutils literal notranslate"><span class="pre">[installation_path]\Host</span></code> directory. This utility works with ELF format files, with Windows PDB directories and symbol servers, or with files where each line is in the format <code class="docutils literal notranslate"><span class="pre"><start><length><name></span></code>.</p>
<table class="table-no-stripes docutils align-default">
<thead>
<tr class="row-odd"><th class="head"><p>Short</p></th>
<th class="head"><p>Long</p></th>
<th class="head"><p>Argument</p></th>
<th class="head"><p>Description</p></th>
</tr>
</thead>
<tbody>
<tr class="row-even"><td><p>-h</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--help</span></code></p></td>
<td></td>
<td><p>Help message providing information about available options.</p></td>
</tr>
<tr class="row-odd"><td><p>-l</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--process-list</span></code></p></td>
<td></td>
<td><p>Print global process IDs list</p></td>
</tr>
<tr class="row-even"><td><p>-s</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--sym-file</span></code></p></td>
<td><p>filename</p></td>
<td><p>Path to symbol file</p></td>
</tr>
<tr class="row-odd"><td><p>-b</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--base-addr</span></code></p></td>
<td><p>address</p></td>
<td><p>If set then <start> in symbol file is treated as relative address starting from this base address</p></td>
</tr>
<tr class="row-even"><td><p>-p</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--global-pid</span></code></p></td>
<td><p>pid</p></td>
<td><p>Which process in the report should be resolved. May be omitted if there is only one process in the report.</p></td>
</tr>
<tr class="row-odd"><td><p>-f</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--force</span></code></p></td>
<td></td>
<td><p>This option forces use of a given symbol file.</p></td>
</tr>
<tr class="row-even"><td><p>-i</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--report</span></code></p></td>
<td><p>filename</p></td>
<td><p>Path to the report with unresolved symbols.</p></td>
</tr>
<tr class="row-odd"><td><p>-o</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--output</span></code></p></td>
<td><p>filename</p></td>
<td><p>Path and name of the output file. If it is omitted then “resolved” suffix is added to the original filename.</p></td>
</tr>
<tr class="row-even"><td><p>-d</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--directories</span></code></p></td>
<td><p>directory paths</p></td>
<td><p>List of symbol folder paths, separated by semi-colon characters. Available only on Windows.</p></td>
</tr>
<tr class="row-odd"><td><p>-v</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--servers</span></code></p></td>
<td><p>server URLs</p></td>
<td><p>List of symbol servers that uses the same format as <code class="docutils literal notranslate"><span class="pre">_NT_SYMBOL_PATH</span></code> environment variable, i.e. <code class="docutils literal notranslate"><span class="pre">srv*<LocalStore>*<SymbolServerURL></span></code>. Available only on Windows.</p></td>
</tr>
<tr class="row-even"><td><p>-n</p></td>
<td><p><code class="docutils literal notranslate"><span class="pre">--ignore-nt-sym-path</span></code></p></td>
<td></td>
<td><p>Ignore the symbol locations stored in the <code class="docutils literal notranslate"><span class="pre">_NT_SYMBOL_PATH</span></code> environment variable. Available only on Windows.</p></td>
</tr>
</tbody>
</table>
<section id="broken-backtraces-on-tegra">
<h4>Broken Backtraces on Tegra<a class="headerlink" href="#broken-backtraces-on-tegra" title="Permalink to this heading"></a></h4>
<p>In Nsight Systems Embedded Platforms Edition, in the symbols table there is a special entry called <strong>Broken backtraces</strong>. This entry is used to denote the point in the call chain where the unwinding algorithms used by Nsight Systems could not determine what is the next (caller) function.</p>
<p>Broken backtraces happen because there is no information related to the current function that the unwinding algorithms can use. In the Top-Down view, these functions are immediate children of the Broken backtraces row.</p>
<p>One can eliminate broken backtraces by modifying the build system to provide at least one kind of unwind information. The types of unwind information, used by the algorithms in Nsight Systems, include the following:</p>
<p>For ARMv7 binaries:</p>
<ul>
<li><p>DWARF information in ELF sections: <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.zdebug_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.eh_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.eh_frame_hdr</span></code>. This information is the most precise. <code class="docutils literal notranslate"><span class="pre">.zdebug_frame</span></code> is a compressed version of <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code>, so at most one of them is typically present. <code class="docutils literal notranslate"><span class="pre">.eh_frame_hdr</span></code> is a companion section for <code class="docutils literal notranslate"><span class="pre">.eh_frame</span></code> and might be absent.</p>
<p>Compiler flag: <code class="docutils literal notranslate"><span class="pre">-g</span></code>.</p>
</li>
<li><p>Exception handling information in EHABI format provided in <code class="docutils literal notranslate"><span class="pre">.ARM.exidx</span></code> and <code class="docutils literal notranslate"><span class="pre">.ARM.extab</span></code> ELF sections. <code class="docutils literal notranslate"><span class="pre">.ARM.extab</span></code> might be absent if all information is compact enough to be encoded into <code class="docutils literal notranslate"><span class="pre">.ARM.exidx</span></code>.</p>
<p>Compiler flag: <code class="docutils literal notranslate"><span class="pre">-funwind-tables</span></code>.</p>
</li>
<li><p>Frame pointers (built into the <code class="docutils literal notranslate"><span class="pre">.text</span></code> section).</p>
<p>Compiler flag: <code class="docutils literal notranslate"><span class="pre">-fno-omit-frame-pointer</span></code>.</p>
</li>
</ul>
<p>For Aarch64 binaries:</p>
<ul>
<li><p>DWARF information in ELF sections: <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.zdebug_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.eh_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.eh_frame_hdr</span></code>. See additional comments above.</p>
<p>Compiler flag: <code class="docutils literal notranslate"><span class="pre">-g</span></code>.</p>
</li>
<li><p>Frame pointers (built into the <code class="docutils literal notranslate"><span class="pre">.text</span></code> section).</p>
<p>Compiler flag: <code class="docutils literal notranslate"><span class="pre">-fno-omit-frame-pointer</span></code>.</p>
</li>
</ul>
<p>The following ELF sections should be considered empty if they have size of 4 bytes: <code class="docutils literal notranslate"><span class="pre">.debug_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.eh_frame</span></code>, <code class="docutils literal notranslate"><span class="pre">.ARM.exidx</span></code>. In this case, these sections only contain termination records and no useful information.</p>
<p>For GCC, use the following compiler invocation to see which compiler flags are enabled in your toolchain by default (for example, to check if <code class="docutils literal notranslate"><span class="pre">-funwind-tables</span></code> is enabled by default):</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="n">gcc</span><span class="w"> </span><span class="o">-</span><span class="n">Q</span><span class="w"> </span><span class="o">--</span><span class="n">help</span><span class="o">=</span><span class="n">common</span><span class="w"></span>
</pre></div>
</div>
<p>For GCC and Clang, add <code class="docutils literal notranslate"><span class="pre">-###</span></code> to the compiler invocation command to see which compiler flags are actually being used.</p>
<p>Since EHABI and DWARF information is compiled on per-unit basis (every <code class="docutils literal notranslate"><span class="pre">.cpp</span></code> or <code class="docutils literal notranslate"><span class="pre">.c</span></code> file, as well as every static library, can be built with or without this information), presence of the ELF sections does not guarantee that every function has necessary unwind information.</p>
<p>Frame pointers are required by the Aarch64 Procedure Call Standard. Adding frame pointers slows down execution time, but in most cases the difference is negligible.</p>
</section>
<section id="debug-versions-of-elf-files">
<h4>Debug Versions of ELF Files<a class="headerlink" href="#debug-versions-of-elf-files" title="Permalink to this heading"></a></h4>
<p>Often, after a binary is built, especially if it is built with debug information (<code class="docutils literal notranslate"><span class="pre">-g</span></code> compiler flag), it gets stripped before deploying or installing. In this case, ELF sections that contain useful information, such as non-export function names or unwind information, can get stripped as well.</p>
<p>One solution is to deploy or install the original unstripped library instead of the stripped one, but in many cases this would be inconvenient. Nsight Systems can use missing information from alternative locations.</p>
<p>For target devices with Ubuntu, see <a class="reference external" href="https://wiki.ubuntu.com/Debug_Symbol_packages">Debug Symbol Packages</a>. These packages typically install debug ELF files with <code class="docutils literal notranslate"><span class="pre">/usr/lib/debug</span></code> prefix. Nsight Systems can find debug libraries there, and if it matches the original library (e.g., the built-in <code class="docutils literal notranslate"><span class="pre">BuildID</span></code> is the same), it will be picked up and used to provide symbol names and unwind information.</p>
<p>Many packages have debug companions in the same repository and can be directly installed with APT (<code class="docutils literal notranslate"><span class="pre">apt-get</span></code>). Look for packages with the <code class="docutils literal notranslate"><span class="pre">-dbg</span></code> suffix. For other packages, refer to the <a class="reference external" href="https://wiki.ubuntu.com/Debug_Symbol_packages">Debug Symbol Packages</a> wiki page on how to add the debs package repository. After setting up the repository and running apt-get update, look for packages with <code class="docutils literal notranslate"><span class="pre">-dbgsym</span></code> suffix.</p>
<p>To verify that a debug version of a library has been picked up and downloaded from the target device, look in the <strong>Module Summary</strong> section of <strong>Analysis Summary</strong>:</p>
<blockquote>
<div><img alt="Debug library has been used" class="image" src="../_images/debug-library-used.png" />
</div></blockquote>
</section>
</section>
<section id="logging">
<h3>Logging<a class="headerlink" href="#logging" title="Permalink to this heading"></a></h3>
<p>To enable logging on the host, refer to this config file:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">host</span><span class="o">-</span><span class="n">linux</span><span class="o">-</span><span class="n">x64</span><span class="o">/</span><span class="n">nvlog</span><span class="p">.</span><span class="n">config</span><span class="p">.</span><span class="k">template</span><span class="w"></span>
</pre></div>
</div>
<p>When reporting any bugs please include the build version number as described in the <strong>Help → About</strong> dialog. If possible, attach log files and report (<code class="docutils literal notranslate"><span class="pre">.nsys-rep</span></code>) files, as they already contain necessary version information.</p>
<section id="verbose-remote-logging-on-linux-targets">
<h4>Verbose Remote Logging on Linux Targets<a class="headerlink" href="#verbose-remote-logging-on-linux-targets" title="Permalink to this heading"></a></h4>
<p>Verbose logging is available when connecting to a Linux-based device from the GUI on the host. This extra debug information is not available when launching via the command line. Nsight Systems installs its executable and library files into the following directory:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="w"></span>
</pre></div>
</div>
<p>To enable verbose logging on the target device, when launched from the host, follow these steps:</p>
<ol class="arabic">
<li><p>Close the host application.</p></li>
<li><p>Restart the target device.</p></li>
<li><p>Place <code class="docutils literal notranslate"><span class="pre">nvlog.config</span></code> from host directory to the <code class="docutils literal notranslate"><span class="pre">/opt/nvidia/nsight_systems</span></code> directory on target.</p></li>
<li><p>From SSH console, launch the following command:</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">sudo</span><span class="w"> </span><span class="o">/</span><span class="n">opt</span><span class="o">/</span><span class="n">nvidia</span><span class="o">/</span><span class="n">nsight_systems</span><span class="o">/</span><span class="n">nsys</span><span class="w"> </span><span class="o">--</span><span class="n">daemon</span><span class="w"> </span><span class="o">--</span><span class="n">debug</span><span class="w"></span>
</pre></div>
</div>
</li>
<li><p>Start the host application and connect to the target device.</p></li>
</ol>
<p>Logs on the target devices are collected into this file (if enabled):</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span><span class="p">.</span><span class="n">log</span><span class="w"></span>
</pre></div>
</div>
<p>in the directory where <code class="docutils literal notranslate"><span class="pre">nsys</span></code> command was launched.</p>
<p>Please note that in some cases, debug logging can significantly slow down the profiler.</p>
</section>
<section id="verbose-cli-logging-on-linux-targets">
<h4>Verbose CLI Logging on Linux Targets<a class="headerlink" href="#verbose-cli-logging-on-linux-targets" title="Permalink to this heading"></a></h4>
<p>To enable verbose logging of the Nsight Systems CLI and the target application’s injection behavior:</p>
<ol class="arabic simple">
<li><p>In the target-linux-x64 directory, rename the nvlog.config.template file to nvlog.config.</p></li>
<li></li>
</ol>
<p>Inside that file, change the line</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="p">}}{{{}</span><span class="n">nsys</span><span class="o">-</span><span class="n">ui</span><span class="p">.</span><span class="n">log</span><span class="w"></span>
</pre></div>
</div>
<p> to</p>
<div class="highlight-c++ notranslate"><div class="highlight"><pre><span></span><span class="n">$</span><span class="w"> </span><span class="p">}}{{{}</span><span class="n">nsys</span><span class="o">-</span><span class="n">agent</span><span class="p">.</span><span class="n">log</span><span class="w"></span>
</pre></div>
</div>
<ol class="arabic simple">
<li><p>Run a collection and the <code class="docutils literal notranslate"><span class="pre">target-linux.x64</span></code> directory should include a file named <code class="docutils literal notranslate"><span class="pre">nsys-agent.log</span></code>.</p></li>
</ol>
<p>Please note that in some cases, debug logging can significantly slow down the profiler.</p>
</section>
<section id="verbose-logging-on-windows-targets">
<h4>Verbose Logging on Windows Targets<a class="headerlink" href="#verbose-logging-on-windows-targets" title="Permalink to this heading"></a></h4>
<p>Verbose logging is available when connecting to a Windows-based device from the GUI on the host. Nsight Systems installs its executable and library files into the following directory by default:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>C:\Program Files\NVIDIA Corporation\Nsight Systems 2023.3
</pre></div>
</div>
<p>To enable verbose logging on the target device, when launched from the host, follow these steps:</p>
<ol class="arabic">
<li><p>Close the host application.</p></li>
<li><p>Terminate the <code class="docutils literal notranslate"><span class="pre">nsys</span></code> process.</p></li>
<li><p>Place <code class="docutils literal notranslate"><span class="pre">nvlog.config</span></code> from host directory next to Nsight Systems Windows agent on the target device</p>
<ul>
<li><p>Local Windows target:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>C:\Program Files\NVIDIA Corporation\Nsight Systems 2023.3\target-windows-x64
</pre></div>
</div>
</li>
<li><p>Remote Windows target:</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>C:\Users\<user name>\AppData\Local\Temp\nvidia\nsight_systems
</pre></div>
</div>
</li>
</ul>
</li>
<li><p>Start the host application and connect to the target device.</p></li>
</ol>
<p>Logs on the target devices are collected into this file (if enabled):</p>
<div class="highlight-text notranslate"><div class="highlight"><pre><span></span>nsight-sys.log
</pre></div>
</div>
<p>in the same directory as Nsight Systems Windows agent.</p>
<p>Please note that in some cases debug logging can significantly slow down the profiler.</p>
</section>
</section>
</section>
<section id="other-resources">
<h2>Other Resources<a class="headerlink" href="#other-resources" title="Permalink to this heading"></a></h2>
<p>Looking for information to help you use Nsight Systems the most effectively? Here are some more resources you might want to review:</p>
<section id="training-seminars">
<h3>Training Seminars<a class="headerlink" href="#training-seminars" title="Permalink to this heading"></a></h3>
<p>NVIDIA Deep Learning Institute Training - Self-Paced Online Course <a class="reference external" href="https://courses.nvidia.com/courses/course-v1:DLI+S-AC-03+V1/">Optimizing CUDA Machine Learning Codes With Nsight Profiling Tools</a></p>
<p>2018 NCSA Blue Waters Webinar - Video Only <a class="reference external" href="https://www.youtube.com/watch?v=WA8C48FJi3c">Introduction to NVIDIA Nsight Systems</a></p>
</section>
<section id="blog-posts">
<h3>Blog Posts<a class="headerlink" href="#blog-posts" title="Permalink to this heading"></a></h3>
<p>NVIDIA developer blogs, these are longer form, technical pieces written by tool and domain experts.</p>
<ul class="simple">
<li><p>2021 - <a class="reference external" href="https://developer.nvidia.com/blog/optimizing-dx12-resource-uploads-to-the-gpu-using-cpu-visible-vram/">Optimizing DX12 Resource Uploads to the GPU Using CPU-Visible VRAM</a></p></li>
<li><p>2019 - <a class="reference external" href="https://devblogs.nvidia.com/migrating-nvidia-nsight-tools-nvvp-nvprof/">Migrating to NVIDIA Nsight Tools from NVVP and nvprof</a></p></li>
<li><p>2019 - <a class="reference external" href="https://devblogs.nvidia.com/transitioning-nsight-systems-nvidia-visual-profiler-nvprof/">Transitioning to Nsight Systems from NVIDIA Visual Profiler / nvprof</a></p></li>
<li><p>2019 - <a class="reference external" href="https://devblogs.nvidia.com/nvidia-nsight-systems-adds-vulkan-support/">NVIDIA Nsight Systems Add Vulkan Support</a></p></li>
<li><p>2019 - <a class="reference external" href="https://devblogs.nvidia.com/tensorflow-performance-logging-plugin-nvtx-plugins-tf-public/">TensorFlow Performance Logging Plugin nvtx-plugins-tf Goes Public</a></p></li>
<li><p>2020 - <a class="reference external" href="https://developer.nvidia.com/blog/understanding-the-visualization-of-overhead-and-latency-in-nsight-systems/">Understanding the Visualization of Overhead and Latency in Nsight Systems</a></p></li>
<li><p>2021 - <a class="reference external" href="https://developer.nvidia.com/blog/optimizing-dx12-resource-uploads-to-the-gpu-using-cpu-visible-vram/">Optimizing DX12 Resource Uploads to the GPU Using CPU-Visible VRAM</a></p></li>
</ul>
</section>
<section id="feature-videos">
<h3>Feature Videos<a class="headerlink" href="#feature-videos" title="Permalink to this heading"></a></h3>
<p>Short videos, only a minute or two, to introduce new features.</p>
<ul class="simple">
<li><p><a class="reference external" href="https://youtu.be/ZeuM2k_hrq0">OpenMP Trace Feature Spotlight</a></p></li>
<li><p><a class="reference external" href="https://youtu.be/r2ewwd4d0vc">Command Line Sessions Video Spotlight</a></p></li>
<li><p><a class="reference external" href="https://youtu.be/DUhzjyBr-wg">Direct3D11 Feature Spotlight</a></p></li>
<li><p><a class="reference external" href="https://youtu.be/witzRF-wu8M">Vulkan Trace</a></p></li>
<li><p><a class="reference external" href="https://www.youtube.com/watch?v=fyhPFTF75tk">Statistics Driven Profiling</a></p></li>
<li><p><a class="reference external" href="https://www.youtube.com/embed/bda5Er27jqo">Analyzing NCCL Usage with NVDIA Nsight Systems</a></p></li>
</ul>
</section>
<section id="conference-presentations">
<h3>Conference Presentations<a class="headerlink" href="#conference-presentations" title="Permalink to this heading"></a></h3>
<ul class="simple">
<li><p>GTC 2023 <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring23-d52660/">Optimize Multi-Node System Workloads With NVIDIA Nsight Systems</a></p></li>
<li><p>GTC 2023 <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring23-dlit51580/">Ray-Tracing Development using NVIDIA Nsight Graphics and NVIDIA Nsight Systems</a></p></li>
<li><p>GTC 2022 - <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring22-s41518/">Killing Cloud Monsters Has Never Been Smoother</a></p></li>
<li><p>GTC 2022 - <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring22-s41500/">Optimizing Communication with Nsight Systems Network Profiling</a></p></li>
<li><p>GTC 2022 - <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/other2021-vlk2201/">Optimizing Vulkan 1.3 Applications with Nsight Graphics & Nsight Systems</a></p></li>
<li><p>GTC 2021 - <a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring21-s31566/">Tuning GPU Network and Memory Usage in Apache Spark</a></p></li>
<li><p>GTC 2020 - <a class="reference external" href="https://developer.nvidia.com/gtc/2020/video/s21547">Rebalancing the Load: Profile-Guided Optimization of the NAMD Molecular Dynamics Program for Modern GPUs using Nsight Systems</a></p></li>
<li><p>GTC 2020 - <a class="reference external" href="https://developer.nvidia.com/gtc/2020/video/s21351">Scaling the Transformer Model Implementation in PyTorch Across Multiple Nodes</a></p></li>
<li><p>GTC 2019 - <a class="reference external" href="https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=s9503-using+nsight+tools+to+optimize+the+namd+molecular+dynamics+simulation+program">Using Nsight Tools to Optimize the NAMD Molecular Dynamics Simulation Program</a></p></li>
<li><p>GTC 2019 - <a class="reference external" href="https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=s9866-optimizing+facebook+ai+workloads+for+nvidia+gpus">Optimizing Facebook AI Workloads for NVIDIA GPUs</a></p></li>
<li><p>GTC 2018 - <a class="reference external" href="https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=s8718-optimizing+hpc+simulation+and+visualization+codes+using+the+nvidia+nsight+systems">Optimizing HPC Simulation and Visualization Codes Using NVIDIA Nsight Systems</a></p></li>
<li><p>GTC 2018 - Israel - <a class="reference external" href="https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=sil8105-boost+dnn+training+performance+using+nvidia+tools">Boost DNN Training Performance using NVIDIA Tools</a></p></li>
<li><p>Siggraph 2018 - <a class="reference external" href="http://on-demand.gputechconf.com/siggraph/2018/video/sig1805-aurelio-reis-yaki-tebeka-taming-the-beast.html">Taming the Beast; Using NVIDIA Tools to Unlock Hidden GPU Performance</a></p></li>
</ul>
</section>
<section id="for-more-support">
<h3>For More Support<a class="headerlink" href="#for-more-support" title="Permalink to this heading"></a></h3>
<p>To file a bug report or to ask a question on the Nsight Systems forums, you will need to register with the NVIDIA Developer Program. See the <a class="reference external" href="https://www.nvidia.com/en-us/account/faq/">FAQ</a>. You do not need to register to read the forums.</p>
<p>After that, you can access Nsight Systems <a class="reference external" href="https://forums.developer.nvidia.com/c/development-tools/nsight-systems/116">Forums</a> and the <a class="reference external" href="https://developer.nvidia.com/nvidia_bug/add">NVIDIA Bug Tracking System</a>.</p>
<p>To submit feedback directly from the GUI, go to <strong>Help->Send Feedback</strong> and fill out the form. Enter your email address if you would like to hear back from the Nsight Systems team.</p>
<blockquote>
<div><img alt="Feedback" class="image" src="../_images/feedback.png" />
</div></blockquote>
</section>
</section>
</section>
</div>
</div>
<footer>
<hr/>
<div role="contentinfo">
<p>© Copyright 2018-2023, NVIDIA Corporation & Affiliates. All rights reserved.
<span class="lastupdated">Last updated on Dec 20, 2023.
</span></p>
</div>
</footer>
</div>
</div>
</section>
</div>
<script>
jQuery(function () {
SphinxRtdTheme.Navigation.enable(true);
});
</script>
</body>
</html>
|
