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
|
/*___INFO__MARK_BEGIN__*/
/*************************************************************************
*
* The Contents of this file are made available subject to the terms of
* the Sun Industry Standards Source License Version 1.2
*
* Sun Microsystems Inc., March, 2001
*
*
* Sun Industry Standards Source License Version 1.2
* =================================================
* The contents of this file are subject to the Sun Industry Standards
* Source License Version 1.2 (the "License"); You may not use this file
* except in compliance with the License. You may obtain a copy of the
* License at http://gridengine.sunsource.net/Gridengine_SISSL_license.html
*
* Software provided under this License is provided on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING,
* WITHOUT LIMITATION, WARRANTIES THAT THE SOFTWARE IS FREE OF DEFECTS,
* MERCHANTABLE, FIT FOR A PARTICULAR PURPOSE, OR NON-INFRINGING.
* See the License for the specific provisions governing your rights and
* obligations concerning the Software.
*
* The Initial Developer of the Original Code is: Sun Microsystems, Inc.
*
* Copyright: 2001 by Sun Microsystems, Inc.
*
* All Rights Reserved.
*
************************************************************************/
/*___INFO__MARK_END__*/
#include <stdarg.h>
#include <time.h>
#include <sys/time.h>
#include <sys/times.h>
#include <string.h>
#include <limits.h>
#include <unistd.h>
#include <pthread.h>
#include "basis_types.h"
#include "uti/sge_rmon.h"
#include "uti/sge_dstring.h"
#include "uti/sge_log.h"
#include "uti/sge_time.h"
#include "uti/sge_string.h"
#include "uti/msg_utilib.h"
#include "uti/sge_profiling.h"
/****** uti/profiling/--Profiling ****************************************
* NAME
* Profiling -- A simple profiling utility
*
* FUNCTION
* The profiling module provides a number of utility functions
* to output some performance metrics about a program, like wallclock,
* busy time, user and system cpu time etc.
*
* Profiling can be started and stopped, the measured data can be
* reset to 0.
*
* Profiling for individual code blocks can be done by starting
* and stopping a measurement.
*
* Measured data can be queried individually for each metric variable,
* or an informational message string can be requested containing the
* most interesting variables.
*
* To distinguish profiling data for different modules of a program,
* (e.g. time/cpu consumed for communication, spooling, etc.),
* multiple levels have been introduced.
*
* Besides the predefined levels, there exist 10 custom levels
* (SGE_PROF_CUSTOM0 up to SGE_PROF_CUSTOM9) that can
* be used to profile individual parts of a program.
*
* The predefined levels shall only be used by the developers of the
* corresponding components.
*
* The level SGE_PROF_OTHER is maintained by the profiling module itself
* and collects the remaining usage not covered by other levels. It shall
* NOT be used outside the profiling module itself.
*
* A measurement is always done for a certain level.
* If a measurement for a certain level is started, while another
* measurement for another level is still active, the data of the
* subordinated measurement is remembered in the superordinated level.
*
* When retrieving profiling information, a parameter defines whether
* information for subordinated measurements shall be included or excluded.
*
* Cyclic starting of measurements is not allowed and will result in
* profiling being switched off.
*
* EXAMPLE
* There is a test program (libs/uti/test_sge_profiling) and the
* module is used in sge_schedd (daemons/schedd/scheduler.c),
* in the event mirror module (libs/gdi/sge_mirror.c), the qmaster
* and his threads, execd and the spooling
* test program (libs/spool/test_sge_spooling.c).
*
* SEE ALSO
* uti/profiling/prof_is_active()
* uti/profiling/prof_start()
* uti/profiling/prof_stop()
* uti/profiling/prof_start_measurement()
* uti/profiling/prof_stop_measurement()
* uti/profiling/prof_reset()
* uti/profiling/prof_get_measurement_wallclock()
* uti/profiling/prof_get_measurement_utime()
* uti/profiling/prof_get_measurement_stime()
* uti/profiling/prof_get_total_wallclock()
* uti/profiling/prof_get_total_busy()
* uti/profiling/prof_get_total_utime()
* uti/profiling/prof_get_total_stime()
* uti/profiling/prof_get_info_string()
*
* NOTES
* MT-NOTE: this module is MT safe, if prof_mt_init() and/or
* sge_prof_set_enabled() are called before
* profiling is started and sge_prof_cleanup() is called after
* all threads which make profiling calls have been stopped!
* sge_prof_cleanup() is used to free the profiling array
*
* BUGS
* In a multithreaded program, the values delivered by the times system call
* for cpu system and user time are per process times.
* This module assumes that the values delivered by times are per thread
* values.
* In the profiling output, the sum of user and system time can be higher
* than the wallclock time, the utilization greater 100%.
*
* TODO
* - Replace all pthread_mutex_(un)lock() calls with sge_mutex_(un)lock()
* - Separate public and private interfaces so that thread index can be
* passed around instead of always being rediscovered
* - Replace static array with thread specific storage?
* - Document static functions
* - Print levels without names
*******************************************************************************/
/****** uti/profiling/-Profiling-Defines ***************************************
* NAME
* Defines -- Defines and macros for profiling
*
* SYNOPSIS
* #define PROF_START_MEASUREMENT ...
* #define PROF_STOP_MEASUREMENT ...
*
* FUNCTION
* Macros to start and stop a profiling measurement.
* They test if profiling is enabled at all and then call the
* functions to start or stop a measurement.
*
* SEE ALSO
* uti/profiling/prof_mt_init()
* uti/profiling/sge_prof_cleanup()
* uti/profiling/prof_is_active()
* uti/profiling/prof_start_measurement()
* uti/profiling/prof_stop_measurement()
*******************************************************************************/
/* this struct is used for pthread name, thread id mapping
It also holds the profiling status information for each thread */
typedef struct {
const char* thrd_name;
pthread_t thrd_id;
bool prof_is_active;
int is_terminated;
} sge_thread_info_t;
static int sge_prof_array_initialized = 0;
static void prof_info_init(prof_level level);
static void prof_info_level_init(prof_level i, int thread_num);
static void prof_reset_thread(int thread_num, prof_level level);
static void init_array(pthread_t num);
static void init_array_first(void);
static void init_thread_info(void);
static int get_prof_info_thread_id(void);
static sge_prof_info_t **theInfo = NULL;
static sge_thread_info_t *thrdInfo = NULL;
static pthread_mutex_t thrdInfo_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_key_t thread_id_key;
static bool profiling_enabled = true;
static pthread_once_t prof_once = PTHREAD_ONCE_INIT;
/*
* TODO: If application starts more than MAX_THREAD_NUM threads,
* this value must be increased:
*
*/
static const int MAX_THREAD_NUM = 64;
/****** uti/profiling/sge_prof_set_enabled() ***********************************
* NAME
* sge_prof_set_enabled() -- enables/disables profiling
*
* SYNOPSIS
* void sge_prof_set_enabled(bool enabled)
*
* FUNCTION
* Enables/disables profiling completely. Profiling is enabled by
* default. This method is the fix to Issue 1471.
*
* INPUTS
* bool enabled Whether profiling should be enabled
* RESULT
*
* NOTES
* MT-NOTE: sge_prof_set_enabled() is MT safe only if called before any other
* profiling calls
*
* SEE ALSO
*******************************************************************************/
void sge_prof_set_enabled(bool enabled) {
profiling_enabled = enabled;
}
/****** uti/profiling/prof_thread_local_once_init() ************************************
* NAME
* prof_thread_local_once_init() -- inititalizes the profiling array
*
* SYNOPSIS
* void prof_thread_local_once_init(void)
*
* FUNCTION
* Initializes the profiling array.
*
* INPUTS
*
* RESULT
*
* NOTES
* MT-NOTE: prof_thread_local_once_init() is MT safe only if called before any other
* profiling calls other than sge_prof_set_enabled()
*
* SEE ALSO
*******************************************************************************/
static void prof_thread_local_once_init(void)
{
if (!profiling_enabled) {
return;
}
init_thread_info();
init_array_first();
init_array(pthread_self());
}
void prof_mt_init(void) {
pthread_once(&prof_once, prof_thread_local_once_init);
}
/****** uti/profiling/prof_set_level_name() ************************************
* NAME
* prof_set_level_name() -- set name of a custom level
*
* SYNOPSIS
* bool prof_set_level_name(prof_level level, const char *name)
*
* FUNCTION
* Set the name of a custom profiling level.
*
* INPUTS
* prof_level level - level to edit
* const char *name - new name for level
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_set_level_name() is MT safe
*
* SEE ALSO
* uti/profiling/--Profiling
*******************************************************************************/
bool prof_set_level_name(prof_level level, const char *name, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_set_level_name", level);
ret = false;
} else if (name == NULL) {
sge_dstring_sprintf_append(error, MSG_PROF_NULLLEVELNAME_S, "prof_set_level_name");
ret = false;
} else {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
ret = false;
} else {
theInfo[thread_num][level].name = name;
}
}
return ret;
}
/****** uti/profiling/prof_is_active() ****************************************
* NAME
* prof_is_active() -- is profiling active?
*
* SYNOPSIS
* bool prof_is_active(void)
*
* FUNCTION
* Returns true, if profiling is active, else false.
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* SEE ALSO
* uti/profiling/prof_stop()
* uti/profiling/prof_stop()
*
* NOTES
* MT-NOTE: prof_is_active() is MT safe
*******************************************************************************/
bool prof_is_active(prof_level level)
{
int thread_num;
pthread_t thread_id;
bool ret = false;
if (profiling_enabled && (level <= SGE_PROF_ALL)) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num >= 0) && (thread_num < MAX_THREAD_NUM)) {
ret = theInfo[thread_num][level].prof_is_started;
}
}
return ret;
}
/****** uti/profiling/prof_start() ****************************************
* NAME
* prof_start() -- start profiling
*
* SYNOPSIS
* bool prof_start(dstring *error)
*
* FUNCTION
* Enables profiling. All internal variables are reset to 0.
* Performance measurement has to be started and stopped by
* calling profiling_start_measurement or profiling_stop_measurement.
*
* INPUTS
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* SEE ALSO
* uti/profiling/prof_stop()
* uti/profiling/prof_start_measurement()
* uti/profiling/prof_stop_measurement()
*
* NOTES
* MT-NOTE: prof_start() is MT safe
*******************************************************************************/
bool prof_start(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_start", level);
ret = false;
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_start");
ret = false;
} else if (theInfo[thread_num][level].prof_is_started) {
sge_dstring_sprintf_append(error, MSG_PROF_ALREADYACTIVE_S, "prof_start");
ret = false;
} else {
struct tms tms_buffer;
prof_level i;
clock_t start_time = times(&tms_buffer);
if (level == SGE_PROF_ALL) {
for (i = SGE_PROF_OTHER; i <= SGE_PROF_ALL; i++) {
/* get start time */
theInfo[thread_num][i].start_clock = start_time;
/* initialize names and reset all data */
ret = prof_reset(i, error);
theInfo[thread_num][i].prof_is_started = true;
theInfo[thread_num][i].ever_started = true;
}
} else {
/* get start time */
theInfo[thread_num][level].start_clock = start_time;
/* initialize names and reset all data */
ret = prof_reset(level, error);
theInfo[thread_num][level].prof_is_started = true;
/* TODO
* Is there a reason we do this? This flag in never cleared. */
theInfo[thread_num][SGE_PROF_ALL].prof_is_started = true;
theInfo[thread_num][level].ever_started = true;
}
/* we have no actual profiling level */
theInfo[thread_num][SGE_PROF_ALL].akt_level = SGE_PROF_NONE;
/* implicitly start the OTHER level */
prof_start_measurement(SGE_PROF_OTHER, error);
}
}
return ret;
}
/****** uti/profiling/prof_stop() *****************************************
* NAME
* prof_stop() -- stop profiling
*
* SYNOPSIS
* bool prof_stop(dstring *error)
*
* FUNCTION
* Profiling is disabled.
*
* Subsequent calls to profiling_start_measurement or
* profiling_stop_measurement will have no effect.
*
* Profiling can be re-enabled by calling profiling_start.
*
* INPUTS
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* SEE ALSO
* uti/profiling/prof_start()
* uti/profiling/prof_start_measurement()
* uti/profiling/prof_stop_measurement()
*
* NOTES
* MT-NOTE: prof_start() is MT safe
*******************************************************************************/
bool prof_stop(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
prof_level i;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_stop", level);
ret = false;
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_stop");
ret = false;
} else if (!theInfo[thread_num][level].prof_is_started) {
sge_dstring_sprintf_append(error, MSG_PROF_NOTACTIVE_S, "prof_stop");
ret = false;
} else {
prof_stop_measurement(SGE_PROF_OTHER, error);
if (level == SGE_PROF_ALL) {
for (i = SGE_PROF_OTHER; i <= SGE_PROF_ALL; i++) {
theInfo[thread_num][i].prof_is_started = false;
}
} else {
theInfo[thread_num][level].prof_is_started = false;
}
}
}
return ret;
}
/****** uti/profiling/prof_start_measurement() ****************************
* NAME
* prof_start_measurement() -- start measurement
*
* SYNOPSIS
* bool prof_start_measurement(prof_level level, dstring *error)
*
* FUNCTION
* Starts measurement of performance data.
* Retrieves and stores current time and usage information.
*
* INPUTS
* prof_level level - level to process
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* SEE ALSO
* uti/profiling/prof_stop_measurement()
*
* NOTES
* MT-NOTE: prof_start_measurement() is MT safe
*******************************************************************************/
bool prof_start_measurement(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_start_measurement", level);
ret = false;
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_start_measurement");
ret = false;
} else if (!theInfo[thread_num][level].prof_is_started) {
sge_dstring_sprintf_append(error, MSG_PROF_NOTACTIVE_S, "prof_start_measurement");
ret = false;
} else if (theInfo[thread_num][SGE_PROF_ALL].akt_level == level) {
/* multiple start_measurement calls within one level are allowed */
theInfo[thread_num][level].nested_calls++;
} else if (theInfo[thread_num][level].pre != SGE_PROF_NONE) {
/* we cannot yet handle cyclic measurements between multiple levels
* produce an error and stop profiling
*/
sge_dstring_sprintf_append(error, MSG_PROF_CYCLICNOTALLOWED_SD, "prof_start_measurement", level);
prof_stop(level, error);
ret = false;
} else {
theInfo[thread_num][level].pre = theInfo[thread_num][SGE_PROF_ALL].akt_level;
theInfo[thread_num][SGE_PROF_ALL].akt_level = level;
theInfo[thread_num][level].start = times(&(theInfo[thread_num][level].tms_start));
/* when we start a level, we have no sub usage */
theInfo[thread_num][level].sub = 0;
theInfo[thread_num][level].sub_utime = 0;
theInfo[thread_num][level].sub_utime= 0;
}
}
return ret;
}
/****** uti/profiling/prof_stop_measurement() *****************************
* NAME
* prof_stop_measurement() -- stop measurement
*
* SYNOPSIS
* bool prof_stop_measurement(prof_level level, dstring *error)
*
* FUNCTION
* Stops measurement for a certain code block.
* Retrieves and stores current time and usage information.
* Sums up global usage information.
*
* INPUTS
* prof_level level - level to process
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* SEE ALSO
* uti/profiling/prof_start_measurement()
*
* NOTES
* MT-NOTE: prof_stop_measurement() is MT safe
*******************************************************************************/
bool prof_stop_measurement(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_stop_measurement", level);
ret = false;
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_stop_measurement");
ret = false;
} else if (!theInfo[thread_num][level].prof_is_started) {
sge_dstring_sprintf_append(error, MSG_PROF_NOTACTIVE_S, "prof_stop_measurement");
ret = false;
} else {
clock_t time, utime, stime;
if (theInfo[thread_num][level].nested_calls > 0) {
theInfo[thread_num][level].nested_calls--;
} else {
theInfo[thread_num][level].end = times(&(theInfo[thread_num][level].tms_end));
time = theInfo[thread_num][level].end - theInfo[thread_num][level].start;
utime = theInfo[thread_num][level].tms_end.tms_utime - theInfo[thread_num][level].tms_start.tms_utime;
stime = theInfo[thread_num][level].tms_end.tms_stime - theInfo[thread_num][level].tms_start.tms_stime;
#if 0
if (time < (utime + stime)) {
DPRINTF(("---> utime + stime > time, difference is %d clock ticks\n", (utime + stime) - time));
}
#endif
theInfo[thread_num][level].total += time;
theInfo[thread_num][level].total_utime += utime;
theInfo[thread_num][level].total_stime += stime;
if (theInfo[thread_num][level].pre != SGE_PROF_NONE) {
prof_level pre = theInfo[thread_num][level].pre;
theInfo[thread_num][pre].sub += time;
theInfo[thread_num][pre].sub_utime += utime;
theInfo[thread_num][pre].sub_stime += stime;
theInfo[thread_num][pre].sub_total += time;
theInfo[thread_num][pre].sub_total_utime += utime;
theInfo[thread_num][pre].sub_total_stime += stime;
theInfo[thread_num][SGE_PROF_ALL].akt_level = theInfo[thread_num][level].pre;
theInfo[thread_num][level].pre = SGE_PROF_NONE;
} else {
theInfo[thread_num][SGE_PROF_ALL].akt_level = SGE_PROF_NONE;
}
}
}
}
return ret;
}
/****** uti/profiling/prof_reset() ****************************************
* NAME
* prof_reset() -- reset usage information
*
* SYNOPSIS
* bool prof_reset(dstring *error)
*
* FUNCTION
* Reset usage and timing information to 0.
*
* INPUTS
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* bool - true on success, else false is returned and an error message
* is returned in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_reset() is MT safe
*******************************************************************************/
bool prof_reset(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
bool ret = true;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_reset", level);
ret = false;
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_reset");
ret = false;
} else if (theInfo[thread_num][SGE_PROF_ALL].akt_level > SGE_PROF_OTHER) {
sge_dstring_sprintf_append(error, MSG_PROF_RESETWHILEMEASUREMENT_S, "prof_reset");
ret = false;
} else {
if (theInfo[thread_num][level].prof_is_started) {
ret = prof_stop_measurement(SGE_PROF_OTHER, error);
}
if (level == SGE_PROF_ALL) {
prof_level c;
for (c = SGE_PROF_OTHER; c <= SGE_PROF_ALL; c++) {
prof_reset_thread (thread_num, c);
}
} else {
prof_reset_thread (thread_num, level);
}
if (theInfo[thread_num][level].prof_is_started) {
ret = prof_start_measurement(SGE_PROF_OTHER, error);
}
}
}
return ret;
}
/****** uti/profiling/prof_reset_thread() **************************************
* NAME
* prof_reset_thread() -- reset usage information
*
* SYNOPSIS
* bool prof_reset_thread(int thread_num, prof_level level)
*
* FUNCTION
* Reset usage and timing information for a single thread and level to 0.
*
* INPUTS
* int thread_num the thread to reset
* prof_level level the profiling level to reset
*
* NOTES
* MT-NOTE: prof_reset() is MT safe
*******************************************************************************/
static void prof_reset_thread(int thread_num, prof_level level)
{
struct tms tms_buffer;
theInfo[thread_num][level].start = 0;
theInfo[thread_num][level].end = 0;
theInfo[thread_num][level].tms_start.tms_utime = 0;
theInfo[thread_num][level].tms_start.tms_stime = 0;
theInfo[thread_num][level].tms_start.tms_cutime = 0;
theInfo[thread_num][level].tms_start.tms_cstime = 0;
theInfo[thread_num][level].tms_end.tms_utime = 0;
theInfo[thread_num][level].tms_end.tms_stime = 0;
theInfo[thread_num][level].tms_end.tms_cutime = 0;
theInfo[thread_num][level].tms_end.tms_cstime = 0;
theInfo[thread_num][level].total = 0;
theInfo[thread_num][level].total_utime = 0;
theInfo[thread_num][level].total_stime = 0;
theInfo[thread_num][level].pre = SGE_PROF_NONE;
theInfo[thread_num][level].sub = 0;
theInfo[thread_num][level].sub_utime = 0;
theInfo[thread_num][level].sub_stime = 0;
theInfo[thread_num][level].sub_total = 0;
theInfo[thread_num][level].sub_total_utime = 0;
theInfo[thread_num][level].sub_total_stime = 0;
theInfo[thread_num][level].start_clock = times(&tms_buffer);
}
/****** uti/profiling/prof_get_measurement_wallclock() ********************
* NAME
* prof_get_measurement_wallclock() -- return wallclock of a measurement
*
* SYNOPSIS
* double prof_get_measurement_wallclock(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns the wallclock of the last measurement in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the wallclock time of the last measurement
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* RESULT
* double - the wallclock time
*
* NOTES
* MT-NOTE: prof_get_measurement_wallclock() is MT safe
*******************************************************************************/
double prof_get_measurement_wallclock(prof_level level, bool with_sub, dstring *error)
{
pthread_t thread_id;
int thread_num;
clock_t clock = 0;
double ret = 0.0;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_measurement_wallclock", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_measurement_wallclock");
} else {
clock = theInfo[thread_num][level].end - theInfo[thread_num][level].start;
if (!with_sub) {
clock -= theInfo[thread_num][level].sub;
}
}
ret = clock * 1.0 / sysconf(_SC_CLK_TCK);
}
return ret;
}
/****** uti/profiling/prof_get_measurement_utime() ************************
* NAME
* prof_get_measurement_utime() -- return user cpu time of measurement
*
* SYNOPSIS
* double prof_get_measurement_utime(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns the user cpu time of the last measurement in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the user cpu time of the last measurement
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_measurement_utime() is not MT safe
*******************************************************************************/
double prof_get_measurement_utime(prof_level level, bool with_sub, dstring *error)
{
pthread_t thread_id;
int thread_num;
clock_t clock = 0;
double ret = 0.0;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_measurement_utime", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_measurement_utime");
} else {
clock = (theInfo[thread_num][level].tms_end.tms_utime - theInfo[thread_num][level].tms_start.tms_utime);
if (!with_sub) {
clock -= theInfo[thread_num][level].sub_utime;
}
}
ret = clock * 1.0 / sysconf(_SC_CLK_TCK);
}
return ret;
}
/****** uti/profiling/prof_get_measurement_stime() ************************
* NAME
* prof_get_measurement_stime() -- return system cpu time of measurement
*
* SYNOPSIS
* double prof_get_measurement_stime(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns the system cpu time of the last measurement in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the system cpu time of the last measurement
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_measurement_stime() is MT safe
*******************************************************************************/
double prof_get_measurement_stime(prof_level level, bool with_sub, dstring *error)
{
pthread_t thread_id;
int thread_num;
clock_t clock = 0;
double ret = 0.0;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_measurement_stime", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_measurement_stime");
} else {
clock = (theInfo[thread_num][level].tms_end.tms_stime - theInfo[thread_num][level].tms_start.tms_stime);
if (!with_sub) {
clock -= theInfo[thread_num][level].sub_stime;
}
}
ret = clock * 1.0 / sysconf(_SC_CLK_TCK);
}
return ret;
}
/****** uti/profiling/prof_get_total_wallclock() **************************
* NAME
* prof_get_total_wallclock() -- get total wallclock time
*
* SYNOPSIS
* double prof_get_total_wallclock(dstring *error)
*
* FUNCTION
* Returns the wallclock time since profiling was enabled in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the total wallclock time of the profiling run
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_total_wallclock() is MT safe
*******************************************************************************/
double prof_get_total_wallclock(prof_level level, dstring *error)
{
pthread_t thread_id;
int thread_num;
double ret = 0.0;
if (level >= SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_total_wallclock", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_total_wallclock");
} else if (!theInfo[thread_num][level].prof_is_started) {
sge_dstring_sprintf_append(error, MSG_PROF_NOTACTIVE_S, "prof_get_total_wallclock");
} else {
struct tms tms_buffer;
clock_t now;
now = times(&tms_buffer);
ret = (now - theInfo[thread_num][level].start_clock) * 1.0 / sysconf(_SC_CLK_TCK);
}
}
return ret;
}
/****** uti/profiling/prof_get_total_busy() *******************************
* NAME
* prof_get_total_busy() -- return total busy time
*
* SYNOPSIS
* double prof_get_total_busy(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns the total busy time since profiling was enabled in seconds.
* Busy time is the time between starting and stopping a measurement.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the total busy time of the profiling run
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_total_busy() is MT safe
*******************************************************************************/
static double _prof_get_total_busy(prof_level level, bool with_sub, dstring *error _UNUSED)
{
int thread_num;
clock_t clock = 0;
thread_num = get_prof_info_thread_id();
clock = theInfo[thread_num][level].total;
if (!with_sub) {
clock -= theInfo[thread_num][level].sub_total;
}
return clock * 1.0 / sysconf(_SC_CLK_TCK);
}
double prof_get_total_busy(prof_level level, bool with_sub, dstring *error)
{
double ret = 0.0;
pthread_t thread_id;
int thread_num;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_total_busy", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_total_busy");
} else if (level == SGE_PROF_ALL) {
prof_level i;
for (i = SGE_PROF_OTHER; i < SGE_PROF_ALL; i++) {
ret += _prof_get_total_busy(i, with_sub, error);
}
} else {
ret = _prof_get_total_busy(level, with_sub, error);
}
}
return ret;
}
/****** uti/profiling/prof_get_total_utime() ******************************
* NAME
* prof_get_total_utime() -- get total user cpu time
*
* SYNOPSIS
* double prof_get_total_utime(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns the user cpu time since profiling was enabled in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the total user cpu time of the profiling run
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_total_utime() is MT safe
*******************************************************************************/
static double _prof_get_total_utime(prof_level level, bool with_sub, dstring *error _UNUSED)
{
int thread_num;
clock_t clock = 0;
thread_num = get_prof_info_thread_id();
clock = theInfo[thread_num][level].total_utime;
if (!with_sub) {
clock -= theInfo[thread_num][level].sub_total_utime;
}
return clock * 1.0 / sysconf(_SC_CLK_TCK);
}
double prof_get_total_utime(prof_level level, bool with_sub, dstring *error)
{
double ret = 0.0;
pthread_t thread_id;
int thread_num;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_total_utime", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_total_utime");
} else if (level == SGE_PROF_ALL) {
prof_level i;
for (i = SGE_PROF_OTHER; i < SGE_PROF_ALL; i++) {
ret += _prof_get_total_utime(i, with_sub, error);
}
} else {
ret = _prof_get_total_utime(level, with_sub, error);
}
}
return ret;
}
/****** uti/profiling/prof_get_total_stime() ******************************
* NAME
* prof_get_total_stime() -- get total system cpu time
*
* SYNOPSIS
* double prof_get_total_stime(prof_level level, bool with_sub)
*
* FUNCTION
* Returns the total system cpu time since profiling was enabled in seconds.
* Resolution is clock ticks (_SC_CLK_TCK).
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* double - the total system cpu time of the profiling run
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* NOTES
* MT-NOTE: prof_get_total_stime() is MT safe
*******************************************************************************/
static double _prof_get_total_stime(prof_level level, bool with_sub, dstring *error _UNUSED)
{
int thread_num;
clock_t clock = 0;
thread_num = get_prof_info_thread_id();
clock = theInfo[thread_num][level].total_stime;
if (!with_sub) {
clock -= theInfo[thread_num][level].sub_total_stime;
}
return clock * 1.0 / sysconf(_SC_CLK_TCK);
}
double prof_get_total_stime(prof_level level, bool with_sub, dstring *error)
{
double ret = 0.0;
pthread_t thread_id;
int thread_num;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_total_stime", level);
} else if (profiling_enabled) {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_total_stime");
} else if (level == SGE_PROF_ALL) {
prof_level i;
for (i = SGE_PROF_OTHER; i < SGE_PROF_ALL; i++) {
ret += _prof_get_total_stime(i, with_sub, error);
}
} else {
ret = _prof_get_total_stime(level, with_sub, error);
}
}
return ret;
}
/****** uti/profiling/prof_get_info_string() ******************************
* NAME
* prof_get_info_string() -- get informational message
*
* SYNOPSIS
* const char* prof_get_info_string(prof_level level, bool with_sub, dstring *error)
*
* FUNCTION
* Returns a string containing the most interesting data, both for the
* last measurement and for the total runtime:
* - wallclock of measurement
* - user cpu time of measurement
* - system cpu time of measurement
* - total wallclock time (runtime)
* - total busys time
* - utilization (busy time / wallclock time) * 100 %
*
* INPUTS
* prof_level level - level to process
* bool with_sub - include usage of subordinated measurements?
* dstring *error - if != NULL, error messages will be put here
*
* RESULT
* const char* - pointer to result string. It is valid until the next
* call of prof_get_info_string()
* on error, 0 is returned and an error message is written
* to the buffer given in parameter error, if error != NULL
*
* EXAMPLE
* The result can look like the following:
* "wc = 0.190s, utime = 0.120s, stime = 0.000s, runtime 9515s, busy 105s,
* utilization 1%"
*
* NOTES
* MT-NOTE: prof_get_info_string() is MT safe
*******************************************************************************/
#define PROF_GET_INFO_FORMAT "%-15.15s: wc = %10.3fs, utime = %10.3fs, stime = %10.3fs, utilization = %3.0f%%\n"
static const char *
_prof_get_info_string(prof_level level, dstring *info_string, bool with_sub, dstring *error)
{
int thread_num;
dstring level_string = DSTRING_INIT;
const char *ret = NULL;
double busy, utime, stime, utilization;
thread_num = get_prof_info_thread_id();
busy = prof_get_total_busy(level, with_sub, error);
utime = prof_get_total_utime(level, with_sub, error);
stime = prof_get_total_stime(level, with_sub, error);
utilization = busy > 0 ? (utime + stime) / busy * 100 : 0;
sge_dstring_sprintf(&level_string, PROF_GET_INFO_FORMAT,
theInfo[thread_num][level].name, busy, utime, stime, utilization);
ret = sge_dstring_append_dstring(info_string, &level_string);
sge_dstring_free(&level_string);
return ret;
}
const char *
prof_get_info_string(prof_level level, bool with_sub, dstring *error)
{
pthread_t thread_id;
int thread_num;
const char *ret = NULL;
if (level > SGE_PROF_ALL) {
sge_dstring_sprintf_append(error, MSG_PROF_INVALIDLEVEL_SD, "prof_get_info_string", level);
ret = sge_dstring_get_string (error);
} else if (!profiling_enabled) {
ret = "Profiling disabled";
} else {
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num < 0) || (thread_num >= MAX_THREAD_NUM)) {
sge_dstring_sprintf_append(error, MSG_PROF_MAXTHREADSEXCEEDED_S, "prof_get_info_string");
/* total summary: one line for each level, one line for summary */
} else if (level == SGE_PROF_ALL) {
double busy, utime, stime, utilization;
prof_level i;
dstring total_string = DSTRING_INIT;
for (i = SGE_PROF_OTHER; i <= SGE_PROF_ALL; i++) {
/* clear previous contents */
sge_dstring_clear(&(theInfo[thread_num][i].info_string));
}
prof_stop_measurement(SGE_PROF_OTHER, error);
busy = prof_get_total_busy(SGE_PROF_ALL, with_sub, error);
utime = prof_get_total_utime(SGE_PROF_ALL, with_sub, error);
stime = prof_get_total_stime(SGE_PROF_ALL, with_sub, error);
utilization = busy > 0 ? (utime + stime) / busy * 100 : 0;
for (i = SGE_PROF_OTHER; i < SGE_PROF_ALL; i++) {
if (theInfo[thread_num][i].name != NULL && theInfo[thread_num][i].ever_started == true) {
_prof_get_info_string(i, &theInfo[thread_num][SGE_PROF_ALL].info_string, with_sub, error);
}
}
prof_start_measurement(SGE_PROF_OTHER, error);
sge_dstring_sprintf(&total_string, PROF_GET_INFO_FORMAT,
"total", busy, utime, stime, utilization);
ret = sge_dstring_append_dstring(&theInfo[thread_num][SGE_PROF_ALL].info_string, &total_string);
sge_dstring_free(&total_string);
} else {
/* clear previous contents */
sge_dstring_clear(&(theInfo[thread_num][level].info_string));
if (theInfo[thread_num][level].name != NULL) {
ret = _prof_get_info_string(level, &theInfo[thread_num][level].info_string, with_sub, error);
}
}
}
return ret;
}
bool prof_output_info(prof_level level, bool with_sub, const char *info)
{
bool ret = false;
DENTER(TOP_LAYER, "prof_output_info");
if (profiling_enabled && (level <= SGE_PROF_ALL)) {
int thread_num;
pthread_t thread_id;
thread_id = pthread_self();
thread_num = get_prof_info_thread_id();
if ((thread_num >= 0) && (thread_num < MAX_THREAD_NUM) && prof_is_active(level)) {
const char *info_message = NULL;
const char *message = NULL;
struct saved_vars_s *context = NULL;
info_message = prof_get_info_string(level, with_sub, NULL);
PROFILING((SGE_EVENT, "PROF(%d): %s%s", (int)thread_id, info, ""));
for (message = sge_strtok_r(info_message, "\n", &context); message != NULL;
message = sge_strtok_r(NULL, "\n", &context)) {
PROFILING((SGE_EVENT, "PROF(%d): %s", (int)thread_id, message));
}
prof_reset(level, NULL);
sge_free_saved_vars(context);
ret = true;
}
}
DRETURN(ret);
}
/****** uti/profiling/prof_info_init() ******************************
* NAME
* prof_info_init() -- initialize the sge_prof_info_t struc array with default values
*
* SYNOPSIS
* static void prof_info_init(prof_level level)
*
* FUNCTION
* initialize the sge_prof_info_t struct array with default values
*
* INPUTS
* prof_level level
*
* RESULT
* initialized sge_prof_info_t array for the given profiling level
*
* EXAMPLE
*
* NOTES
* MT-NOTE: prof_info_init() is MT safe
*******************************************************************************/
static void prof_info_init(prof_level level)
{
int thread_num;
thread_num = get_prof_info_thread_id();
if (level <= SGE_PROF_ALL) {
if (level == SGE_PROF_ALL) {
prof_level i;
for (i = SGE_PROF_OTHER; i <= SGE_PROF_ALL; i++) {
prof_info_level_init (i, thread_num);
}
} else {
prof_info_level_init (SGE_PROF_OTHER, thread_num);
}
theInfo[thread_num][SGE_PROF_ALL].akt_level = SGE_PROF_NONE;
}
return;
}
/****** uti/profiling/prof_info_level_init() ******************************
* NAME
* prof_info_level_init() -- initialize the sge_prof_info_t struc array with default values
*
* SYNOPSIS
* static void prof_info_level_init(prof_level level)
*
* FUNCTION
* initialize the sge_prof_info_t struct array with default values
*
* INPUTS
* prof_level i
* int thread_num
*
* RESULT
* initialized sge_prof_info_t array for the given profiling level
*
* EXAMPLE
*
* NOTES
* MT-NOTE: prof_info_level_init() is MT safe
*******************************************************************************/
static void prof_info_level_init(prof_level i, int thread_num)
{
switch (i) {
case SGE_PROF_OTHER:
theInfo[thread_num][i].name = "other";
break;
case SGE_PROF_COMMUNICATION:
theInfo[thread_num][i].name = "communication";
break;
case SGE_PROF_PACKING:
theInfo[thread_num][i].name = "packing";
break;
case SGE_PROF_EVENTCLIENT:
theInfo[thread_num][i].name = "eventclient";
break;
case SGE_PROF_EVENTMASTER:
theInfo[thread_num][i].name = "eventmaster";
break;
case SGE_PROF_MIRROR:
theInfo[thread_num][i].name = "mirror";
break;
case SGE_PROF_SPOOLING:
theInfo[thread_num][i].name = "spooling";
break;
case SGE_PROF_SPOOLINGIO:
theInfo[thread_num][i].name = "spooling-io";
break;
case SGE_PROF_JOBSCRIPT:
theInfo[thread_num][i].name = "spooling-script";
break;
case SGE_PROF_GDI:
theInfo[thread_num][i].name = "gdi";
break;
case SGE_PROF_GDI_REQUEST:
theInfo[thread_num][i].name = "gdi_request";
break;
case SGE_PROF_HT_RESIZE:
theInfo[thread_num][i].name = "ht-resize";
break;
case SGE_PROF_ALL:
theInfo[thread_num][i].name = "all";
break;
default:
theInfo[thread_num][i].name = NULL; /* "custom"*/
break;
}
theInfo[thread_num][i].nested_calls = 0;
theInfo[thread_num][i].start = 0;
theInfo[thread_num][i].end = 0;
theInfo[thread_num][i].tms_start.tms_utime = 0;
theInfo[thread_num][i].tms_start.tms_stime = 0;
theInfo[thread_num][i].tms_start.tms_cutime = 0;
theInfo[thread_num][i].tms_start.tms_cstime = 0;
theInfo[thread_num][i].tms_end.tms_utime = 0;
theInfo[thread_num][i].tms_end.tms_stime = 0;
theInfo[thread_num][i].tms_end.tms_cutime = 0;
theInfo[thread_num][i].tms_end.tms_cstime = 0;
theInfo[thread_num][i].total = 0;
theInfo[thread_num][i].total_utime = 0;
theInfo[thread_num][i].total_stime = 0;
theInfo[thread_num][i].pre = SGE_PROF_NONE;
theInfo[thread_num][i].sub = 0;
theInfo[thread_num][i].sub_utime = 0;
theInfo[thread_num][i].sub_stime = 0;
theInfo[thread_num][i].sub_total = 0;
theInfo[thread_num][i].sub_total_utime = 0;
theInfo[thread_num][i].sub_total_stime = 0;
theInfo[thread_num][i].prof_is_started = false;
theInfo[thread_num][i].start_clock = 0;
theInfo[thread_num][i].ever_started = false;
theInfo[thread_num][i].info_string.s = NULL;
theInfo[thread_num][i].info_string.length = 0;
theInfo[thread_num][i].info_string.size = 0;
theInfo[thread_num][i].info_string.is_static = false;
}
/****** uti/profiling/init_array() ******************************
* NAME
* init_array() -- mallocs memory for the sge_prof_info_t array
*
* SYNOPSIS
* void init_array(pthread_t num)
*
* FUNCTION
* mallocs memory for sge_prof_info_t array for the number
* of MAX_THREAD_NUM threads
* mallocs memory for each thread if nedded
*
* INPUTS
* the thread id, which needs malloced memory
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: init_array() is MT safe
*******************************************************************************/
static void init_array(pthread_t num) {
int i, c;
DENTER(CULL_LAYER,"init_array");
if (sge_prof_array_initialized == 0) {
CRITICAL((SGE_EVENT,"Profiling array is not initialized!"));
abort();
}
pthread_mutex_lock(&thrdInfo_mutex);
for (i = 0; i < MAX_THREAD_NUM; i++) {
if (theInfo[i] != NULL && theInfo[i][SGE_PROF_ALL].thread_id == num) {
break;
} else if (theInfo[i] == NULL) {
long storage = 0;
theInfo[i] = (sge_prof_info_t*)sge_malloc((SGE_PROF_ALL + 1) * sizeof(sge_prof_info_t));
memset (theInfo[i], 0, (SGE_PROF_ALL + 1) * sizeof(sge_prof_info_t));
for (c = 0; c <= SGE_PROF_ALL; c++) {
theInfo[i][c].thread_id = num;
}
/* Rather than malloc'ing space for an int that we'll have to clean up
* later, we just store the int directly in the thread local storage.
* In order to keep Linux from complaining, the value we store must be
* the same size as a void pointer, i.e. a long. */
storage = i;
pthread_setspecific (thread_id_key, (void *)storage);
prof_info_init(SGE_PROF_ALL);
break;
}
}
pthread_mutex_unlock(&thrdInfo_mutex);
DEXIT;
return;
}
/* per process initialization */
static void init_array_first(void) {
if (sge_prof_array_initialized == 0) {
pthread_mutex_lock(&thrdInfo_mutex);
if (pthread_key_create (&thread_id_key, NULL) == 0) {
theInfo = (sge_prof_info_t**)sge_malloc(MAX_THREAD_NUM * sizeof(sge_prof_info_t*));
memset (theInfo, 0, MAX_THREAD_NUM * sizeof(sge_prof_info_t*));
sge_prof_array_initialized = 1;
}
pthread_mutex_unlock(&thrdInfo_mutex);
}
}
/****** uti/profiling/init_thread_info() ******************************
* NAME
* init_thread_info() -- mallocs memory for the thread_info_t array
*
* SYNOPSIS
* void init_thread_info(void)
*
* FUNCTION
* mallocs memory for thread_info_t array (thread name/id mapping)
* for the number of MAX_THREAD_NUM threads. Must be called once
* per process.
*
* INPUTS
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: init_thread_info() is MT safe
*******************************************************************************/
static void init_thread_info(void) {
if (!profiling_enabled) {
return;
}
pthread_mutex_lock(&thrdInfo_mutex);
if (thrdInfo == NULL) {
thrdInfo = (sge_thread_info_t*)sge_malloc(MAX_THREAD_NUM * sizeof(sge_thread_info_t));
memset(thrdInfo, 0, MAX_THREAD_NUM * sizeof(sge_thread_info_t));
}
pthread_mutex_unlock(&thrdInfo_mutex);
}
/****** uti/profiling/get_thread_info() ******************************
* NAME
* get_thread_info() -- get the thread name/id mapping array
*
* SYNOPSIS
* thread_info_t* get_thread_info(void)
*
* FUNCTION
* if the thread name/id mapping array is not initialized
* it will be done
*
* INPUTS
*
* RESULT
* returns a pointer to the thread name/id mapping array
*
* EXAMPLE
*
* NOTES
* MT-NOTE: get_thread_info() is MT safe
*******************************************************************************/
sge_thread_info_t* get_thread_info(void) {
if (!profiling_enabled) {
return NULL;
}
init_thread_info();
return thrdInfo;
}
/****** uti/profiling/set_thread_name() ******************************
* NAME
* set_thread_name() -- set the thread name mapped to its id
*
* SYNOPSIS
* void set_thread_name(pthread_t thread_id, const char* thread_name)
*
* FUNCTION
* maps the name and the id of a thread
* set the thread profiling status to false
*
* INPUTS
* pthread_t thread_id
* const char* thread_name
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: set_thread_info() is MT safe
*******************************************************************************/
void set_thread_name(pthread_t thread_id, const char* thread_name) {
int thread_num;
if (!profiling_enabled) {
return;
}
init_thread_info();
init_array(thread_id);
thread_num = get_prof_info_thread_id();
if ((thread_num >= 0) && (thread_num < MAX_THREAD_NUM)) {
pthread_mutex_lock(&thrdInfo_mutex);
thrdInfo[thread_num].thrd_id = thread_id;
thrdInfo[thread_num].thrd_name = thread_name;
thrdInfo[thread_num].prof_is_active = false;
thrdInfo[thread_num].is_terminated = 0;
pthread_mutex_unlock(&thrdInfo_mutex);
}
}
/****** uti/profiling/set_thread_prof_status_by_id() ******************************
* NAME
* set_thread_prof_status_by_id() -- sets the profiling status for the thread
* with the given thread id
*
* SYNOPSIS
* void set_thread_prof_status_by_id(pthread_t thread_id, bool prof_status)
*
* FUNCTION
* set the thread profiling status of the thread with the given id
*
* INPUTS
* pthread_t thread_id
* bool prof_status
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: set_thread_prof_status_by_id() is MT safe
*******************************************************************************/
void set_thread_prof_status_by_id(pthread_t thread_id, bool prof_status) {
int thread_num;
if (!profiling_enabled) {
return;
}
init_thread_info();
thread_num = get_prof_info_thread_id();
pthread_mutex_lock(&thrdInfo_mutex);
if (thrdInfo[thread_num].thrd_id == thread_id) {
thrdInfo[thread_num].prof_is_active = prof_status;
}
pthread_mutex_unlock(&thrdInfo_mutex);
}
/****** uti/profiling/set_thread_prof_status_by_name() ******************************
* NAME
* set_thread_prof_status_by_name() -- sets the profiling status for the thread
* with the given thread id and thread name
*
* SYNOPSIS
* void set_thread_prof_status_by_name(pthread_t thread_id, const char* thread_name, bool prof_status)
*
* FUNCTION
* set the thread profiling status of the thread with the given id and name
*
* INPUTS
* pthread_t thread_id
* const char* thread_name
* bool prof_status
*
* RESULTS
* return 0 - ok
* return 1 - thread_name = NULL
*
* EXAMPLE
*
* NOTES
* MT-NOTE: set_thread_prof_status_by_name() is MT safe
*******************************************************************************/
int set_thread_prof_status_by_name(const char* thread_name, bool prof_status) {
int i;
if (!profiling_enabled) {
return 0;
} else if (thread_name == NULL) {
return 1;
}
init_thread_info();
pthread_mutex_lock(&thrdInfo_mutex);
for (i = 0; i < MAX_THREAD_NUM; i++) {
if (thrdInfo[i].thrd_name != NULL) {
if (strcmp(thrdInfo[i].thrd_name, thread_name) == 0) {
thrdInfo[i].prof_is_active = prof_status;
}
}
}
pthread_mutex_unlock(&thrdInfo_mutex);
return 0;
}
/****** uti/profiling/sge_prof_cleanup() ******************************
* NAME
* sge_prof_cleanup() -- frees the profiling array
*
* SYNOPSIS
* void sge_prof_cleanup(void)
*
* FUNCTION
* frees the profiling array
*
* INPUTS
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: sge_prof_cleanup() is MT safe
*******************************************************************************/
void sge_prof_cleanup(void) {
if (!profiling_enabled) {
return;
}
pthread_mutex_lock(&thrdInfo_mutex);
pthread_key_delete (thread_id_key);
if (theInfo != NULL) {
int c, i;
for (c = 0; c < MAX_THREAD_NUM; c++) {
for (i = 0; i <= SGE_PROF_ALL; i++) {
if ( theInfo[c] != NULL) {
sge_dstring_free(&theInfo[c][i].info_string);
}
}
sge_free(&(theInfo[c]));
}
sge_free(&theInfo);
}
sge_free(&thrdInfo);
sge_prof_array_initialized = 0;
pthread_mutex_unlock(&thrdInfo_mutex);
}
/****** uti/profiling/thread_prof_active_by_id() ******************************
* NAME
* thread_prof_active_by_id() -- returns the status of a thread
*
* SYNOPSIS
* bool thread_prof_active_by_id(void)
*
* FUNCTION
* returns the profiling status of the thread
*
* INPUTS
* void
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: thread_prof_active_by_id() is MT safe
*******************************************************************************/
bool thread_prof_active_by_id(void) {
int thread_num;
bool ret = false;
if (profiling_enabled) {
init_thread_info();
thread_num = get_prof_info_thread_id();
if ((thread_num >= 0) && (thread_num < MAX_THREAD_NUM)) {
pthread_mutex_lock(&thrdInfo_mutex);
ret = thrdInfo[thread_num].prof_is_active;
pthread_mutex_unlock(&thrdInfo_mutex);
}
}
return ret;
}
/****** uti/profiling/thread_prof_active_by_name() ******************************
* NAME
* thread_prof_active_by_name() -- returns the status of a thread
*
* SYNOPSIS
* bool thread_prof_active_by_name(pthread_t thread_id)
*
* FUNCTION
* returns the profiling status of a thread
*
* INPUTS
* pthread_t thread_id
*
* RESULT
*
* EXAMPLE
*
* NOTES
* MT-NOTE: thread_prof_active_by_name() is MT safe
*******************************************************************************/
bool thread_prof_active_by_name(const char* thread_name) {
bool ret = false;
if (profiling_enabled && (thread_name != NULL)) {
int c;
init_thread_info();
pthread_mutex_lock(&thrdInfo_mutex);
for (c = 0; c < MAX_THREAD_NUM; c++) {
if (thrdInfo[c].thrd_name != NULL && strstr(thrdInfo[c].thrd_name, thread_name)) {
ret = thrdInfo[c].prof_is_active;
break;
}
}
pthread_mutex_unlock(&thrdInfo_mutex);
}
return ret;
}
static int get_prof_info_thread_id(void) {
long ret = (long)pthread_getspecific (thread_id_key);
return ret;
}
/****** uti/profiling/thread_start_stop_profiling() ****************************
* NAME
* thread_start_stop_profiling() -- start profiling for thread
*
* SYNOPSIS
* void
* thread_start_stop_profiling(void)
*
* FUNCTION
* Checks if profiling has been enabled for the current thread.
* If yes, starts profiling for all levels.
* If profiling has been disabled for the current thread, profiling
* is disabled for all levels.
*
* NOTES
* MT-NOTE: thread_start_stop_profiling() is MT safe
*******************************************************************************/
void
thread_start_stop_profiling(void)
{
if (!profiling_enabled) {
return;
}
if (thread_prof_active_by_id()) {
prof_start(SGE_PROF_ALL, NULL);
} else {
prof_stop(SGE_PROF_ALL, NULL);
}
}
/****** uti/profiling/thread_output_profiling() ********************************
* NAME
* thread_output_profiling() -- output profiling info for thread
*
* SYNOPSIS
* void
* thread_output_profiling(const char *title, time_t *next_prof_output)
*
* FUNCTION
* Outputs profiling information for the current thread.
* Information for all active profiling levels is dumped.
* The first line dumped is a sort of title, that can/should be used
* to identify the thread.
*
* The variable next_prof_output will be set by this function.
* This variable should be initialized to 0 (zero) by the caller before the
* first call of this function.
*
* INPUTS
* const char *title - title to print as first line
* time_t *next_prof_output - time of next profiling output
*
* NOTES
* MT-NOTE: thread_output_profiling() is MT safe
*
* SEE ALSO
* uti/profiling/prof_output_info()
*******************************************************************************/
void
thread_output_profiling(const char *title, time_t *next_prof_output)
{
if (prof_is_active(SGE_PROF_ALL)) {
time_t now = (time_t)sge_get_gmt();
if (*next_prof_output == 0) {
unsigned int seed = (unsigned int)pthread_self();
*next_prof_output = now + (rand_r(&seed) % 20);
} else {
if (now >= *next_prof_output) {
prof_output_info(SGE_PROF_ALL, false, title);
*next_prof_output = now + 60;
}
}
}
}
|