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
|
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta http-equiv="X-UA-Compatible" content="IE=9"/>
<meta name="generator" content="Doxygen 1.8.6"/>
<title>ViennaCL - The Vienna Computing Library: viennacl/linalg/cuda/spgemm_rmerge.hpp Source File</title>
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<link href="navtree.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="resize.js"></script>
<script type="text/javascript" src="navtree.js"></script>
<script type="text/javascript">
$(document).ready(initResizable);
$(window).load(resizeHeight);
</script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/search.js"></script>
<script type="text/javascript">
$(document).ready(function() { searchBox.OnSelectItem(0); });
</script>
<link href="doxygen.css" rel="stylesheet" type="text/css" />
</head>
<body>
<div id="top"><!-- do not remove this div, it is closed by doxygen! -->
<div id="titlearea">
<table cellspacing="0" cellpadding="0">
<tbody>
<tr style="height: 56px;">
<td style="padding-left: 0.5em;">
<div id="projectname">ViennaCL - The Vienna Computing Library
 <span id="projectnumber">1.7.1</span>
</div>
<div id="projectbrief">Free open-source GPU-accelerated linear algebra and solver library.</div>
</td>
<td> <div id="MSearchBox" class="MSearchBoxInactive">
<span class="left">
<img id="MSearchSelect" src="search/mag_sel.png"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
alt=""/>
<input type="text" id="MSearchField" value="Search" accesskey="S"
onfocus="searchBox.OnSearchFieldFocus(true)"
onblur="searchBox.OnSearchFieldFocus(false)"
onkeyup="searchBox.OnSearchFieldChange(event)"/>
</span><span class="right">
<a id="MSearchClose" href="javascript:searchBox.CloseResultsWindow()"><img id="MSearchCloseImg" border="0" src="search/close.png" alt=""/></a>
</span>
</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
<!-- Generated by Doxygen 1.8.6 -->
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
</div><!-- top -->
<div id="side-nav" class="ui-resizable side-nav-resizable">
<div id="nav-tree">
<div id="nav-tree-contents">
<div id="nav-sync" class="sync"></div>
</div>
</div>
<div id="splitbar" style="-moz-user-select:none;"
class="ui-resizable-handle">
</div>
</div>
<script type="text/javascript">
$(document).ready(function(){initNavTree('spgemm__rmerge_8hpp_source.html','');});
</script>
<div id="doc-content">
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
onkeydown="return searchBox.OnSearchSelectKey(event)">
<a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(0)"><span class="SelectionMark"> </span>All</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(1)"><span class="SelectionMark"> </span>Classes</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(2)"><span class="SelectionMark"> </span>Namespaces</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(3)"><span class="SelectionMark"> </span>Files</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(4)"><span class="SelectionMark"> </span>Functions</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(5)"><span class="SelectionMark"> </span>Variables</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(6)"><span class="SelectionMark"> </span>Typedefs</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(7)"><span class="SelectionMark"> </span>Enumerations</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(8)"><span class="SelectionMark"> </span>Enumerator</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(9)"><span class="SelectionMark"> </span>Friends</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(10)"><span class="SelectionMark"> </span>Macros</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(11)"><span class="SelectionMark"> </span>Pages</a></div>
<!-- iframe showing the search results (closed by default) -->
<div id="MSearchResultsWindow">
<iframe src="javascript:void(0)" frameborder="0"
name="MSearchResults" id="MSearchResults">
</iframe>
</div>
<div class="header">
<div class="headertitle">
<div class="title">spgemm_rmerge.hpp</div> </div>
</div><!--header-->
<div class="contents">
<a href="spgemm__rmerge_8hpp.html">Go to the documentation of this file.</a><div class="fragment"><div class="line"><a name="l00001"></a><span class="lineno"> 1</span> <span class="preprocessor">#ifndef VIENNACL_LINALG_CUDA_SPGEMM_RMERGE_HPP_</span></div>
<div class="line"><a name="l00002"></a><span class="lineno"> 2</span> <span class="preprocessor">#define VIENNACL_LINALG_CUDA_SPGEMM_RMERGE_HPP_</span></div>
<div class="line"><a name="l00003"></a><span class="lineno"> 3</span> </div>
<div class="line"><a name="l00004"></a><span class="lineno"> 4</span> <span class="comment">/* =========================================================================</span></div>
<div class="line"><a name="l00005"></a><span class="lineno"> 5</span> <span class="comment"> Copyright (c) 2010-2016, Institute for Microelectronics,</span></div>
<div class="line"><a name="l00006"></a><span class="lineno"> 6</span> <span class="comment"> Institute for Analysis and Scientific Computing,</span></div>
<div class="line"><a name="l00007"></a><span class="lineno"> 7</span> <span class="comment"> TU Wien.</span></div>
<div class="line"><a name="l00008"></a><span class="lineno"> 8</span> <span class="comment"> Portions of this software are copyright by UChicago Argonne, LLC.</span></div>
<div class="line"><a name="l00009"></a><span class="lineno"> 9</span> <span class="comment"></span></div>
<div class="line"><a name="l00010"></a><span class="lineno"> 10</span> <span class="comment"> -----------------</span></div>
<div class="line"><a name="l00011"></a><span class="lineno"> 11</span> <span class="comment"> ViennaCL - The Vienna Computing Library</span></div>
<div class="line"><a name="l00012"></a><span class="lineno"> 12</span> <span class="comment"> -----------------</span></div>
<div class="line"><a name="l00013"></a><span class="lineno"> 13</span> <span class="comment"></span></div>
<div class="line"><a name="l00014"></a><span class="lineno"> 14</span> <span class="comment"> Project Head: Karl Rupp rupp@iue.tuwien.ac.at</span></div>
<div class="line"><a name="l00015"></a><span class="lineno"> 15</span> <span class="comment"></span></div>
<div class="line"><a name="l00016"></a><span class="lineno"> 16</span> <span class="comment"> (A list of authors and contributors can be found in the manual)</span></div>
<div class="line"><a name="l00017"></a><span class="lineno"> 17</span> <span class="comment"></span></div>
<div class="line"><a name="l00018"></a><span class="lineno"> 18</span> <span class="comment"> License: MIT (X11), see file LICENSE in the base directory</span></div>
<div class="line"><a name="l00019"></a><span class="lineno"> 19</span> <span class="comment">============================================================================= */</span></div>
<div class="line"><a name="l00020"></a><span class="lineno"> 20</span> </div>
<div class="line"><a name="l00025"></a><span class="lineno"> 25</span> <span class="preprocessor">#include <stdexcept></span></div>
<div class="line"><a name="l00026"></a><span class="lineno"> 26</span> </div>
<div class="line"><a name="l00027"></a><span class="lineno"> 27</span> <span class="preprocessor">#include "<a class="code" href="forwards_8h.html">viennacl/forwards.h</a>"</span></div>
<div class="line"><a name="l00028"></a><span class="lineno"> 28</span> <span class="preprocessor">#include "<a class="code" href="scalar_8hpp.html">viennacl/scalar.hpp</a>"</span></div>
<div class="line"><a name="l00029"></a><span class="lineno"> 29</span> <span class="preprocessor">#include "<a class="code" href="vector_8hpp.html">viennacl/vector.hpp</a>"</span></div>
<div class="line"><a name="l00030"></a><span class="lineno"> 30</span> <span class="preprocessor">#include "<a class="code" href="tools_8hpp.html">viennacl/tools/tools.hpp</a>"</span></div>
<div class="line"><a name="l00031"></a><span class="lineno"> 31</span> <span class="preprocessor">#include "<a class="code" href="linalg_2cuda_2common_8hpp.html">viennacl/linalg/cuda/common.hpp</a>"</span></div>
<div class="line"><a name="l00032"></a><span class="lineno"> 32</span> </div>
<div class="line"><a name="l00033"></a><span class="lineno"> 33</span> <span class="preprocessor">#include "<a class="code" href="timer_8hpp.html">viennacl/tools/timer.hpp</a>"</span></div>
<div class="line"><a name="l00034"></a><span class="lineno"> 34</span> </div>
<div class="line"><a name="l00035"></a><span class="lineno"> 35</span> <span class="preprocessor">#include "<a class="code" href="sparse__matrix__operations__solve_8hpp.html">viennacl/linalg/cuda/sparse_matrix_operations_solve.hpp</a>"</span></div>
<div class="line"><a name="l00036"></a><span class="lineno"> 36</span> </div>
<div class="line"><a name="l00037"></a><span class="lineno"> 37</span> <span class="keyword">namespace </span>viennacl</div>
<div class="line"><a name="l00038"></a><span class="lineno"> 38</span> {</div>
<div class="line"><a name="l00039"></a><span class="lineno"> 39</span> <span class="keyword">namespace </span>linalg</div>
<div class="line"><a name="l00040"></a><span class="lineno"> 40</span> {</div>
<div class="line"><a name="l00041"></a><span class="lineno"> 41</span> <span class="keyword">namespace </span>cuda</div>
<div class="line"><a name="l00042"></a><span class="lineno"> 42</span> {</div>
<div class="line"><a name="l00043"></a><span class="lineno"> 43</span> </div>
<div class="line"><a name="l00045"></a><span class="lineno"> 45</span> <span class="keyword">template</span><<span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00046"></a><span class="lineno"> 46</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> load_and_cache(<span class="keyword">const</span> <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *address)</div>
<div class="line"><a name="l00047"></a><span class="lineno"> 47</span> {</div>
<div class="line"><a name="l00048"></a><span class="lineno"> 48</span> <span class="preprocessor">#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 350</span></div>
<div class="line"><a name="l00049"></a><span class="lineno"> 49</span>  <span class="keywordflow">return</span> __ldg(address);</div>
<div class="line"><a name="l00050"></a><span class="lineno"> 50</span> <span class="preprocessor">#else</span></div>
<div class="line"><a name="l00051"></a><span class="lineno"> 51</span>  <span class="keywordflow">return</span> *address;</div>
<div class="line"><a name="l00052"></a><span class="lineno"> 52</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00053"></a><span class="lineno"> 53</span> }</div>
<div class="line"><a name="l00054"></a><span class="lineno"> 54</span> </div>
<div class="line"><a name="l00055"></a><span class="lineno"> 55</span> </div>
<div class="line"><a name="l00056"></a><span class="lineno"> 56</span> <span class="comment">//</span></div>
<div class="line"><a name="l00057"></a><span class="lineno"> 57</span> <span class="comment">// Stage 1: Obtain upper bound for number of elements per row in C:</span></div>
<div class="line"><a name="l00058"></a><span class="lineno"> 58</span> <span class="comment">//</span></div>
<div class="line"><a name="l00059"></a><span class="lineno"> 59</span> <span class="keyword">template</span><<span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00060"></a><span class="lineno"> 60</span> __device__ IndexT <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#ac681c14ce17df9e05ee22df27353fcb1">round_to_next_power_of_2</a>(IndexT val)</div>
<div class="line"><a name="l00061"></a><span class="lineno"> 61</span> {</div>
<div class="line"><a name="l00062"></a><span class="lineno"> 62</span>  <span class="keywordflow">if</span> (val > 32)</div>
<div class="line"><a name="l00063"></a><span class="lineno"> 63</span>  <span class="keywordflow">return</span> 64; <span class="comment">// just to indicate that we need to split/factor the matrix!</span></div>
<div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (val > 16)</div>
<div class="line"><a name="l00065"></a><span class="lineno"> 65</span>  <span class="keywordflow">return</span> 32;</div>
<div class="line"><a name="l00066"></a><span class="lineno"> 66</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (val > 8)</div>
<div class="line"><a name="l00067"></a><span class="lineno"> 67</span>  <span class="keywordflow">return</span> 16;</div>
<div class="line"><a name="l00068"></a><span class="lineno"> 68</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (val > 4)</div>
<div class="line"><a name="l00069"></a><span class="lineno"> 69</span>  <span class="keywordflow">return</span> 8;</div>
<div class="line"><a name="l00070"></a><span class="lineno"> 70</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (val > 2)</div>
<div class="line"><a name="l00071"></a><span class="lineno"> 71</span>  <span class="keywordflow">return</span> 4;</div>
<div class="line"><a name="l00072"></a><span class="lineno"> 72</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (val > 1)</div>
<div class="line"><a name="l00073"></a><span class="lineno"> 73</span>  <span class="keywordflow">return</span> 2;</div>
<div class="line"><a name="l00074"></a><span class="lineno"> 74</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00075"></a><span class="lineno"> 75</span>  <span class="keywordflow">return</span> 1;</div>
<div class="line"><a name="l00076"></a><span class="lineno"> 76</span> }</div>
<div class="line"><a name="l00077"></a><span class="lineno"> 77</span> </div>
<div class="line"><a name="l00078"></a><span class="lineno"> 78</span> <span class="keyword">template</span><<span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00079"></a><span class="lineno"> 79</span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#af55c923f7bcf1fb5bef3ac98eee818d7">compressed_matrix_gemm_stage_1</a>(</div>
<div class="line"><a name="l00080"></a><span class="lineno"> 80</span>  <span class="keyword">const</span> IndexT * A_row_indices,</div>
<div class="line"><a name="l00081"></a><span class="lineno"> 81</span>  <span class="keyword">const</span> IndexT * A_col_indices,</div>
<div class="line"><a name="l00082"></a><span class="lineno"> 82</span>  IndexT A_size1,</div>
<div class="line"><a name="l00083"></a><span class="lineno"> 83</span>  <span class="keyword">const</span> IndexT * B_row_indices,</div>
<div class="line"><a name="l00084"></a><span class="lineno"> 84</span>  IndexT *subwarpsize_per_group,</div>
<div class="line"><a name="l00085"></a><span class="lineno"> 85</span>  IndexT *max_nnz_row_A_per_group,</div>
<div class="line"><a name="l00086"></a><span class="lineno"> 86</span>  IndexT *max_nnz_row_B_per_group)</div>
<div class="line"><a name="l00087"></a><span class="lineno"> 87</span> {</div>
<div class="line"><a name="l00088"></a><span class="lineno"> 88</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> subwarpsize_in_thread = 0;</div>
<div class="line"><a name="l00089"></a><span class="lineno"> 89</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> max_nnz_row_A = 0;</div>
<div class="line"><a name="l00090"></a><span class="lineno"> 90</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> max_nnz_row_B = 0;</div>
<div class="line"><a name="l00091"></a><span class="lineno"> 91</span> </div>
<div class="line"><a name="l00092"></a><span class="lineno"> 92</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> rows_per_group = (A_size1 - 1) / gridDim.x + 1;</div>
<div class="line"><a name="l00093"></a><span class="lineno"> 93</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_per_group_end = <a class="code" href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">min</a>(A_size1, rows_per_group * (blockIdx.x + 1));</div>
<div class="line"><a name="l00094"></a><span class="lineno"> 94</span> </div>
<div class="line"><a name="l00095"></a><span class="lineno"> 95</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> = rows_per_group * blockIdx.x + threadIdx.x; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> < row_per_group_end; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> += blockDim.x)</div>
<div class="line"><a name="l00096"></a><span class="lineno"> 96</span>  {</div>
<div class="line"><a name="l00097"></a><span class="lineno"> 97</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_row_start = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>];</div>
<div class="line"><a name="l00098"></a><span class="lineno"> 98</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_row_end = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>+1];</div>
<div class="line"><a name="l00099"></a><span class="lineno"> 99</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_num = A_row_end - A_row_start;</div>
<div class="line"><a name="l00100"></a><span class="lineno"> 100</span>  subwarpsize_in_thread = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>(A_row_end - A_row_start, subwarpsize_in_thread);</div>
<div class="line"><a name="l00101"></a><span class="lineno"> 101</span>  max_nnz_row_A = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>(max_nnz_row_A, row_num);</div>
<div class="line"><a name="l00102"></a><span class="lineno"> 102</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> j = A_row_start; j < A_row_end; ++j)</div>
<div class="line"><a name="l00103"></a><span class="lineno"> 103</span>  {</div>
<div class="line"><a name="l00104"></a><span class="lineno"> 104</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> col = A_col_indices[j];</div>
<div class="line"><a name="l00105"></a><span class="lineno"> 105</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_len_B = B_row_indices[col + 1] - B_row_indices[col];</div>
<div class="line"><a name="l00106"></a><span class="lineno"> 106</span>  max_nnz_row_B = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>(row_len_B, max_nnz_row_B);</div>
<div class="line"><a name="l00107"></a><span class="lineno"> 107</span>  }</div>
<div class="line"><a name="l00108"></a><span class="lineno"> 108</span>  }</div>
<div class="line"><a name="l00109"></a><span class="lineno"> 109</span> </div>
<div class="line"><a name="l00110"></a><span class="lineno"> 110</span>  <span class="comment">// reduction to obtain maximum in thread block</span></div>
<div class="line"><a name="l00111"></a><span class="lineno"> 111</span>  __shared__ <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shared_subwarpsize[256];</div>
<div class="line"><a name="l00112"></a><span class="lineno"> 112</span>  __shared__ <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shared_max_nnz_row_A[256];</div>
<div class="line"><a name="l00113"></a><span class="lineno"> 113</span>  __shared__ <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shared_max_nnz_row_B[256];</div>
<div class="line"><a name="l00114"></a><span class="lineno"> 114</span> </div>
<div class="line"><a name="l00115"></a><span class="lineno"> 115</span>  shared_subwarpsize[threadIdx.x] = subwarpsize_in_thread;</div>
<div class="line"><a name="l00116"></a><span class="lineno"> 116</span>  shared_max_nnz_row_A[threadIdx.x] = max_nnz_row_A;</div>
<div class="line"><a name="l00117"></a><span class="lineno"> 117</span>  shared_max_nnz_row_B[threadIdx.x] = max_nnz_row_B;</div>
<div class="line"><a name="l00118"></a><span class="lineno"> 118</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a> = blockDim.x/2; <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a> > 0; <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a> /= 2)</div>
<div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  {</div>
<div class="line"><a name="l00120"></a><span class="lineno"> 120</span>  __syncthreads();</div>
<div class="line"><a name="l00121"></a><span class="lineno"> 121</span>  <span class="keywordflow">if</span> (threadIdx.x < <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a>)</div>
<div class="line"><a name="l00122"></a><span class="lineno"> 122</span>  {</div>
<div class="line"><a name="l00123"></a><span class="lineno"> 123</span>  shared_subwarpsize[threadIdx.x] = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>( shared_subwarpsize[threadIdx.x], shared_subwarpsize[threadIdx.x + <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a>]);</div>
<div class="line"><a name="l00124"></a><span class="lineno"> 124</span>  shared_max_nnz_row_A[threadIdx.x] = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>(shared_max_nnz_row_A[threadIdx.x], shared_max_nnz_row_A[threadIdx.x + <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a>]);</div>
<div class="line"><a name="l00125"></a><span class="lineno"> 125</span>  shared_max_nnz_row_B[threadIdx.x] = <a class="code" href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">max</a>(shared_max_nnz_row_B[threadIdx.x], shared_max_nnz_row_B[threadIdx.x + <a class="code" href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">stride</a>]);</div>
<div class="line"><a name="l00126"></a><span class="lineno"> 126</span>  }</div>
<div class="line"><a name="l00127"></a><span class="lineno"> 127</span>  }</div>
<div class="line"><a name="l00128"></a><span class="lineno"> 128</span> </div>
<div class="line"><a name="l00129"></a><span class="lineno"> 129</span>  <span class="keywordflow">if</span> (threadIdx.x == 0)</div>
<div class="line"><a name="l00130"></a><span class="lineno"> 130</span>  {</div>
<div class="line"><a name="l00131"></a><span class="lineno"> 131</span>  subwarpsize_per_group[blockIdx.x] = <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#ac681c14ce17df9e05ee22df27353fcb1">round_to_next_power_of_2</a>(shared_subwarpsize[0]);</div>
<div class="line"><a name="l00132"></a><span class="lineno"> 132</span>  max_nnz_row_A_per_group[blockIdx.x] = shared_max_nnz_row_A[0];</div>
<div class="line"><a name="l00133"></a><span class="lineno"> 133</span>  max_nnz_row_B_per_group[blockIdx.x] = shared_max_nnz_row_B[0];</div>
<div class="line"><a name="l00134"></a><span class="lineno"> 134</span>  }</div>
<div class="line"><a name="l00135"></a><span class="lineno"> 135</span> }</div>
<div class="line"><a name="l00136"></a><span class="lineno"> 136</span> </div>
<div class="line"><a name="l00137"></a><span class="lineno"> 137</span> <span class="comment">//</span></div>
<div class="line"><a name="l00138"></a><span class="lineno"> 138</span> <span class="comment">// Stage 2: Determine sparsity pattern of C</span></div>
<div class="line"><a name="l00139"></a><span class="lineno"> 139</span> <span class="comment">//</span></div>
<div class="line"><a name="l00140"></a><span class="lineno"> 140</span> </div>
<div class="line"><a name="l00141"></a><span class="lineno"> 141</span> <span class="comment">// Using warp shuffle routines (CUDA arch 3.5)</span></div>
<div class="line"><a name="l00142"></a><span class="lineno"> 142</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00143"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#a5151284a26947b0204440ab90b1f4f4b"> 143</a></span> __device__ IndexT <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a5151284a26947b0204440ab90b1f4f4b">subwarp_minimum_shuffle</a>(IndexT min_index)</div>
<div class="line"><a name="l00144"></a><span class="lineno"> 144</span> {</div>
<div class="line"><a name="l00145"></a><span class="lineno"> 145</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i = SubWarpSizeV/2; i >= 1; i /= 2)</div>
<div class="line"><a name="l00146"></a><span class="lineno"> 146</span>  min_index = <a class="code" href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">min</a>(min_index, __shfl_xor((<span class="keywordtype">int</span>)min_index, (<span class="keywordtype">int</span>)i));</div>
<div class="line"><a name="l00147"></a><span class="lineno"> 147</span>  <span class="keywordflow">return</span> min_index;</div>
<div class="line"><a name="l00148"></a><span class="lineno"> 148</span> }</div>
<div class="line"><a name="l00149"></a><span class="lineno"> 149</span> </div>
<div class="line"><a name="l00150"></a><span class="lineno"> 150</span> <span class="comment">// Using shared memory</span></div>
<div class="line"><a name="l00151"></a><span class="lineno"> 151</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00152"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#a247433ee260c185ea2f08c3fdf164953"> 152</a></span> __device__ IndexT <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a247433ee260c185ea2f08c3fdf164953">subwarp_minimum_shared</a>(IndexT min_index, IndexT id_in_warp, IndexT *shared_buffer)</div>
<div class="line"><a name="l00153"></a><span class="lineno"> 153</span> {</div>
<div class="line"><a name="l00154"></a><span class="lineno"> 154</span>  shared_buffer[threadIdx.x] = min_index;</div>
<div class="line"><a name="l00155"></a><span class="lineno"> 155</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i = SubWarpSizeV/2; i >= 1; i /= 2)</div>
<div class="line"><a name="l00156"></a><span class="lineno"> 156</span>  shared_buffer[threadIdx.x] = <a class="code" href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">min</a>(shared_buffer[threadIdx.x], shared_buffer[(threadIdx.x + i) % 512]);</div>
<div class="line"><a name="l00157"></a><span class="lineno"> 157</span>  <span class="keywordflow">return</span> shared_buffer[threadIdx.x - id_in_warp];</div>
<div class="line"><a name="l00158"></a><span class="lineno"> 158</span> }</div>
<div class="line"><a name="l00159"></a><span class="lineno"> 159</span> </div>
<div class="line"><a name="l00160"></a><span class="lineno"> 160</span> </div>
<div class="line"><a name="l00161"></a><span class="lineno"> 161</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00162"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#a69ef499581661acb0831820266a8a7e1"> 162</a></span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a2526f31c8e9c88587938b341497e5666">compressed_matrix_gemm_stage_2</a>(</div>
<div class="line"><a name="l00163"></a><span class="lineno"> 163</span>  <span class="keyword">const</span> IndexT * A_row_indices,</div>
<div class="line"><a name="l00164"></a><span class="lineno"> 164</span>  <span class="keyword">const</span> IndexT * A_col_indices,</div>
<div class="line"><a name="l00165"></a><span class="lineno"> 165</span>  IndexT A_size1,</div>
<div class="line"><a name="l00166"></a><span class="lineno"> 166</span>  <span class="keyword">const</span> IndexT * B_row_indices,</div>
<div class="line"><a name="l00167"></a><span class="lineno"> 167</span>  <span class="keyword">const</span> IndexT * B_col_indices,</div>
<div class="line"><a name="l00168"></a><span class="lineno"> 168</span>  IndexT B_size2,</div>
<div class="line"><a name="l00169"></a><span class="lineno"> 169</span>  IndexT * C_row_indices)</div>
<div class="line"><a name="l00170"></a><span class="lineno"> 170</span> {</div>
<div class="line"><a name="l00171"></a><span class="lineno"> 171</span>  __shared__ <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shared_buffer[512];</div>
<div class="line"><a name="l00172"></a><span class="lineno"> 172</span> </div>
<div class="line"><a name="l00173"></a><span class="lineno"> 173</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> num_warps = blockDim.x / SubWarpSizeV;</div>
<div class="line"><a name="l00174"></a><span class="lineno"> 174</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> warp_id = threadIdx.x / SubWarpSizeV;</div>
<div class="line"><a name="l00175"></a><span class="lineno"> 175</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> id_in_warp = threadIdx.x % SubWarpSizeV;</div>
<div class="line"><a name="l00176"></a><span class="lineno"> 176</span> </div>
<div class="line"><a name="l00177"></a><span class="lineno"> 177</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> rows_per_group = (A_size1 - 1) / gridDim.x + 1;</div>
<div class="line"><a name="l00178"></a><span class="lineno"> 178</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_per_group_end = <a class="code" href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">min</a>(A_size1, rows_per_group * (blockIdx.x + 1));</div>
<div class="line"><a name="l00179"></a><span class="lineno"> 179</span> </div>
<div class="line"><a name="l00180"></a><span class="lineno"> 180</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> = rows_per_group * blockIdx.x + warp_id; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> < row_per_group_end; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> += num_warps)</div>
<div class="line"><a name="l00181"></a><span class="lineno"> 181</span>  {</div>
<div class="line"><a name="l00182"></a><span class="lineno"> 182</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_A_start = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>];</div>
<div class="line"><a name="l00183"></a><span class="lineno"> 183</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_A_end = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>+1];</div>
<div class="line"><a name="l00184"></a><span class="lineno"> 184</span> </div>
<div class="line"><a name="l00185"></a><span class="lineno"> 185</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> my_row_B = row_A_start + id_in_warp;</div>
<div class="line"><a name="l00186"></a><span class="lineno"> 186</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_index = (my_row_B < row_A_end) ? A_col_indices[my_row_B] : 0;</div>
<div class="line"><a name="l00187"></a><span class="lineno"> 187</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_start = (my_row_B < row_A_end) ? load_and_cache(B_row_indices + row_B_index) : 0;</div>
<div class="line"><a name="l00188"></a><span class="lineno"> 188</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_end = (my_row_B < row_A_end) ? load_and_cache(B_row_indices + row_B_index + 1) : 0;</div>
<div class="line"><a name="l00189"></a><span class="lineno"> 189</span> </div>
<div class="line"><a name="l00190"></a><span class="lineno"> 190</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> num_nnz = 0;</div>
<div class="line"><a name="l00191"></a><span class="lineno"> 191</span>  <span class="keywordflow">if</span> (row_A_end - row_A_start > 1) <span class="comment">// zero or no row can be processed faster</span></div>
<div class="line"><a name="l00192"></a><span class="lineno"> 192</span>  {</div>
<div class="line"><a name="l00193"></a><span class="lineno"> 193</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> current_front_index = (row_B_start < row_B_end) ? load_and_cache(B_col_indices + row_B_start) : B_size2;</div>
<div class="line"><a name="l00194"></a><span class="lineno"> 194</span> </div>
<div class="line"><a name="l00195"></a><span class="lineno"> 195</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00196"></a><span class="lineno"> 196</span>  {</div>
<div class="line"><a name="l00197"></a><span class="lineno"> 197</span>  <span class="comment">// determine current minimum (warp shuffle)</span></div>
<div class="line"><a name="l00198"></a><span class="lineno"> 198</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index = current_front_index;</div>
<div class="line"><a name="l00199"></a><span class="lineno"> 199</span>  min_index = subwarp_minimum_shared<SubWarpSizeV>(min_index, id_in_warp, shared_buffer);</div>
<div class="line"><a name="l00200"></a><span class="lineno"> 200</span> </div>
<div class="line"><a name="l00201"></a><span class="lineno"> 201</span>  <span class="keywordflow">if</span> (min_index == B_size2)</div>
<div class="line"><a name="l00202"></a><span class="lineno"> 202</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00203"></a><span class="lineno"> 203</span> </div>
<div class="line"><a name="l00204"></a><span class="lineno"> 204</span>  <span class="comment">// update front:</span></div>
<div class="line"><a name="l00205"></a><span class="lineno"> 205</span>  <span class="keywordflow">if</span> (current_front_index == min_index)</div>
<div class="line"><a name="l00206"></a><span class="lineno"> 206</span>  {</div>
<div class="line"><a name="l00207"></a><span class="lineno"> 207</span>  ++row_B_start;</div>
<div class="line"><a name="l00208"></a><span class="lineno"> 208</span>  current_front_index = (row_B_start < row_B_end) ? load_and_cache(B_col_indices + row_B_start) : B_size2;</div>
<div class="line"><a name="l00209"></a><span class="lineno"> 209</span>  }</div>
<div class="line"><a name="l00210"></a><span class="lineno"> 210</span> </div>
<div class="line"><a name="l00211"></a><span class="lineno"> 211</span>  ++num_nnz;</div>
<div class="line"><a name="l00212"></a><span class="lineno"> 212</span>  }</div>
<div class="line"><a name="l00213"></a><span class="lineno"> 213</span>  }</div>
<div class="line"><a name="l00214"></a><span class="lineno"> 214</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00215"></a><span class="lineno"> 215</span>  {</div>
<div class="line"><a name="l00216"></a><span class="lineno"> 216</span>  num_nnz = row_B_end - row_B_start;</div>
<div class="line"><a name="l00217"></a><span class="lineno"> 217</span>  }</div>
<div class="line"><a name="l00218"></a><span class="lineno"> 218</span> </div>
<div class="line"><a name="l00219"></a><span class="lineno"> 219</span>  <span class="keywordflow">if</span> (id_in_warp == 0)</div>
<div class="line"><a name="l00220"></a><span class="lineno"> 220</span>  C_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>] = num_nnz;</div>
<div class="line"><a name="l00221"></a><span class="lineno"> 221</span>  }</div>
<div class="line"><a name="l00222"></a><span class="lineno"> 222</span> </div>
<div class="line"><a name="l00223"></a><span class="lineno"> 223</span> }</div>
<div class="line"><a name="l00224"></a><span class="lineno"> 224</span> </div>
<div class="line"><a name="l00225"></a><span class="lineno"> 225</span> </div>
<div class="line"><a name="l00226"></a><span class="lineno"> 226</span> <span class="comment">//</span></div>
<div class="line"><a name="l00227"></a><span class="lineno"> 227</span> <span class="comment">// Stage 3: Fill C with values</span></div>
<div class="line"><a name="l00228"></a><span class="lineno"> 228</span> <span class="comment">//</span></div>
<div class="line"><a name="l00229"></a><span class="lineno"> 229</span> </div>
<div class="line"><a name="l00230"></a><span class="lineno"> 230</span> <span class="comment">// Using warp shuffle routines (CUDA arch 3.5)</span></div>
<div class="line"><a name="l00231"></a><span class="lineno"> 231</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00232"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#aea6aef114eb07f9ddf935c7441b232ca"> 232</a></span> __device__ <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#aea6aef114eb07f9ddf935c7441b232ca">subwarp_accumulate_shuffle</a>(<a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> output_value)</div>
<div class="line"><a name="l00233"></a><span class="lineno"> 233</span> {</div>
<div class="line"><a name="l00234"></a><span class="lineno"> 234</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i = SubWarpSizeV/2; i >= 1; i /= 2)</div>
<div class="line"><a name="l00235"></a><span class="lineno"> 235</span>  output_value += __shfl_xor((<span class="keywordtype">int</span>)output_value, (<span class="keywordtype">int</span>)i);</div>
<div class="line"><a name="l00236"></a><span class="lineno"> 236</span>  <span class="keywordflow">return</span> output_value;</div>
<div class="line"><a name="l00237"></a><span class="lineno"> 237</span> }</div>
<div class="line"><a name="l00238"></a><span class="lineno"> 238</span> </div>
<div class="line"><a name="l00239"></a><span class="lineno"> 239</span> <span class="comment">// Using shared memory</span></div>
<div class="line"><a name="l00240"></a><span class="lineno"> 240</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00241"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#a466d08cbedce419dde70a2114660b15f"> 241</a></span> __device__ <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a466d08cbedce419dde70a2114660b15f">subwarp_accumulate_shared</a>(<a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> output_value, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> id_in_warp, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *shared_buffer)</div>
<div class="line"><a name="l00242"></a><span class="lineno"> 242</span> {</div>
<div class="line"><a name="l00243"></a><span class="lineno"> 243</span>  shared_buffer[threadIdx.x] = output_value;</div>
<div class="line"><a name="l00244"></a><span class="lineno"> 244</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i = SubWarpSizeV/2; i >= 1; i /= 2)</div>
<div class="line"><a name="l00245"></a><span class="lineno"> 245</span>  shared_buffer[threadIdx.x] += shared_buffer[(threadIdx.x + i) % 512];</div>
<div class="line"><a name="l00246"></a><span class="lineno"> 246</span>  <span class="keywordflow">return</span> shared_buffer[threadIdx.x - id_in_warp];</div>
<div class="line"><a name="l00247"></a><span class="lineno"> 247</span> }</div>
<div class="line"><a name="l00248"></a><span class="lineno"> 248</span> </div>
<div class="line"><a name="l00249"></a><span class="lineno"> 249</span> </div>
<div class="line"><a name="l00250"></a><span class="lineno"> 250</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> SubWarpSizeV, <span class="keyword">typename</span> IndexT, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00251"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1cuda.html#a7c3aeafbfe928dd8044944d802f0b767"> 251</a></span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a07ca20b5c90e81c92d0e2004b457b10d">compressed_matrix_gemm_stage_3</a>(</div>
<div class="line"><a name="l00252"></a><span class="lineno"> 252</span>  <span class="keyword">const</span> IndexT * A_row_indices,</div>
<div class="line"><a name="l00253"></a><span class="lineno"> 253</span>  <span class="keyword">const</span> IndexT * A_col_indices,</div>
<div class="line"><a name="l00254"></a><span class="lineno"> 254</span>  <span class="keyword">const</span> <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> * A_elements,</div>
<div class="line"><a name="l00255"></a><span class="lineno"> 255</span>  IndexT A_size1,</div>
<div class="line"><a name="l00256"></a><span class="lineno"> 256</span>  <span class="keyword">const</span> IndexT * B_row_indices,</div>
<div class="line"><a name="l00257"></a><span class="lineno"> 257</span>  <span class="keyword">const</span> IndexT * B_col_indices,</div>
<div class="line"><a name="l00258"></a><span class="lineno"> 258</span>  <span class="keyword">const</span> <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> * B_elements,</div>
<div class="line"><a name="l00259"></a><span class="lineno"> 259</span>  IndexT B_size2,</div>
<div class="line"><a name="l00260"></a><span class="lineno"> 260</span>  IndexT <span class="keyword">const</span> * C_row_indices,</div>
<div class="line"><a name="l00261"></a><span class="lineno"> 261</span>  IndexT * C_col_indices,</div>
<div class="line"><a name="l00262"></a><span class="lineno"> 262</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> * C_elements)</div>
<div class="line"><a name="l00263"></a><span class="lineno"> 263</span> {</div>
<div class="line"><a name="l00264"></a><span class="lineno"> 264</span>  __shared__ <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shared_indices[512];</div>
<div class="line"><a name="l00265"></a><span class="lineno"> 265</span>  __shared__ <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> shared_values[512];</div>
<div class="line"><a name="l00266"></a><span class="lineno"> 266</span> </div>
<div class="line"><a name="l00267"></a><span class="lineno"> 267</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> num_warps = blockDim.x / SubWarpSizeV;</div>
<div class="line"><a name="l00268"></a><span class="lineno"> 268</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> warp_id = threadIdx.x / SubWarpSizeV;</div>
<div class="line"><a name="l00269"></a><span class="lineno"> 269</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> id_in_warp = threadIdx.x % SubWarpSizeV;</div>
<div class="line"><a name="l00270"></a><span class="lineno"> 270</span> </div>
<div class="line"><a name="l00271"></a><span class="lineno"> 271</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> rows_per_group = (A_size1 - 1) / gridDim.x + 1;</div>
<div class="line"><a name="l00272"></a><span class="lineno"> 272</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_per_group_end = <a class="code" href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">min</a>(A_size1, rows_per_group * (blockIdx.x + 1));</div>
<div class="line"><a name="l00273"></a><span class="lineno"> 273</span> </div>
<div class="line"><a name="l00274"></a><span class="lineno"> 274</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> = rows_per_group * blockIdx.x + warp_id; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> < row_per_group_end; <a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a> += num_warps)</div>
<div class="line"><a name="l00275"></a><span class="lineno"> 275</span>  {</div>
<div class="line"><a name="l00276"></a><span class="lineno"> 276</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_A_start = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>];</div>
<div class="line"><a name="l00277"></a><span class="lineno"> 277</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_A_end = A_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>+1];</div>
<div class="line"><a name="l00278"></a><span class="lineno"> 278</span> </div>
<div class="line"><a name="l00279"></a><span class="lineno"> 279</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> my_row_B = row_A_start + ((row_A_end - row_A_start > 1) ? id_in_warp : 0); <span class="comment">// special case: single row</span></div>
<div class="line"><a name="l00280"></a><span class="lineno"> 280</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_index = (my_row_B < row_A_end) ? A_col_indices[my_row_B] : 0;</div>
<div class="line"><a name="l00281"></a><span class="lineno"> 281</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_start = (my_row_B < row_A_end) ? load_and_cache(B_row_indices + row_B_index) : 0;</div>
<div class="line"><a name="l00282"></a><span class="lineno"> 282</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B_end = (my_row_B < row_A_end) ? load_and_cache(B_row_indices + row_B_index + 1) : 0;</div>
<div class="line"><a name="l00283"></a><span class="lineno"> 283</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> val_A = (my_row_B < row_A_end) ? A_elements[my_row_B] : 0;</div>
<div class="line"><a name="l00284"></a><span class="lineno"> 284</span> </div>
<div class="line"><a name="l00285"></a><span class="lineno"> 285</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_in_C = C_row_indices[<a class="code" href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">row</a>];</div>
<div class="line"><a name="l00286"></a><span class="lineno"> 286</span> </div>
<div class="line"><a name="l00287"></a><span class="lineno"> 287</span>  <span class="keywordflow">if</span> (row_A_end - row_A_start > 1)</div>
<div class="line"><a name="l00288"></a><span class="lineno"> 288</span>  {</div>
<div class="line"><a name="l00289"></a><span class="lineno"> 289</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> current_front_index = (row_B_start < row_B_end) ? load_and_cache(B_col_indices + row_B_start) : B_size2;</div>
<div class="line"><a name="l00290"></a><span class="lineno"> 290</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> current_front_value = (row_B_start < row_B_end) ? load_and_cache(B_elements + row_B_start) : 0;</div>
<div class="line"><a name="l00291"></a><span class="lineno"> 291</span> </div>
<div class="line"><a name="l00292"></a><span class="lineno"> 292</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_buffer = 0;</div>
<div class="line"><a name="l00293"></a><span class="lineno"> 293</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> value_buffer = 0;</div>
<div class="line"><a name="l00294"></a><span class="lineno"> 294</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> buffer_size = 0;</div>
<div class="line"><a name="l00295"></a><span class="lineno"> 295</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00296"></a><span class="lineno"> 296</span>  {</div>
<div class="line"><a name="l00297"></a><span class="lineno"> 297</span>  <span class="comment">// determine current minimum:</span></div>
<div class="line"><a name="l00298"></a><span class="lineno"> 298</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index = subwarp_minimum_shared<SubWarpSizeV>(current_front_index, id_in_warp, shared_indices);</div>
<div class="line"><a name="l00299"></a><span class="lineno"> 299</span> </div>
<div class="line"><a name="l00300"></a><span class="lineno"> 300</span>  <span class="keywordflow">if</span> (min_index == B_size2) <span class="comment">// done</span></div>
<div class="line"><a name="l00301"></a><span class="lineno"> 301</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00302"></a><span class="lineno"> 302</span> </div>
<div class="line"><a name="l00303"></a><span class="lineno"> 303</span>  <span class="comment">// compute entry in C:</span></div>
<div class="line"><a name="l00304"></a><span class="lineno"> 304</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> output_value = (current_front_index == min_index) ? val_A * current_front_value : 0;</div>
<div class="line"><a name="l00305"></a><span class="lineno"> 305</span>  output_value = subwarp_accumulate_shared<SubWarpSizeV>(output_value, id_in_warp, shared_values);</div>
<div class="line"><a name="l00306"></a><span class="lineno"> 306</span> </div>
<div class="line"><a name="l00307"></a><span class="lineno"> 307</span>  <span class="comment">// update front:</span></div>
<div class="line"><a name="l00308"></a><span class="lineno"> 308</span>  <span class="keywordflow">if</span> (current_front_index == min_index)</div>
<div class="line"><a name="l00309"></a><span class="lineno"> 309</span>  {</div>
<div class="line"><a name="l00310"></a><span class="lineno"> 310</span>  ++row_B_start;</div>
<div class="line"><a name="l00311"></a><span class="lineno"> 311</span>  current_front_index = (row_B_start < row_B_end) ? load_and_cache(B_col_indices + row_B_start) : B_size2;</div>
<div class="line"><a name="l00312"></a><span class="lineno"> 312</span>  current_front_value = (row_B_start < row_B_end) ? load_and_cache(B_elements + row_B_start) : 0;</div>
<div class="line"><a name="l00313"></a><span class="lineno"> 313</span>  }</div>
<div class="line"><a name="l00314"></a><span class="lineno"> 314</span> </div>
<div class="line"><a name="l00315"></a><span class="lineno"> 315</span>  <span class="comment">// write current front to register buffer:</span></div>
<div class="line"><a name="l00316"></a><span class="lineno"> 316</span>  index_buffer = (id_in_warp == buffer_size) ? min_index : index_buffer;</div>
<div class="line"><a name="l00317"></a><span class="lineno"> 317</span>  value_buffer = (id_in_warp == buffer_size) ? output_value : value_buffer;</div>
<div class="line"><a name="l00318"></a><span class="lineno"> 318</span>  ++buffer_size;</div>
<div class="line"><a name="l00319"></a><span class="lineno"> 319</span> </div>
<div class="line"><a name="l00320"></a><span class="lineno"> 320</span>  <span class="comment">// flush register buffer via a coalesced write once full:</span></div>
<div class="line"><a name="l00321"></a><span class="lineno"> 321</span>  <span class="keywordflow">if</span> (buffer_size == SubWarpSizeV)</div>
<div class="line"><a name="l00322"></a><span class="lineno"> 322</span>  {</div>
<div class="line"><a name="l00323"></a><span class="lineno"> 323</span>  C_col_indices[index_in_C + id_in_warp] = index_buffer;</div>
<div class="line"><a name="l00324"></a><span class="lineno"> 324</span>  C_elements[index_in_C + id_in_warp] = value_buffer;</div>
<div class="line"><a name="l00325"></a><span class="lineno"> 325</span>  }</div>
<div class="line"><a name="l00326"></a><span class="lineno"> 326</span> </div>
<div class="line"><a name="l00327"></a><span class="lineno"> 327</span>  index_in_C += (buffer_size == SubWarpSizeV) ? SubWarpSizeV : 0;</div>
<div class="line"><a name="l00328"></a><span class="lineno"> 328</span>  buffer_size = (buffer_size == SubWarpSizeV) ? 0 : buffer_size;</div>
<div class="line"><a name="l00329"></a><span class="lineno"> 329</span>  }</div>
<div class="line"><a name="l00330"></a><span class="lineno"> 330</span> </div>
<div class="line"><a name="l00331"></a><span class="lineno"> 331</span>  <span class="comment">// write remaining entries in register buffer to C:</span></div>
<div class="line"><a name="l00332"></a><span class="lineno"> 332</span>  <span class="keywordflow">if</span> (id_in_warp < buffer_size)</div>
<div class="line"><a name="l00333"></a><span class="lineno"> 333</span>  {</div>
<div class="line"><a name="l00334"></a><span class="lineno"> 334</span>  C_col_indices[index_in_C + id_in_warp] = index_buffer;</div>
<div class="line"><a name="l00335"></a><span class="lineno"> 335</span>  C_elements[index_in_C + id_in_warp] = value_buffer;</div>
<div class="line"><a name="l00336"></a><span class="lineno"> 336</span>  }</div>
<div class="line"><a name="l00337"></a><span class="lineno"> 337</span>  }</div>
<div class="line"><a name="l00338"></a><span class="lineno"> 338</span>  <span class="keywordflow">else</span> <span class="comment">// write respective row using the full subwarp:</span></div>
<div class="line"><a name="l00339"></a><span class="lineno"> 339</span>  {</div>
<div class="line"><a name="l00340"></a><span class="lineno"> 340</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i = row_B_start + id_in_warp; i < row_B_end; i += SubWarpSizeV)</div>
<div class="line"><a name="l00341"></a><span class="lineno"> 341</span>  {</div>
<div class="line"><a name="l00342"></a><span class="lineno"> 342</span>  C_col_indices[index_in_C + id_in_warp] = load_and_cache(B_col_indices + i);</div>
<div class="line"><a name="l00343"></a><span class="lineno"> 343</span>  C_elements[index_in_C + id_in_warp] = val_A * load_and_cache(B_elements + i);</div>
<div class="line"><a name="l00344"></a><span class="lineno"> 344</span>  index_in_C += SubWarpSizeV;</div>
<div class="line"><a name="l00345"></a><span class="lineno"> 345</span>  }</div>
<div class="line"><a name="l00346"></a><span class="lineno"> 346</span>  }</div>
<div class="line"><a name="l00347"></a><span class="lineno"> 347</span> </div>
<div class="line"><a name="l00348"></a><span class="lineno"> 348</span>  }</div>
<div class="line"><a name="l00349"></a><span class="lineno"> 349</span> </div>
<div class="line"><a name="l00350"></a><span class="lineno"> 350</span> }</div>
<div class="line"><a name="l00351"></a><span class="lineno"> 351</span> </div>
<div class="line"><a name="l00352"></a><span class="lineno"> 352</span> </div>
<div class="line"><a name="l00353"></a><span class="lineno"> 353</span> </div>
<div class="line"><a name="l00354"></a><span class="lineno"> 354</span> <span class="comment">//</span></div>
<div class="line"><a name="l00355"></a><span class="lineno"> 355</span> <span class="comment">// Decomposition kernels:</span></div>
<div class="line"><a name="l00356"></a><span class="lineno"> 356</span> <span class="comment">//</span></div>
<div class="line"><a name="l00357"></a><span class="lineno"> 357</span> <span class="keyword">template</span><<span class="keyword">typename</span> IndexT></div>
<div class="line"><a name="l00358"></a><span class="lineno"> 358</span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#ab759723d26a2f8b81377d4a4b0d92db9">compressed_matrix_gemm_decompose_1</a>(</div>
<div class="line"><a name="l00359"></a><span class="lineno"> 359</span>  <span class="keyword">const</span> IndexT * A_row_indices,</div>
<div class="line"><a name="l00360"></a><span class="lineno"> 360</span>  IndexT A_size1,</div>
<div class="line"><a name="l00361"></a><span class="lineno"> 361</span>  IndexT max_per_row,</div>
<div class="line"><a name="l00362"></a><span class="lineno"> 362</span>  IndexT *chunks_per_row)</div>
<div class="line"><a name="l00363"></a><span class="lineno"> 363</span> {</div>
<div class="line"><a name="l00364"></a><span class="lineno"> 364</span>  <span class="keywordflow">for</span> (IndexT i = blockIdx.x * blockDim.x + threadIdx.x; i < A_size1; i += blockDim.x * gridDim.x)</div>
<div class="line"><a name="l00365"></a><span class="lineno"> 365</span>  {</div>
<div class="line"><a name="l00366"></a><span class="lineno"> 366</span>  IndexT num_entries = A_row_indices[i+1] - A_row_indices[i];</div>
<div class="line"><a name="l00367"></a><span class="lineno"> 367</span>  chunks_per_row[i] = (num_entries < max_per_row) ? 1 : ((num_entries - 1)/ max_per_row + 1);</div>
<div class="line"><a name="l00368"></a><span class="lineno"> 368</span>  }</div>
<div class="line"><a name="l00369"></a><span class="lineno"> 369</span> }</div>
<div class="line"><a name="l00370"></a><span class="lineno"> 370</span> </div>
<div class="line"><a name="l00371"></a><span class="lineno"> 371</span> </div>
<div class="line"><a name="l00372"></a><span class="lineno"> 372</span> <span class="keyword">template</span><<span class="keyword">typename</span> IndexT, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00373"></a><span class="lineno"> 373</span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a80758e8edbc3622467f98d5a9bf2e826">compressed_matrix_gemm_A2</a>(</div>
<div class="line"><a name="l00374"></a><span class="lineno"> 374</span>  IndexT * A2_row_indices,</div>
<div class="line"><a name="l00375"></a><span class="lineno"> 375</span>  IndexT * A2_col_indices,</div>
<div class="line"><a name="l00376"></a><span class="lineno"> 376</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> * A2_elements,</div>
<div class="line"><a name="l00377"></a><span class="lineno"> 377</span>  IndexT A2_size1,</div>
<div class="line"><a name="l00378"></a><span class="lineno"> 378</span>  IndexT *new_row_buffer)</div>
<div class="line"><a name="l00379"></a><span class="lineno"> 379</span> {</div>
<div class="line"><a name="l00380"></a><span class="lineno"> 380</span>  <span class="keywordflow">for</span> (IndexT i = blockIdx.x * blockDim.x + threadIdx.x; i < A2_size1; i += blockDim.x * gridDim.x)</div>
<div class="line"><a name="l00381"></a><span class="lineno"> 381</span>  {</div>
<div class="line"><a name="l00382"></a><span class="lineno"> 382</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_start = new_row_buffer[i];</div>
<div class="line"><a name="l00383"></a><span class="lineno"> 383</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_stop = new_row_buffer[i+1];</div>
<div class="line"><a name="l00384"></a><span class="lineno"> 384</span> </div>
<div class="line"><a name="l00385"></a><span class="lineno"> 385</span>  A2_row_indices[i] = index_start;</div>
<div class="line"><a name="l00386"></a><span class="lineno"> 386</span> </div>
<div class="line"><a name="l00387"></a><span class="lineno"> 387</span>  <span class="keywordflow">for</span> (IndexT j = index_start; j < index_stop; ++j)</div>
<div class="line"><a name="l00388"></a><span class="lineno"> 388</span>  {</div>
<div class="line"><a name="l00389"></a><span class="lineno"> 389</span>  A2_col_indices[j] = j;</div>
<div class="line"><a name="l00390"></a><span class="lineno"> 390</span>  A2_elements[j] = <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1);</div>
<div class="line"><a name="l00391"></a><span class="lineno"> 391</span>  }</div>
<div class="line"><a name="l00392"></a><span class="lineno"> 392</span>  }</div>
<div class="line"><a name="l00393"></a><span class="lineno"> 393</span> </div>
<div class="line"><a name="l00394"></a><span class="lineno"> 394</span>  <span class="comment">// write last entry in row_buffer with global thread 0:</span></div>
<div class="line"><a name="l00395"></a><span class="lineno"> 395</span>  <span class="keywordflow">if</span> (threadIdx.x == 0 && blockIdx.x == 0)</div>
<div class="line"><a name="l00396"></a><span class="lineno"> 396</span>  A2_row_indices[A2_size1] = new_row_buffer[A2_size1];</div>
<div class="line"><a name="l00397"></a><span class="lineno"> 397</span> }</div>
<div class="line"><a name="l00398"></a><span class="lineno"> 398</span> </div>
<div class="line"><a name="l00399"></a><span class="lineno"> 399</span> <span class="keyword">template</span><<span class="keyword">typename</span> IndexT, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00400"></a><span class="lineno"> 400</span> __global__ <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a2ead3da9d15125a029d1ae747c35d915">compressed_matrix_gemm_G1</a>(</div>
<div class="line"><a name="l00401"></a><span class="lineno"> 401</span>  IndexT * G1_row_indices,</div>
<div class="line"><a name="l00402"></a><span class="lineno"> 402</span>  IndexT * G1_col_indices,</div>
<div class="line"><a name="l00403"></a><span class="lineno"> 403</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> * G1_elements,</div>
<div class="line"><a name="l00404"></a><span class="lineno"> 404</span>  IndexT G1_size1,</div>
<div class="line"><a name="l00405"></a><span class="lineno"> 405</span>  IndexT <span class="keyword">const</span> *A_row_indices,</div>
<div class="line"><a name="l00406"></a><span class="lineno"> 406</span>  IndexT <span class="keyword">const</span> *A_col_indices,</div>
<div class="line"><a name="l00407"></a><span class="lineno"> 407</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *A_elements,</div>
<div class="line"><a name="l00408"></a><span class="lineno"> 408</span>  IndexT A_size1,</div>
<div class="line"><a name="l00409"></a><span class="lineno"> 409</span>  IndexT A_nnz,</div>
<div class="line"><a name="l00410"></a><span class="lineno"> 410</span>  IndexT max_per_row,</div>
<div class="line"><a name="l00411"></a><span class="lineno"> 411</span>  IndexT *new_row_buffer)</div>
<div class="line"><a name="l00412"></a><span class="lineno"> 412</span> {</div>
<div class="line"><a name="l00413"></a><span class="lineno"> 413</span>  <span class="comment">// Part 1: Copy column indices and entries:</span></div>
<div class="line"><a name="l00414"></a><span class="lineno"> 414</span>  <span class="keywordflow">for</span> (IndexT i = blockIdx.x * blockDim.x + threadIdx.x; i < A_nnz; i += blockDim.x * gridDim.x)</div>
<div class="line"><a name="l00415"></a><span class="lineno"> 415</span>  {</div>
<div class="line"><a name="l00416"></a><span class="lineno"> 416</span>  G1_col_indices[i] = A_col_indices[i];</div>
<div class="line"><a name="l00417"></a><span class="lineno"> 417</span>  G1_elements[i] = A_elements[i];</div>
<div class="line"><a name="l00418"></a><span class="lineno"> 418</span>  }</div>
<div class="line"><a name="l00419"></a><span class="lineno"> 419</span> </div>
<div class="line"><a name="l00420"></a><span class="lineno"> 420</span>  <span class="comment">// Part 2: Derive new row indicies:</span></div>
<div class="line"><a name="l00421"></a><span class="lineno"> 421</span>  <span class="keywordflow">for</span> (IndexT i = blockIdx.x * blockDim.x + threadIdx.x; i < A_size1; i += blockDim.x * gridDim.x)</div>
<div class="line"><a name="l00422"></a><span class="lineno"> 422</span>  {</div>
<div class="line"><a name="l00423"></a><span class="lineno"> 423</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> old_start = A_row_indices[i];</div>
<div class="line"><a name="l00424"></a><span class="lineno"> 424</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> new_start = new_row_buffer[i];</div>
<div class="line"><a name="l00425"></a><span class="lineno"> 425</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_chunks = new_row_buffer[i+1] - new_start;</div>
<div class="line"><a name="l00426"></a><span class="lineno"> 426</span> </div>
<div class="line"><a name="l00427"></a><span class="lineno"> 427</span>  <span class="keywordflow">for</span> (IndexT j=0; j<row_chunks; ++j)</div>
<div class="line"><a name="l00428"></a><span class="lineno"> 428</span>  G1_row_indices[new_start + j] = old_start + j * max_per_row;</div>
<div class="line"><a name="l00429"></a><span class="lineno"> 429</span>  }</div>
<div class="line"><a name="l00430"></a><span class="lineno"> 430</span> </div>
<div class="line"><a name="l00431"></a><span class="lineno"> 431</span>  <span class="comment">// write last entry in row_buffer with global thread 0:</span></div>
<div class="line"><a name="l00432"></a><span class="lineno"> 432</span>  <span class="keywordflow">if</span> (threadIdx.x == 0 && blockIdx.x == 0)</div>
<div class="line"><a name="l00433"></a><span class="lineno"> 433</span>  G1_row_indices[G1_size1] = A_row_indices[A_size1];</div>
<div class="line"><a name="l00434"></a><span class="lineno"> 434</span> }</div>
<div class="line"><a name="l00435"></a><span class="lineno"> 435</span> </div>
<div class="line"><a name="l00436"></a><span class="lineno"> 436</span> </div>
<div class="line"><a name="l00437"></a><span class="lineno"> 437</span> </div>
<div class="line"><a name="l00447"></a><span class="lineno"> 447</span> <span class="keyword">template</span><<span class="keyword">class</span> NumericT, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> AlignmentV></div>
<div class="line"><a name="l00448"></a><span class="lineno"> 448</span> <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#afd7583ae441f66185764ae2fed763feb">prod_impl</a>(<a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> <span class="keyword">const</span> & A,</div>
<div class="line"><a name="l00449"></a><span class="lineno"> 449</span>  <a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> <span class="keyword">const</span> & B,</div>
<div class="line"><a name="l00450"></a><span class="lineno"> 450</span>  <a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> & C)</div>
<div class="line"><a name="l00451"></a><span class="lineno"> 451</span> {</div>
<div class="line"><a name="l00452"></a><span class="lineno"> 452</span>  C.<a class="code" href="classviennacl_1_1compressed__matrix.html#afcf49945a63836e1edfca536fcd191fb">resize</a>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>(), B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>(), <span class="keyword">false</span>);</div>
<div class="line"><a name="l00453"></a><span class="lineno"> 453</span> </div>
<div class="line"><a name="l00454"></a><span class="lineno"> 454</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> blocknum = 256;</div>
<div class="line"><a name="l00455"></a><span class="lineno"> 455</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> threadnum = 128;</div>
<div class="line"><a name="l00456"></a><span class="lineno"> 456</span> </div>
<div class="line"><a name="l00457"></a><span class="lineno"> 457</span>  <a class="code" href="classviennacl_1_1vector.html">viennacl::vector<unsigned int></a> subwarp_sizes(blocknum, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A)); <span class="comment">// upper bound for the nonzeros per row encountered for each work group</span></div>
<div class="line"><a name="l00458"></a><span class="lineno"> 458</span>  <a class="code" href="classviennacl_1_1vector.html">viennacl::vector<unsigned int></a> max_nnz_row_A(blocknum, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A)); <span class="comment">// upper bound for the nonzeros per row encountered for each work group</span></div>
<div class="line"><a name="l00459"></a><span class="lineno"> 459</span>  <a class="code" href="classviennacl_1_1vector.html">viennacl::vector<unsigned int></a> max_nnz_row_B(blocknum, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A)); <span class="comment">// upper bound for the nonzeros per row encountered for each work group</span></div>
<div class="line"><a name="l00460"></a><span class="lineno"> 460</span> </div>
<div class="line"><a name="l00461"></a><span class="lineno"> 461</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00462"></a><span class="lineno"> 462</span>  <span class="comment">// Stage 1: Determine upper bound for number of nonzeros</span></div>
<div class="line"><a name="l00463"></a><span class="lineno"> 463</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00464"></a><span class="lineno"> 464</span>  compressed_matrix_gemm_stage_1<<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00465"></a><span class="lineno"> 465</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00466"></a><span class="lineno"> 466</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00467"></a><span class="lineno"> 467</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00468"></a><span class="lineno"> 468</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(subwarp_sizes),</div>
<div class="line"><a name="l00469"></a><span class="lineno"> 469</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(max_nnz_row_A),</div>
<div class="line"><a name="l00470"></a><span class="lineno"> 470</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(max_nnz_row_B)</div>
<div class="line"><a name="l00471"></a><span class="lineno"> 471</span>  );</div>
<div class="line"><a name="l00472"></a><span class="lineno"> 472</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_1"</span>);</div>
<div class="line"><a name="l00473"></a><span class="lineno"> 473</span> </div>
<div class="line"><a name="l00474"></a><span class="lineno"> 474</span>  subwarp_sizes.switch_memory_context(<a class="code" href="classviennacl_1_1context.html">viennacl::context</a>(<a class="code" href="namespaceviennacl.html#a00b40450b6b2fd87aad3527d9b2084b8a427356f0fb1b8d32b28f37e36b272df4">MAIN_MEMORY</a>));</div>
<div class="line"><a name="l00475"></a><span class="lineno"> 475</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> * subwarp_sizes_ptr = viennacl::linalg::host_based::detail::extract_raw_pointer<unsigned int>(subwarp_sizes.handle());</div>
<div class="line"><a name="l00476"></a><span class="lineno"> 476</span> </div>
<div class="line"><a name="l00477"></a><span class="lineno"> 477</span>  max_nnz_row_A.switch_memory_context(<a class="code" href="classviennacl_1_1context.html">viennacl::context</a>(<a class="code" href="namespaceviennacl.html#a00b40450b6b2fd87aad3527d9b2084b8a427356f0fb1b8d32b28f37e36b272df4">MAIN_MEMORY</a>));</div>
<div class="line"><a name="l00478"></a><span class="lineno"> 478</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> * max_nnz_row_A_ptr = viennacl::linalg::host_based::detail::extract_raw_pointer<unsigned int>(max_nnz_row_A.handle());</div>
<div class="line"><a name="l00479"></a><span class="lineno"> 479</span> </div>
<div class="line"><a name="l00480"></a><span class="lineno"> 480</span>  max_nnz_row_B.switch_memory_context(<a class="code" href="classviennacl_1_1context.html">viennacl::context</a>(<a class="code" href="namespaceviennacl.html#a00b40450b6b2fd87aad3527d9b2084b8a427356f0fb1b8d32b28f37e36b272df4">MAIN_MEMORY</a>));</div>
<div class="line"><a name="l00481"></a><span class="lineno"> 481</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> * max_nnz_row_B_ptr = viennacl::linalg::host_based::detail::extract_raw_pointer<unsigned int>(max_nnz_row_B.handle());</div>
<div class="line"><a name="l00482"></a><span class="lineno"> 482</span> </div>
<div class="line"><a name="l00483"></a><span class="lineno"> 483</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> max_subwarp_size = 0;</div>
<div class="line"><a name="l00484"></a><span class="lineno"> 484</span>  <span class="comment">//std::cout << "Scratchpad offsets: " << std::endl;</span></div>
<div class="line"><a name="l00485"></a><span class="lineno"> 485</span>  <span class="keywordflow">for</span> (std::size_t i=0; i<subwarp_sizes.size(); ++i)</div>
<div class="line"><a name="l00486"></a><span class="lineno"> 486</span>  max_subwarp_size = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#a5d46fe9558b0e462f10fd44942ad4fc6">std::max</a>(max_subwarp_size, subwarp_sizes_ptr[i]);</div>
<div class="line"><a name="l00487"></a><span class="lineno"> 487</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_max_nnz_per_row = 0;</div>
<div class="line"><a name="l00488"></a><span class="lineno"> 488</span>  <span class="keywordflow">for</span> (std::size_t i=0; i<max_nnz_row_A.size(); ++i)</div>
<div class="line"><a name="l00489"></a><span class="lineno"> 489</span>  A_max_nnz_per_row = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#a5d46fe9558b0e462f10fd44942ad4fc6">std::max</a>(A_max_nnz_per_row, max_nnz_row_A_ptr[i]);</div>
<div class="line"><a name="l00490"></a><span class="lineno"> 490</span> </div>
<div class="line"><a name="l00491"></a><span class="lineno"> 491</span>  <span class="keywordflow">if</span> (max_subwarp_size > 32)</div>
<div class="line"><a name="l00492"></a><span class="lineno"> 492</span>  {</div>
<div class="line"><a name="l00493"></a><span class="lineno"> 493</span>  <span class="comment">// determine augmented size:</span></div>
<div class="line"><a name="l00494"></a><span class="lineno"> 494</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> max_entries_in_G = 32;</div>
<div class="line"><a name="l00495"></a><span class="lineno"> 495</span>  <span class="keywordflow">if</span> (A_max_nnz_per_row <= 256)</div>
<div class="line"><a name="l00496"></a><span class="lineno"> 496</span>  max_entries_in_G = 16;</div>
<div class="line"><a name="l00497"></a><span class="lineno"> 497</span>  <span class="keywordflow">if</span> (A_max_nnz_per_row <= 64)</div>
<div class="line"><a name="l00498"></a><span class="lineno"> 498</span>  max_entries_in_G = 8;</div>
<div class="line"><a name="l00499"></a><span class="lineno"> 499</span> </div>
<div class="line"><a name="l00500"></a><span class="lineno"> 500</span>  <a class="code" href="classviennacl_1_1vector.html">viennacl::vector<unsigned int></a> exclusive_scan_helper(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>() + 1, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00501"></a><span class="lineno"> 501</span>  compressed_matrix_gemm_decompose_1<<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00502"></a><span class="lineno"> 502</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00503"></a><span class="lineno"> 503</span>  static_cast<unsigned int>(max_entries_in_G),</div>
<div class="line"><a name="l00504"></a><span class="lineno"> 504</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(exclusive_scan_helper)</div>
<div class="line"><a name="l00505"></a><span class="lineno"> 505</span>  );</div>
<div class="line"><a name="l00506"></a><span class="lineno"> 506</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_decompose_1"</span>);</div>
<div class="line"><a name="l00507"></a><span class="lineno"> 507</span> </div>
<div class="line"><a name="l00508"></a><span class="lineno"> 508</span>  <a class="code" href="namespaceviennacl_1_1linalg.html#a7f1b00757223e6feecf70a8657a9f096">viennacl::linalg::exclusive_scan</a>(exclusive_scan_helper);</div>
<div class="line"><a name="l00509"></a><span class="lineno"> 509</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> augmented_size = exclusive_scan_helper[A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()];</div>
<div class="line"><a name="l00510"></a><span class="lineno"> 510</span> </div>
<div class="line"><a name="l00511"></a><span class="lineno"> 511</span>  <span class="comment">// split A = A2 * G1</span></div>
<div class="line"><a name="l00512"></a><span class="lineno"> 512</span>  <a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> A2(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>(), augmented_size, augmented_size, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00513"></a><span class="lineno"> 513</span>  <a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> G1(augmented_size, A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>(), A.<a class="code" href="classviennacl_1_1compressed__matrix.html#ae69ca21ded644fdd0c7a5168011b13ed">nnz</a>(), <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00514"></a><span class="lineno"> 514</span> </div>
<div class="line"><a name="l00515"></a><span class="lineno"> 515</span>  <span class="comment">// fill A2:</span></div>
<div class="line"><a name="l00516"></a><span class="lineno"> 516</span>  compressed_matrix_gemm_A2<<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A2.handle1()),</div>
<div class="line"><a name="l00517"></a><span class="lineno"> 517</span>  viennacl::cuda_arg<unsigned int>(A2.handle2()),</div>
<div class="line"><a name="l00518"></a><span class="lineno"> 518</span>  viennacl::cuda_arg<NumericT>(A2.handle()),</div>
<div class="line"><a name="l00519"></a><span class="lineno"> 519</span>  static_cast<unsigned int>(A2.size1()),</div>
<div class="line"><a name="l00520"></a><span class="lineno"> 520</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(exclusive_scan_helper)</div>
<div class="line"><a name="l00521"></a><span class="lineno"> 521</span>  );</div>
<div class="line"><a name="l00522"></a><span class="lineno"> 522</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_A2"</span>);</div>
<div class="line"><a name="l00523"></a><span class="lineno"> 523</span> </div>
<div class="line"><a name="l00524"></a><span class="lineno"> 524</span>  <span class="comment">// fill G1:</span></div>
<div class="line"><a name="l00525"></a><span class="lineno"> 525</span>  compressed_matrix_gemm_G1<<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(G1.handle1()),</div>
<div class="line"><a name="l00526"></a><span class="lineno"> 526</span>  viennacl::cuda_arg<unsigned int>(G1.handle2()),</div>
<div class="line"><a name="l00527"></a><span class="lineno"> 527</span>  viennacl::cuda_arg<NumericT>(G1.handle()),</div>
<div class="line"><a name="l00528"></a><span class="lineno"> 528</span>  static_cast<unsigned int>(G1.size1()),</div>
<div class="line"><a name="l00529"></a><span class="lineno"> 529</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00530"></a><span class="lineno"> 530</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00531"></a><span class="lineno"> 531</span>  viennacl::cuda_arg<NumericT>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00532"></a><span class="lineno"> 532</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00533"></a><span class="lineno"> 533</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#ae69ca21ded644fdd0c7a5168011b13ed">nnz</a>()),</div>
<div class="line"><a name="l00534"></a><span class="lineno"> 534</span>  static_cast<unsigned int>(max_entries_in_G),</div>
<div class="line"><a name="l00535"></a><span class="lineno"> 535</span>  <a class="code" href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a>(exclusive_scan_helper)</div>
<div class="line"><a name="l00536"></a><span class="lineno"> 536</span>  );</div>
<div class="line"><a name="l00537"></a><span class="lineno"> 537</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_G1"</span>);</div>
<div class="line"><a name="l00538"></a><span class="lineno"> 538</span> </div>
<div class="line"><a name="l00539"></a><span class="lineno"> 539</span>  <span class="comment">// compute tmp = G1 * B;</span></div>
<div class="line"><a name="l00540"></a><span class="lineno"> 540</span>  <span class="comment">// C = A2 * tmp;</span></div>
<div class="line"><a name="l00541"></a><span class="lineno"> 541</span>  <a class="code" href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix<NumericT, AlignmentV></a> tmp(G1.size1(), B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>(), 0, <a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00542"></a><span class="lineno"> 542</span>  <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#afd7583ae441f66185764ae2fed763feb">prod_impl</a>(G1, B, tmp); <span class="comment">// this runs a standard RMerge without decomposition of G1</span></div>
<div class="line"><a name="l00543"></a><span class="lineno"> 543</span>  <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#afd7583ae441f66185764ae2fed763feb">prod_impl</a>(A2, tmp, C); <span class="comment">// this may split A2 again</span></div>
<div class="line"><a name="l00544"></a><span class="lineno"> 544</span>  <span class="keywordflow">return</span>;</div>
<div class="line"><a name="l00545"></a><span class="lineno"> 545</span>  }</div>
<div class="line"><a name="l00546"></a><span class="lineno"> 546</span> </div>
<div class="line"><a name="l00547"></a><span class="lineno"> 547</span>  <span class="comment">//std::cout << "Running RMerge with subwarp size " << max_subwarp_size << std::endl;</span></div>
<div class="line"><a name="l00548"></a><span class="lineno"> 548</span> </div>
<div class="line"><a name="l00549"></a><span class="lineno"> 549</span>  subwarp_sizes.switch_memory_context(<a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00550"></a><span class="lineno"> 550</span>  max_nnz_row_A.switch_memory_context(<a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00551"></a><span class="lineno"> 551</span>  max_nnz_row_B.switch_memory_context(<a class="code" href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a>(A));</div>
<div class="line"><a name="l00552"></a><span class="lineno"> 552</span> </div>
<div class="line"><a name="l00553"></a><span class="lineno"> 553</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00554"></a><span class="lineno"> 554</span>  <span class="comment">// Stage 2: Determine pattern of C</span></div>
<div class="line"><a name="l00555"></a><span class="lineno"> 555</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00556"></a><span class="lineno"> 556</span> </div>
<div class="line"><a name="l00557"></a><span class="lineno"> 557</span>  <span class="keywordflow">if</span> (max_subwarp_size == 32)</div>
<div class="line"><a name="l00558"></a><span class="lineno"> 558</span>  {</div>
<div class="line"><a name="l00559"></a><span class="lineno"> 559</span>  compressed_matrix_gemm_stage_2<32><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00560"></a><span class="lineno"> 560</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00561"></a><span class="lineno"> 561</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00562"></a><span class="lineno"> 562</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00563"></a><span class="lineno"> 563</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00564"></a><span class="lineno"> 564</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00565"></a><span class="lineno"> 565</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>())</div>
<div class="line"><a name="l00566"></a><span class="lineno"> 566</span>  );</div>
<div class="line"><a name="l00567"></a><span class="lineno"> 567</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_2"</span>);</div>
<div class="line"><a name="l00568"></a><span class="lineno"> 568</span>  }</div>
<div class="line"><a name="l00569"></a><span class="lineno"> 569</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (max_subwarp_size == 16)</div>
<div class="line"><a name="l00570"></a><span class="lineno"> 570</span>  {</div>
<div class="line"><a name="l00571"></a><span class="lineno"> 571</span>  compressed_matrix_gemm_stage_2<16><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00572"></a><span class="lineno"> 572</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00573"></a><span class="lineno"> 573</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00574"></a><span class="lineno"> 574</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00575"></a><span class="lineno"> 575</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00576"></a><span class="lineno"> 576</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00577"></a><span class="lineno"> 577</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>())</div>
<div class="line"><a name="l00578"></a><span class="lineno"> 578</span>  );</div>
<div class="line"><a name="l00579"></a><span class="lineno"> 579</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_2"</span>);</div>
<div class="line"><a name="l00580"></a><span class="lineno"> 580</span>  }</div>
<div class="line"><a name="l00581"></a><span class="lineno"> 581</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00582"></a><span class="lineno"> 582</span>  {</div>
<div class="line"><a name="l00583"></a><span class="lineno"> 583</span>  compressed_matrix_gemm_stage_2<8><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00584"></a><span class="lineno"> 584</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00585"></a><span class="lineno"> 585</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00586"></a><span class="lineno"> 586</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00587"></a><span class="lineno"> 587</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00588"></a><span class="lineno"> 588</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00589"></a><span class="lineno"> 589</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>())</div>
<div class="line"><a name="l00590"></a><span class="lineno"> 590</span>  );</div>
<div class="line"><a name="l00591"></a><span class="lineno"> 591</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_2"</span>);</div>
<div class="line"><a name="l00592"></a><span class="lineno"> 592</span>  }</div>
<div class="line"><a name="l00593"></a><span class="lineno"> 593</span> </div>
<div class="line"><a name="l00594"></a><span class="lineno"> 594</span>  <span class="comment">// exclusive scan on C.handle1(), ultimately allowing to allocate remaining memory for C</span></div>
<div class="line"><a name="l00595"></a><span class="lineno"> 595</span>  <a class="code" href="classviennacl_1_1backend_1_1typesafe__host__array.html">viennacl::backend::typesafe_host_array<unsigned int></a> row_buffer(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>(), C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>() + 1);</div>
<div class="line"><a name="l00596"></a><span class="lineno"> 596</span>  <a class="code" href="namespaceviennacl_1_1backend.html#a62854cfd6f04404b274f8ede36f63e2d">viennacl::backend::memory_read</a>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>(), 0, row_buffer.<a class="code" href="classviennacl_1_1backend_1_1mem__handle.html#ac8373f0d899b89c843e14de4cb7a1c4a">raw_size</a>(), row_buffer.get());</div>
<div class="line"><a name="l00597"></a><span class="lineno"> 597</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> current_offset = 0;</div>
<div class="line"><a name="l00598"></a><span class="lineno"> 598</span>  <span class="keywordflow">for</span> (std::size_t i=0; i<C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>(); ++i)</div>
<div class="line"><a name="l00599"></a><span class="lineno"> 599</span>  {</div>
<div class="line"><a name="l00600"></a><span class="lineno"> 600</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> tmp = row_buffer[i];</div>
<div class="line"><a name="l00601"></a><span class="lineno"> 601</span>  row_buffer.set(i, current_offset);</div>
<div class="line"><a name="l00602"></a><span class="lineno"> 602</span>  current_offset += tmp;</div>
<div class="line"><a name="l00603"></a><span class="lineno"> 603</span>  }</div>
<div class="line"><a name="l00604"></a><span class="lineno"> 604</span>  row_buffer.<a class="code" href="classviennacl_1_1backend_1_1typesafe__host__array.html#ac31bfde381eb40e0349678252c444af9">set</a>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>(), current_offset);</div>
<div class="line"><a name="l00605"></a><span class="lineno"> 605</span>  <a class="code" href="namespaceviennacl_1_1backend.html#a06bdedb2bc72dc1922cada91e9bbbd61">viennacl::backend::memory_write</a>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>(), 0, row_buffer.<a class="code" href="classviennacl_1_1backend_1_1mem__handle.html#ac8373f0d899b89c843e14de4cb7a1c4a">raw_size</a>(), row_buffer.get());</div>
<div class="line"><a name="l00606"></a><span class="lineno"> 606</span> </div>
<div class="line"><a name="l00607"></a><span class="lineno"> 607</span> </div>
<div class="line"><a name="l00608"></a><span class="lineno"> 608</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00609"></a><span class="lineno"> 609</span>  <span class="comment">// Stage 3: Compute entries in C</span></div>
<div class="line"><a name="l00610"></a><span class="lineno"> 610</span>  <span class="comment">//</span></div>
<div class="line"><a name="l00611"></a><span class="lineno"> 611</span>  C.<a class="code" href="classviennacl_1_1compressed__matrix.html#aba08ea5f9d1c12dff5da58e0dff7a6fe">reserve</a>(current_offset, <span class="keyword">false</span>);</div>
<div class="line"><a name="l00612"></a><span class="lineno"> 612</span> </div>
<div class="line"><a name="l00613"></a><span class="lineno"> 613</span>  <span class="keywordflow">if</span> (max_subwarp_size == 32)</div>
<div class="line"><a name="l00614"></a><span class="lineno"> 614</span>  {</div>
<div class="line"><a name="l00615"></a><span class="lineno"> 615</span>  compressed_matrix_gemm_stage_3<32><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00616"></a><span class="lineno"> 616</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00617"></a><span class="lineno"> 617</span>  viennacl::cuda_arg<NumericT>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00618"></a><span class="lineno"> 618</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00619"></a><span class="lineno"> 619</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00620"></a><span class="lineno"> 620</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00621"></a><span class="lineno"> 621</span>  viennacl::cuda_arg<NumericT>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00622"></a><span class="lineno"> 622</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00623"></a><span class="lineno"> 623</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00624"></a><span class="lineno"> 624</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00625"></a><span class="lineno"> 625</span>  viennacl::cuda_arg<NumericT>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>())</div>
<div class="line"><a name="l00626"></a><span class="lineno"> 626</span>  );</div>
<div class="line"><a name="l00627"></a><span class="lineno"> 627</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_3"</span>);</div>
<div class="line"><a name="l00628"></a><span class="lineno"> 628</span>  }</div>
<div class="line"><a name="l00629"></a><span class="lineno"> 629</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (max_subwarp_size == 16)</div>
<div class="line"><a name="l00630"></a><span class="lineno"> 630</span>  {</div>
<div class="line"><a name="l00631"></a><span class="lineno"> 631</span>  compressed_matrix_gemm_stage_3<16><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00632"></a><span class="lineno"> 632</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00633"></a><span class="lineno"> 633</span>  viennacl::cuda_arg<NumericT>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00634"></a><span class="lineno"> 634</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00635"></a><span class="lineno"> 635</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00636"></a><span class="lineno"> 636</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00637"></a><span class="lineno"> 637</span>  viennacl::cuda_arg<NumericT>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00638"></a><span class="lineno"> 638</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00639"></a><span class="lineno"> 639</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00640"></a><span class="lineno"> 640</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00641"></a><span class="lineno"> 641</span>  viennacl::cuda_arg<NumericT>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>())</div>
<div class="line"><a name="l00642"></a><span class="lineno"> 642</span>  );</div>
<div class="line"><a name="l00643"></a><span class="lineno"> 643</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_3"</span>);</div>
<div class="line"><a name="l00644"></a><span class="lineno"> 644</span>  }</div>
<div class="line"><a name="l00645"></a><span class="lineno"> 645</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00646"></a><span class="lineno"> 646</span>  {</div>
<div class="line"><a name="l00647"></a><span class="lineno"> 647</span>  compressed_matrix_gemm_stage_3<8><<<blocknum, threadnum>>>(viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00648"></a><span class="lineno"> 648</span>  viennacl::cuda_arg<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00649"></a><span class="lineno"> 649</span>  viennacl::cuda_arg<NumericT>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00650"></a><span class="lineno"> 650</span>  static_cast<unsigned int>(A.<a class="code" href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">size1</a>()),</div>
<div class="line"><a name="l00651"></a><span class="lineno"> 651</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00652"></a><span class="lineno"> 652</span>  viennacl::cuda_arg<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00653"></a><span class="lineno"> 653</span>  viennacl::cuda_arg<NumericT>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>()),</div>
<div class="line"><a name="l00654"></a><span class="lineno"> 654</span>  static_cast<unsigned int>(B.<a class="code" href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">size2</a>()),</div>
<div class="line"><a name="l00655"></a><span class="lineno"> 655</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">handle1</a>()),</div>
<div class="line"><a name="l00656"></a><span class="lineno"> 656</span>  viennacl::cuda_arg<unsigned int>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">handle2</a>()),</div>
<div class="line"><a name="l00657"></a><span class="lineno"> 657</span>  viennacl::cuda_arg<NumericT>(C.<a class="code" href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">handle</a>())</div>
<div class="line"><a name="l00658"></a><span class="lineno"> 658</span>  );</div>
<div class="line"><a name="l00659"></a><span class="lineno"> 659</span>  <a class="code" href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a>(<span class="stringliteral">"compressed_matrix_gemm_stage_3"</span>);</div>
<div class="line"><a name="l00660"></a><span class="lineno"> 660</span>  }</div>
<div class="line"><a name="l00661"></a><span class="lineno"> 661</span> </div>
<div class="line"><a name="l00662"></a><span class="lineno"> 662</span> }</div>
<div class="line"><a name="l00663"></a><span class="lineno"> 663</span> </div>
<div class="line"><a name="l00664"></a><span class="lineno"> 664</span> } <span class="comment">// namespace cuda</span></div>
<div class="line"><a name="l00665"></a><span class="lineno"> 665</span> } <span class="comment">//namespace linalg</span></div>
<div class="line"><a name="l00666"></a><span class="lineno"> 666</span> } <span class="comment">//namespace viennacl</span></div>
<div class="line"><a name="l00667"></a><span class="lineno"> 667</span> </div>
<div class="line"><a name="l00668"></a><span class="lineno"> 668</span> </div>
<div class="line"><a name="l00669"></a><span class="lineno"> 669</span> <span class="preprocessor">#endif</span></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_a4fc12fc4abfef4a1426575a2d73f18ab"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#a4fc12fc4abfef4a1426575a2d73f18ab">viennacl::compressed_matrix::size2</a></div><div class="ttdeci">const vcl_size_t & size2() const </div><div class="ttdoc">Returns the number of columns. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00929">compressed_matrix.hpp:929</a></div></div>
<div class="ttc" id="classviennacl_1_1backend_1_1typesafe__host__array_html"><div class="ttname"><a href="classviennacl_1_1backend_1_1typesafe__host__array.html">viennacl::backend::typesafe_host_array</a></div><div class="ttdoc">Helper class implementing an array on the host. Default case: No conversion necessary. </div><div class="ttdef"><b>Definition:</b> <a href="backend_2util_8hpp_source.html#l00092">util.hpp:92</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1backend_html_a06bdedb2bc72dc1922cada91e9bbbd61"><div class="ttname"><a href="namespaceviennacl_1_1backend.html#a06bdedb2bc72dc1922cada91e9bbbd61">viennacl::backend::memory_write</a></div><div class="ttdeci">void memory_write(mem_handle &dst_buffer, vcl_size_t dst_offset, vcl_size_t bytes_to_write, const void *ptr, bool async=false)</div><div class="ttdoc">Writes data from main RAM identified by 'ptr' to the buffer identified by 'dst_buffer'. </div><div class="ttdef"><b>Definition:</b> <a href="memory_8hpp_source.html#l00220">memory.hpp:220</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a247433ee260c185ea2f08c3fdf164953"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a247433ee260c185ea2f08c3fdf164953">viennacl::linalg::cuda::subwarp_minimum_shared</a></div><div class="ttdeci">__device__ IndexT subwarp_minimum_shared(IndexT min_index, IndexT id_in_warp, IndexT *shared_buffer)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__rmerge_8hpp_source.html#l00152">spgemm_rmerge.hpp:152</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_a463cf1739f9cdd387aa185cb574db183"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#a463cf1739f9cdd387aa185cb574db183">viennacl::compressed_matrix::size1</a></div><div class="ttdeci">const vcl_size_t & size1() const </div><div class="ttdoc">Returns the number of rows. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00927">compressed_matrix.hpp:927</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_ab759723d26a2f8b81377d4a4b0d92db9"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#ab759723d26a2f8b81377d4a4b0d92db9">viennacl::linalg::cuda::compressed_matrix_gemm_decompose_1</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_decompose_1(const IndexT *A_row_indices, IndexT A_size1, IndexT max_per_row, IndexT *chunks_per_row)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00469">spgemm.hpp:469</a></div></div>
<div class="ttc" id="tools_8hpp_html"><div class="ttname"><a href="tools_8hpp.html">tools.hpp</a></div><div class="ttdoc">Various little tools used here and there in ViennaCL. </div></div>
<div class="ttc" id="namespaceviennacl_1_1traits_html_a09997870f4802fa5d4ac2c43cf4020d1"><div class="ttname"><a href="namespaceviennacl_1_1traits.html#a09997870f4802fa5d4ac2c43cf4020d1">viennacl::traits::stride</a></div><div class="ttdeci">result_of::size_type< viennacl::vector_base< T > >::type stride(viennacl::vector_base< T > const &s)</div><div class="ttdef"><b>Definition:</b> <a href="stride_8hpp_source.html#l00045">stride.hpp:45</a></div></div>
<div class="ttc" id="forwards_8h_html"><div class="ttname"><a href="forwards_8h.html">forwards.h</a></div><div class="ttdoc">This file provides the forward declarations for the main types used within ViennaCL. </div></div>
<div class="ttc" id="namespaceviennacl_1_1backend_html_a62854cfd6f04404b274f8ede36f63e2d"><div class="ttname"><a href="namespaceviennacl_1_1backend.html#a62854cfd6f04404b274f8ede36f63e2d">viennacl::backend::memory_read</a></div><div class="ttdeci">void memory_read(mem_handle const &src_buffer, vcl_size_t src_offset, vcl_size_t bytes_to_read, void *ptr, bool async=false)</div><div class="ttdoc">Reads data from a buffer back to main RAM. </div><div class="ttdef"><b>Definition:</b> <a href="memory_8hpp_source.html#l00261">memory.hpp:261</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1detail_html_a5d46fe9558b0e462f10fd44942ad4fc6"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1detail.html#a5d46fe9558b0e462f10fd44942ad4fc6">viennacl::linalg::detail::max</a></div><div class="ttdeci">T max(const T &lhs, const T &rhs)</div><div class="ttdoc">Maximum. </div><div class="ttdef"><b>Definition:</b> <a href="linalg_2detail_2bisect_2util_8hpp_source.html#l00059">util.hpp:59</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_afd7583ae441f66185764ae2fed763feb"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#afd7583ae441f66185764ae2fed763feb">viennacl::linalg::cuda::prod_impl</a></div><div class="ttdeci">void prod_impl(const matrix_base< NumericT > &mat, bool mat_transpose, const vector_base< NumericT > &vec, vector_base< NumericT > &result)</div><div class="ttdoc">Carries out matrix-vector multiplication. </div><div class="ttdef"><b>Definition:</b> <a href="cuda_2matrix__operations_8hpp_source.html#l01464">matrix_operations.hpp:1464</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_a87a0ad5f26983b1c2d24ee302d886562"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#a87a0ad5f26983b1c2d24ee302d886562">viennacl::compressed_matrix::handle</a></div><div class="ttdeci">const handle_type & handle() const </div><div class="ttdoc">Returns the OpenCL handle to the matrix entry array. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00942">compressed_matrix.hpp:942</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_af71dec61a70e8df4f78a527aa989a106"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#af71dec61a70e8df4f78a527aa989a106">viennacl::compressed_matrix::handle1</a></div><div class="ttdeci">const handle_type & handle1() const </div><div class="ttdoc">Returns the OpenCL handle to the row index array. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00936">compressed_matrix.hpp:936</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_ae69ca21ded644fdd0c7a5168011b13ed"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#ae69ca21ded644fdd0c7a5168011b13ed">viennacl::compressed_matrix::nnz</a></div><div class="ttdeci">const vcl_size_t & nnz() const </div><div class="ttdoc">Returns the number of nonzero entries. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00931">compressed_matrix.hpp:931</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a07ca20b5c90e81c92d0e2004b457b10d"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a07ca20b5c90e81c92d0e2004b457b10d">viennacl::linalg::cuda::compressed_matrix_gemm_stage_3</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_stage_3(const IndexT *A_row_indices, const IndexT *A_col_indices, const NumericT *A_elements, IndexT A_size1, const IndexT *B_row_indices, const IndexT *B_col_indices, const NumericT *B_elements, IndexT B_size2, IndexT const *C_row_indices, IndexT *C_col_indices, NumericT *C_elements, unsigned int *subwarpsize_array, unsigned int *max_row_size_A, unsigned int *max_row_size_B, unsigned int *scratchpad_offsets, unsigned int *scratchpad_indices, NumericT *scratchpad_values)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00365">spgemm.hpp:365</a></div></div>
<div class="ttc" id="tests_2src_2bisect_8cpp_html_a52b5d30a2d7b064678644a3bf49b7f6c"><div class="ttname"><a href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a></div><div class="ttdeci">float NumericT</div><div class="ttdef"><b>Definition:</b> <a href="tests_2src_2bisect_8cpp_source.html#l00040">bisect.cpp:40</a></div></div>
<div class="ttc" id="classviennacl_1_1context_html"><div class="ttname"><a href="classviennacl_1_1context.html">viennacl::context</a></div><div class="ttdoc">Represents a generic 'context' similar to an OpenCL context, but is backend-agnostic and thus also su...</div><div class="ttdef"><b>Definition:</b> <a href="context_8hpp_source.html#l00039">context.hpp:39</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_ac681c14ce17df9e05ee22df27353fcb1"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#ac681c14ce17df9e05ee22df27353fcb1">viennacl::linalg::cuda::round_to_next_power_of_2</a></div><div class="ttdeci">__device__ IndexT round_to_next_power_of_2(IndexT val)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00063">spgemm.hpp:63</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_af55c923f7bcf1fb5bef3ac98eee818d7"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#af55c923f7bcf1fb5bef3ac98eee818d7">viennacl::linalg::cuda::compressed_matrix_gemm_stage_1</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_stage_1(const IndexT *A_row_indices, const IndexT *A_col_indices, IndexT A_size1, const IndexT *B_row_indices, IndexT *subwarpsize_per_group, IndexT *max_nnz_row_A_per_group, IndexT *max_nnz_row_B_per_group)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00082">spgemm.hpp:82</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a5151284a26947b0204440ab90b1f4f4b"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a5151284a26947b0204440ab90b1f4f4b">viennacl::linalg::cuda::subwarp_minimum_shuffle</a></div><div class="ttdeci">__device__ IndexT subwarp_minimum_shuffle(IndexT min_index)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__rmerge_8hpp_source.html#l00143">spgemm_rmerge.hpp:143</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a80758e8edbc3622467f98d5a9bf2e826"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a80758e8edbc3622467f98d5a9bf2e826">viennacl::linalg::cuda::compressed_matrix_gemm_A2</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_A2(IndexT *A2_row_indices, IndexT *A2_col_indices, NumericT *A2_elements, IndexT A2_size1, IndexT *new_row_buffer)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00484">spgemm.hpp:484</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a466d08cbedce419dde70a2114660b15f"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a466d08cbedce419dde70a2114660b15f">viennacl::linalg::cuda::subwarp_accumulate_shared</a></div><div class="ttdeci">__device__ NumericT subwarp_accumulate_shared(NumericT output_value, unsigned int id_in_warp, NumericT *shared_buffer)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__rmerge_8hpp_source.html#l00241">spgemm_rmerge.hpp:241</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_a91f5145351151a66f916bdc3901206f2"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#a91f5145351151a66f916bdc3901206f2">viennacl::compressed_matrix::handle2</a></div><div class="ttdeci">const handle_type & handle2() const </div><div class="ttdoc">Returns the OpenCL handle to the column index array. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00938">compressed_matrix.hpp:938</a></div></div>
<div class="ttc" id="classviennacl_1_1vector_html"><div class="ttname"><a href="classviennacl_1_1vector.html">viennacl::vector< unsigned int ></a></div></div>
<div class="ttc" id="namespaceviennacl_html_a0a574e6cd04ca0e42298b4ab845700e4"><div class="ttname"><a href="namespaceviennacl.html#a0a574e6cd04ca0e42298b4ab845700e4">viennacl::row</a></div><div class="ttdeci">vector_expression< const matrix_base< NumericT, F >, const unsigned int, op_row > row(const matrix_base< NumericT, F > &A, unsigned int i)</div><div class="ttdef"><b>Definition:</b> <a href="matrix_8hpp_source.html#l00910">matrix.hpp:910</a></div></div>
<div class="ttc" id="sparse__matrix__operations__solve_8hpp_html"><div class="ttname"><a href="sparse__matrix__operations__solve_8hpp.html">sparse_matrix_operations_solve.hpp</a></div><div class="ttdoc">Implementations of direct triangular solvers for sparse matrices using CUDA. </div></div>
<div class="ttc" id="namespaceviennacl_html_a00b40450b6b2fd87aad3527d9b2084b8a427356f0fb1b8d32b28f37e36b272df4"><div class="ttname"><a href="namespaceviennacl.html#a00b40450b6b2fd87aad3527d9b2084b8a427356f0fb1b8d32b28f37e36b272df4">viennacl::MAIN_MEMORY</a></div><div class="ttdef"><b>Definition:</b> <a href="forwards_8h_source.html#l00348">forwards.h:348</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_aba08ea5f9d1c12dff5da58e0dff7a6fe"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#aba08ea5f9d1c12dff5da58e0dff7a6fe">viennacl::compressed_matrix::reserve</a></div><div class="ttdeci">void reserve(vcl_size_t new_nonzeros, bool preserve=true)</div><div class="ttdoc">Allocate memory for the supplied number of nonzeros in the matrix. Old values are preserved...</div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00794">compressed_matrix.hpp:794</a></div></div>
<div class="ttc" id="timer_8hpp_html"><div class="ttname"><a href="timer_8hpp.html">timer.hpp</a></div><div class="ttdoc">A simple, yet (mostly) sufficiently accurate timer for benchmarking and profiling. </div></div>
<div class="ttc" id="namespaceviennacl_1_1traits_html_a6707f5dab8f482170d2046a605f46ef8"><div class="ttname"><a href="namespaceviennacl_1_1traits.html#a6707f5dab8f482170d2046a605f46ef8">viennacl::traits::context</a></div><div class="ttdeci">viennacl::context context(T const &t)</div><div class="ttdoc">Returns an ID for the currently active memory domain of an object. </div><div class="ttdef"><b>Definition:</b> <a href="traits_2context_8hpp_source.html#l00040">context.hpp:40</a></div></div>
<div class="ttc" id="linalg_2cuda_2common_8hpp_html"><div class="ttname"><a href="linalg_2cuda_2common_8hpp.html">common.hpp</a></div><div class="ttdoc">Common routines for CUDA execution. </div></div>
<div class="ttc" id="vector_8hpp_html"><div class="ttname"><a href="vector_8hpp.html">vector.hpp</a></div><div class="ttdoc">The vector type with operator-overloads and proxy classes is defined here. Linear algebra operations ...</div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_html_adfd5b21910a692a78c547b22b9157c2e"><div class="ttname"><a href="namespaceviennacl_1_1linalg.html#adfd5b21910a692a78c547b22b9157c2e">viennacl::linalg::max</a></div><div class="ttdeci">NumericT max(std::vector< NumericT > const &v1)</div><div class="ttdef"><b>Definition:</b> <a href="maxmin_8hpp_source.html#l00047">maxmin.hpp:47</a></div></div>
<div class="ttc" id="classviennacl_1_1backend_1_1typesafe__host__array_html_ac31bfde381eb40e0349678252c444af9"><div class="ttname"><a href="classviennacl_1_1backend_1_1typesafe__host__array.html#ac31bfde381eb40e0349678252c444af9">viennacl::backend::typesafe_host_array::set</a></div><div class="ttdeci">void set(vcl_size_t index, U value)</div><div class="ttdef"><b>Definition:</b> <a href="backend_2util_8hpp_source.html#l00115">util.hpp:115</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a2ead3da9d15125a029d1ae747c35d915"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a2ead3da9d15125a029d1ae747c35d915">viennacl::linalg::cuda::compressed_matrix_gemm_G1</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_G1(IndexT *G1_row_indices, IndexT *G1_col_indices, NumericT *G1_elements, IndexT G1_size1, IndexT const *A_row_indices, IndexT const *A_col_indices, NumericT const *A_elements, IndexT A_size1, IndexT A_nnz, IndexT max_per_row, IndexT *new_row_buffer)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00511">spgemm.hpp:511</a></div></div>
<div class="ttc" id="classviennacl_1_1backend_1_1mem__handle_html_ac8373f0d899b89c843e14de4cb7a1c4a"><div class="ttname"><a href="classviennacl_1_1backend_1_1mem__handle.html#ac8373f0d899b89c843e14de4cb7a1c4a">viennacl::backend::mem_handle::raw_size</a></div><div class="ttdeci">vcl_size_t raw_size() const </div><div class="ttdoc">Returns the number of bytes of the currently active buffer. </div><div class="ttdef"><b>Definition:</b> <a href="mem__handle_8hpp_source.html#l00230">mem_handle.hpp:230</a></div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html">viennacl::compressed_matrix</a></div><div class="ttdoc">A sparse square matrix in compressed sparse rows format. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00559">compressed_matrix.hpp:559</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_html_a7f1b00757223e6feecf70a8657a9f096"><div class="ttname"><a href="namespaceviennacl_1_1linalg.html#a7f1b00757223e6feecf70a8657a9f096">viennacl::linalg::exclusive_scan</a></div><div class="ttdeci">void exclusive_scan(vector_base< NumericT > &vec1, vector_base< NumericT > &vec2)</div><div class="ttdoc">This function implements an exclusive scan. </div><div class="ttdef"><b>Definition:</b> <a href="vector__operations_8hpp_source.html#l01240">vector_operations.hpp:1240</a></div></div>
<div class="ttc" id="linalg_2cuda_2common_8hpp_html_acdb31f22f4d1e12f1c2a27d4c4aa6865"><div class="ttname"><a href="linalg_2cuda_2common_8hpp.html#acdb31f22f4d1e12f1c2a27d4c4aa6865">VIENNACL_CUDA_LAST_ERROR_CHECK</a></div><div class="ttdeci">#define VIENNACL_CUDA_LAST_ERROR_CHECK(message)</div><div class="ttdef"><b>Definition:</b> <a href="linalg_2cuda_2common_8hpp_source.html#l00030">common.hpp:30</a></div></div>
<div class="ttc" id="namespaceviennacl_html_ae7d5db0c2c91be75218db5b52c4d13da"><div class="ttname"><a href="namespaceviennacl.html#ae7d5db0c2c91be75218db5b52c4d13da">viennacl::cuda_arg</a></div><div class="ttdeci">NumericT * cuda_arg(scalar< NumericT > &obj)</div><div class="ttdoc">Convenience helper function for extracting the CUDA handle from a ViennaCL scalar. Non-const version. </div><div class="ttdef"><b>Definition:</b> <a href="linalg_2cuda_2common_8hpp_source.html#l00039">common.hpp:39</a></div></div>
<div class="ttc" id="scalar_8hpp_html"><div class="ttname"><a href="scalar_8hpp.html">scalar.hpp</a></div><div class="ttdoc">Implementation of the ViennaCL scalar class. </div></div>
<div class="ttc" id="classviennacl_1_1compressed__matrix_html_afcf49945a63836e1edfca536fcd191fb"><div class="ttname"><a href="classviennacl_1_1compressed__matrix.html#afcf49945a63836e1edfca536fcd191fb">viennacl::compressed_matrix::resize</a></div><div class="ttdeci">void resize(vcl_size_t new_size1, vcl_size_t new_size2, bool preserve=true)</div><div class="ttdoc">Resize the matrix. </div><div class="ttdef"><b>Definition:</b> <a href="compressed__matrix_8hpp_source.html#l00829">compressed_matrix.hpp:829</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_aea6aef114eb07f9ddf935c7441b232ca"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#aea6aef114eb07f9ddf935c7441b232ca">viennacl::linalg::cuda::subwarp_accumulate_shuffle</a></div><div class="ttdeci">__device__ NumericT subwarp_accumulate_shuffle(NumericT output_value)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__rmerge_8hpp_source.html#l00232">spgemm_rmerge.hpp:232</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a2526f31c8e9c88587938b341497e5666"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a2526f31c8e9c88587938b341497e5666">viennacl::linalg::cuda::compressed_matrix_gemm_stage_2</a></div><div class="ttdeci">__global__ void compressed_matrix_gemm_stage_2(const IndexT *A_row_indices, const IndexT *A_col_indices, IndexT A_size1, const IndexT *B_row_indices, const IndexT *B_col_indices, IndexT B_size2, IndexT *C_row_indices, unsigned int *subwarpsize_array, unsigned int *max_row_size_A, unsigned int *max_row_size_B, unsigned int *scratchpad_offsets, unsigned int *scratchpad_indices)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm_8hpp_source.html#l00217">spgemm.hpp:217</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_html_ae0a4445a6d0f1d75e7157cdc23239027"><div class="ttname"><a href="namespaceviennacl_1_1linalg.html#ae0a4445a6d0f1d75e7157cdc23239027">viennacl::linalg::min</a></div><div class="ttdeci">NumericT min(std::vector< NumericT > const &v1)</div><div class="ttdef"><b>Definition:</b> <a href="maxmin_8hpp_source.html#l00091">maxmin.hpp:91</a></div></div>
</div><!-- fragment --></div><!-- contents -->
</div><!-- doc-content -->
<!-- start footer part -->
<div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
<ul>
<li class="navelem"><a class="el" href="dir_c82e3d11dd171600f4a6e0cab1ec1e0d.html">viennacl</a></li><li class="navelem"><a class="el" href="dir_63cde087767c4ed65c7901ffc6e293fe.html">linalg</a></li><li class="navelem"><a class="el" href="dir_2aede027c3fab12899418e3db60d7e7a.html">cuda</a></li><li class="navelem"><a class="el" href="spgemm__rmerge_8hpp.html">spgemm_rmerge.hpp</a></li>
<li class="footer">Generated on Wed Jan 20 2016 22:32:41 for ViennaCL - The Vienna Computing Library by
<a href="http://www.doxygen.org/index.html">
<img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.8.6 </li>
</ul>
</div>
</body>
</html>
|
