1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
|
<!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/host_based/spgemm_vector.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__vector_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_vector.hpp</div> </div>
</div><!--header-->
<div class="contents">
<a href="spgemm__vector_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_HOST_BASED_SPGEMM_VECTOR_HPP_</span></div>
<div class="line"><a name="l00002"></a><span class="lineno"> 2</span> <span class="preprocessor">#define VIENNACL_LINALG_HOST_BASED_SPGEMM_VECTOR_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 "<a class="code" href="forwards_8h.html">viennacl/forwards.h</a>"</span></div>
<div class="line"><a name="l00026"></a><span class="lineno"> 26</span> <span class="preprocessor">#include "<a class="code" href="linalg_2host__based_2common_8hpp.html">viennacl/linalg/host_based/common.hpp</a>"</span></div>
<div class="line"><a name="l00027"></a><span class="lineno"> 27</span> </div>
<div class="line"><a name="l00028"></a><span class="lineno"> 28</span> </div>
<div class="line"><a name="l00029"></a><span class="lineno"> 29</span> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00030"></a><span class="lineno"> 30</span> <span class="preprocessor">#include "immintrin.h"</span></div>
<div class="line"><a name="l00031"></a><span class="lineno"> 31</span> <span class="preprocessor">#endif</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> </div>
<div class="line"><a name="l00034"></a><span class="lineno"> 34</span> <span class="keyword">namespace </span>viennacl</div>
<div class="line"><a name="l00035"></a><span class="lineno"> 35</span> {</div>
<div class="line"><a name="l00036"></a><span class="lineno"> 36</span> <span class="keyword">namespace </span>linalg</div>
<div class="line"><a name="l00037"></a><span class="lineno"> 37</span> {</div>
<div class="line"><a name="l00038"></a><span class="lineno"> 38</span> <span class="keyword">namespace </span>host_based</div>
<div class="line"><a name="l00039"></a><span class="lineno"> 39</span> {</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> </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> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00044"></a><span class="lineno"> 44</span> <span class="keyword">inline</span></div>
<div class="line"><a name="l00045"></a><span class="lineno"> 45</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_C_scan_symbolic_vector_AVX2(<span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B_begin, <span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B_end,</div>
<div class="line"><a name="l00046"></a><span class="lineno"> 46</span>  <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00047"></a><span class="lineno"> 47</span>  <span class="keywordtype">int</span> *row_C_vector_output)</div>
<div class="line"><a name="l00048"></a><span class="lineno"> 48</span> {</div>
<div class="line"><a name="l00049"></a><span class="lineno"> 49</span>  __m256i avx_all_ones = _mm256_set_epi32(1, 1, 1, 1, 1, 1, 1, 1);</div>
<div class="line"><a name="l00050"></a><span class="lineno"> 50</span>  __m256i avx_all_bsize2 = _mm256_set_epi32(B_size2, B_size2, B_size2, B_size2, B_size2, B_size2, B_size2, B_size2);</div>
<div class="line"><a name="l00051"></a><span class="lineno"> 51</span> </div>
<div class="line"><a name="l00052"></a><span class="lineno"> 52</span>  __m256i avx_row_indices_offsets = _mm256_set_epi32(7, 6, 5, 4, 3, 2, 1, 0);</div>
<div class="line"><a name="l00053"></a><span class="lineno"> 53</span>  __m256i avx_load_mask = _mm256_sub_epi32(avx_row_indices_offsets, _mm256_set1_epi32(row_indices_B_end - row_indices_B_begin));</div>
<div class="line"><a name="l00054"></a><span class="lineno"> 54</span>  __m256i avx_load_mask2 = avx_load_mask;</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>  __m256i avx_row_indices = _mm256_set1_epi32(0);</div>
<div class="line"><a name="l00057"></a><span class="lineno"> 57</span>  avx_row_indices = _mm256_mask_i32gather_epi32(avx_row_indices, row_indices_B_begin, avx_row_indices_offsets, avx_load_mask, 4);</div>
<div class="line"><a name="l00058"></a><span class="lineno"> 58</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00059"></a><span class="lineno"> 59</span>  __m256i avx_row_start = _mm256_mask_i32gather_epi32(avx_all_ones, B_row_buffer, avx_row_indices, avx_load_mask, 4);</div>
<div class="line"><a name="l00060"></a><span class="lineno"> 60</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00061"></a><span class="lineno"> 61</span>  __m256i avx_row_end = _mm256_mask_i32gather_epi32(avx_all_ones, B_row_buffer+1, avx_row_indices, avx_load_mask, 4);</div>
<div class="line"><a name="l00062"></a><span class="lineno"> 62</span> </div>
<div class="line"><a name="l00063"></a><span class="lineno"> 63</span>  avx_load_mask = _mm256_cmpgt_epi32(avx_row_end, avx_row_start);</div>
<div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  __m256i avx_index_front = avx_all_bsize2;</div>
<div class="line"><a name="l00065"></a><span class="lineno"> 65</span>  avx_index_front = _mm256_mask_i32gather_epi32(avx_index_front, B_col_buffer, avx_row_start, avx_load_mask, 4);</div>
<div class="line"><a name="l00066"></a><span class="lineno"> 66</span> </div>
<div class="line"><a name="l00067"></a><span class="lineno"> 67</span>  <span class="keywordtype">int</span> *output_ptr = row_C_vector_output;</div>
<div class="line"><a name="l00068"></a><span class="lineno"> 68</span> </div>
<div class="line"><a name="l00069"></a><span class="lineno"> 69</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00070"></a><span class="lineno"> 70</span>  {</div>
<div class="line"><a name="l00071"></a><span class="lineno"> 71</span>  <span class="comment">// get minimum index in current front:</span></div>
<div class="line"><a name="l00072"></a><span class="lineno"> 72</span>  __m256i avx_index_min1 = avx_index_front;</div>
<div class="line"><a name="l00073"></a><span class="lineno"> 73</span>  __m256i avx_temp = _mm256_permutevar8x32_epi32(avx_index_min1, _mm256_set_epi32(3, 2, 1, 0, 7, 6, 5, 4));</div>
<div class="line"><a name="l00074"></a><span class="lineno"> 74</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// first four elements compared against last four elements</span></div>
<div class="line"><a name="l00075"></a><span class="lineno"> 75</span> </div>
<div class="line"><a name="l00076"></a><span class="lineno"> 76</span>  avx_temp = _mm256_shuffle_epi32(avx_index_min1, <span class="keywordtype">int</span>(78)); <span class="comment">// 0b01001110 = 78, using shuffle instead of permutevar here because of lower latency</span></div>
<div class="line"><a name="l00077"></a><span class="lineno"> 77</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// first two elements compared against elements three and four (same for upper half of register)</span></div>
<div class="line"><a name="l00078"></a><span class="lineno"> 78</span> </div>
<div class="line"><a name="l00079"></a><span class="lineno"> 79</span>  avx_temp = _mm256_shuffle_epi32(avx_index_min1, <span class="keywordtype">int</span>(177)); <span class="comment">// 0b10110001 = 177, using shuffle instead of permutevar here because of lower latency</span></div>
<div class="line"><a name="l00080"></a><span class="lineno"> 80</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// now all entries of avx_index_min1 hold the minimum</span></div>
<div class="line"><a name="l00081"></a><span class="lineno"> 81</span> </div>
<div class="line"><a name="l00082"></a><span class="lineno"> 82</span>  <span class="keywordtype">int</span> min_index_in_front = ((<span class="keywordtype">int</span>*)&avx_index_min1)[0];</div>
<div class="line"><a name="l00083"></a><span class="lineno"> 83</span>  <span class="comment">// check for end of merge operation:</span></div>
<div class="line"><a name="l00084"></a><span class="lineno"> 84</span>  <span class="keywordflow">if</span> (min_index_in_front == B_size2)</div>
<div class="line"><a name="l00085"></a><span class="lineno"> 85</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00086"></a><span class="lineno"> 86</span> </div>
<div class="line"><a name="l00087"></a><span class="lineno"> 87</span>  <span class="comment">// write current entry:</span></div>
<div class="line"><a name="l00088"></a><span class="lineno"> 88</span>  *output_ptr = min_index_in_front;</div>
<div class="line"><a name="l00089"></a><span class="lineno"> 89</span>  ++output_ptr;</div>
<div class="line"><a name="l00090"></a><span class="lineno"> 90</span> </div>
<div class="line"><a name="l00091"></a><span class="lineno"> 91</span>  <span class="comment">// advance index front where equal to minimum index:</span></div>
<div class="line"><a name="l00092"></a><span class="lineno"> 92</span>  avx_load_mask = _mm256_cmpeq_epi32(avx_index_front, avx_index_min1);</div>
<div class="line"><a name="l00093"></a><span class="lineno"> 93</span>  <span class="comment">// first part: set index to B_size2 if equal to minimum index:</span></div>
<div class="line"><a name="l00094"></a><span class="lineno"> 94</span>  avx_temp = _mm256_and_si256(avx_all_bsize2, avx_load_mask);</div>
<div class="line"><a name="l00095"></a><span class="lineno"> 95</span>  avx_index_front = _mm256_max_epi32(avx_index_front, avx_temp);</div>
<div class="line"><a name="l00096"></a><span class="lineno"> 96</span>  <span class="comment">// second part: increment row_start registers where minimum found:</span></div>
<div class="line"><a name="l00097"></a><span class="lineno"> 97</span>  avx_temp = _mm256_and_si256(avx_all_ones, avx_load_mask); <span class="comment">//ones only where the minimum was found</span></div>
<div class="line"><a name="l00098"></a><span class="lineno"> 98</span>  avx_row_start = _mm256_add_epi32(avx_row_start, avx_temp);</div>
<div class="line"><a name="l00099"></a><span class="lineno"> 99</span>  <span class="comment">// third part part: load new data where more entries available:</span></div>
<div class="line"><a name="l00100"></a><span class="lineno"> 100</span>  avx_load_mask = _mm256_cmpgt_epi32(avx_row_end, avx_row_start);</div>
<div class="line"><a name="l00101"></a><span class="lineno"> 101</span>  avx_index_front = _mm256_mask_i32gather_epi32(avx_index_front, B_col_buffer, avx_row_start, avx_load_mask, 4);</div>
<div class="line"><a name="l00102"></a><span class="lineno"> 102</span>  }</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="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - row_C_vector_output);</div>
<div class="line"><a name="l00105"></a><span class="lineno"> 105</span> }</div>
<div class="line"><a name="l00106"></a><span class="lineno"> 106</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00107"></a><span class="lineno"> 107</span> </div>
<div class="line"><a name="l00112"></a><span class="lineno"> 112</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> IndexNum></div>
<div class="line"><a name="l00113"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a52dc2d661efa35953871e706c2fe0cba"> 113</a></span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a52dc2d661efa35953871e706c2fe0cba">row_C_scan_symbolic_vector_N</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B,</div>
<div class="line"><a name="l00114"></a><span class="lineno"> 114</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00115"></a><span class="lineno"> 115</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_C_vector_input, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_C_vector_input_end,</div>
<div class="line"><a name="l00116"></a><span class="lineno"> 116</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_output)</div>
<div class="line"><a name="l00117"></a><span class="lineno"> 117</span> {</div>
<div class="line"><a name="l00118"></a><span class="lineno"> 118</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_front[IndexNum+1];</div>
<div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *index_front_start[IndexNum+1];</div>
<div class="line"><a name="l00120"></a><span class="lineno"> 120</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *index_front_end[IndexNum+1];</div>
<div class="line"><a name="l00121"></a><span class="lineno"> 121</span> </div>
<div class="line"><a name="l00122"></a><span class="lineno"> 122</span>  <span class="comment">// Set up pointers for loading the indices:</span></div>
<div class="line"><a name="l00123"></a><span class="lineno"> 123</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<IndexNum; ++i, ++row_indices_B)</div>
<div class="line"><a name="l00124"></a><span class="lineno"> 124</span>  {</div>
<div class="line"><a name="l00125"></a><span class="lineno"> 125</span>  index_front_start[i] = B_col_buffer + B_row_buffer[*row_indices_B];</div>
<div class="line"><a name="l00126"></a><span class="lineno"> 126</span>  index_front_end[i] = B_col_buffer + B_row_buffer[*row_indices_B + 1];</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>  index_front_start[IndexNum] = row_C_vector_input;</div>
<div class="line"><a name="l00129"></a><span class="lineno"> 129</span>  index_front_end[IndexNum] = row_C_vector_input_end;</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>  <span class="comment">// load indices:</span></div>
<div class="line"><a name="l00132"></a><span class="lineno"> 132</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</div>
<div class="line"><a name="l00133"></a><span class="lineno"> 133</span>  index_front[i] = (index_front_start[i] < index_front_end[i]) ? *index_front_start[i] : B_size2;</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>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_ptr = row_C_vector_output;</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="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00138"></a><span class="lineno"> 138</span>  {</div>
<div class="line"><a name="l00139"></a><span class="lineno"> 139</span>  <span class="comment">// get minimum index in current front:</span></div>
<div class="line"><a name="l00140"></a><span class="lineno"> 140</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index_in_front = B_size2;</div>
<div class="line"><a name="l00141"></a><span class="lineno"> 141</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</div>
<div class="line"><a name="l00142"></a><span class="lineno"> 142</span>  min_index_in_front = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#aef015779a92be597e546c3491cebf6c1">std::min</a>(min_index_in_front, index_front[i]);</div>
<div class="line"><a name="l00143"></a><span class="lineno"> 143</span> </div>
<div class="line"><a name="l00144"></a><span class="lineno"> 144</span>  <span class="keywordflow">if</span> (min_index_in_front == B_size2) <span class="comment">// we're done</span></div>
<div class="line"><a name="l00145"></a><span class="lineno"> 145</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00146"></a><span class="lineno"> 146</span> </div>
<div class="line"><a name="l00147"></a><span class="lineno"> 147</span>  <span class="comment">// advance index front where equal to minimum index:</span></div>
<div class="line"><a name="l00148"></a><span class="lineno"> 148</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</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="keywordflow">if</span> (index_front[i] == min_index_in_front)</div>
<div class="line"><a name="l00151"></a><span class="lineno"> 151</span>  {</div>
<div class="line"><a name="l00152"></a><span class="lineno"> 152</span>  index_front_start[i] += 1;</div>
<div class="line"><a name="l00153"></a><span class="lineno"> 153</span>  index_front[i] = (index_front_start[i] < index_front_end[i]) ? *index_front_start[i] : B_size2;</div>
<div class="line"><a name="l00154"></a><span class="lineno"> 154</span>  }</div>
<div class="line"><a name="l00155"></a><span class="lineno"> 155</span>  }</div>
<div class="line"><a name="l00156"></a><span class="lineno"> 156</span> </div>
<div class="line"><a name="l00157"></a><span class="lineno"> 157</span>  <span class="comment">// write current entry:</span></div>
<div class="line"><a name="l00158"></a><span class="lineno"> 158</span>  *output_ptr = min_index_in_front;</div>
<div class="line"><a name="l00159"></a><span class="lineno"> 159</span>  ++output_ptr;</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> </div>
<div class="line"><a name="l00162"></a><span class="lineno"> 162</span>  <span class="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - row_C_vector_output);</div>
<div class="line"><a name="l00163"></a><span class="lineno"> 163</span> }</div>
<div class="line"><a name="l00164"></a><span class="lineno"> 164</span> </div>
<div class="line"><a name="l00165"></a><span class="lineno"><a class="line" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled.html#a4d2edbd6f51ca00712c5b3ea2bc64fa4"> 165</a></span> <span class="keyword">struct </span><a class="code" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled.html">spgemm_output_write_enabled</a> { <span class="keyword">static</span> <span class="keywordtype">void</span> <a class="code" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled.html#a4d2edbd6f51ca00712c5b3ea2bc64fa4">apply</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *ptr, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> value) { *ptr = value; } };</div>
<div class="line"><a name="l00166"></a><span class="lineno"><a class="line" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled.html#a4fdb2b5a1b4907488e6fc629e7a6195f"> 166</a></span> <span class="keyword">struct </span><a class="code" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled.html">spgemm_output_write_disabled</a> { <span class="keyword">static</span> <span class="keywordtype">void</span> <a class="code" href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled.html#a4fdb2b5a1b4907488e6fc629e7a6195f">apply</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> * , <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> ) { } };</div>
<div class="line"><a name="l00167"></a><span class="lineno"> 167</span> </div>
<div class="line"><a name="l00168"></a><span class="lineno"> 168</span> <span class="keyword">template</span><<span class="keyword">typename</span> OutputWriterT></div>
<div class="line"><a name="l00169"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a16889bdb3e98178a23e6848bacae8fb1"> 169</a></span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a16889bdb3e98178a23e6848bacae8fb1">row_C_scan_symbolic_vector_1</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input1_begin, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input1_end,</div>
<div class="line"><a name="l00170"></a><span class="lineno"> 170</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input2_begin, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input2_end,</div>
<div class="line"><a name="l00171"></a><span class="lineno"> 171</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> termination_index,</div>
<div class="line"><a name="l00172"></a><span class="lineno"> 172</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_begin)</div>
<div class="line"><a name="l00173"></a><span class="lineno"> 173</span> {</div>
<div class="line"><a name="l00174"></a><span class="lineno"> 174</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_ptr = output_begin;</div>
<div class="line"><a name="l00175"></a><span class="lineno"> 175</span> </div>
<div class="line"><a name="l00176"></a><span class="lineno"> 176</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> val_1 = (input1_begin < input1_end) ? *input1_begin : termination_index;</div>
<div class="line"><a name="l00177"></a><span class="lineno"> 177</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> val_2 = (input2_begin < input2_end) ? *input2_begin : termination_index;</div>
<div class="line"><a name="l00178"></a><span class="lineno"> 178</span>  <span class="keywordflow">while</span> (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="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#aef015779a92be597e546c3491cebf6c1">std::min</a>(val_1, val_2);</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="keywordflow">if</span> (min_index == termination_index)</div>
<div class="line"><a name="l00183"></a><span class="lineno"> 183</span>  <span class="keywordflow">break</span>;</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="keywordflow">if</span> (min_index == val_1)</div>
<div class="line"><a name="l00186"></a><span class="lineno"> 186</span>  {</div>
<div class="line"><a name="l00187"></a><span class="lineno"> 187</span>  ++input1_begin;</div>
<div class="line"><a name="l00188"></a><span class="lineno"> 188</span>  val_1 = (input1_begin < input1_end) ? *input1_begin : termination_index;</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> </div>
<div class="line"><a name="l00191"></a><span class="lineno"> 191</span>  <span class="keywordflow">if</span> (min_index == val_2)</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>  ++input2_begin;</div>
<div class="line"><a name="l00194"></a><span class="lineno"> 194</span>  val_2 = (input2_begin < input2_end) ? *input2_begin : termination_index;</div>
<div class="line"><a name="l00195"></a><span class="lineno"> 195</span>  }</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">// write current entry:</span></div>
<div class="line"><a name="l00198"></a><span class="lineno"> 198</span>  OutputWriterT::apply(output_ptr, min_index); <span class="comment">// *output_ptr = min_index; if necessary</span></div>
<div class="line"><a name="l00199"></a><span class="lineno"> 199</span>  ++output_ptr;</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> </div>
<div class="line"><a name="l00202"></a><span class="lineno"> 202</span>  <span class="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - output_begin);</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> </div>
<div class="line"><a name="l00205"></a><span class="lineno"> 205</span> <span class="keyword">inline</span></div>
<div class="line"><a name="l00206"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a258b78fbff5101b5c6703d1e8db07847"> 206</a></span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a258b78fbff5101b5c6703d1e8db07847">row_C_scan_symbolic_vector</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_start_A, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_end_A, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *A_col_buffer,</div>
<div class="line"><a name="l00207"></a><span class="lineno"> 207</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00208"></a><span class="lineno"> 208</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_1, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_2, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_3)</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>  <span class="comment">// Trivial case: row length 0:</span></div>
<div class="line"><a name="l00211"></a><span class="lineno"> 211</span>  <span class="keywordflow">if</span> (row_start_A == row_end_A)</div>
<div class="line"><a name="l00212"></a><span class="lineno"> 212</span>  <span class="keywordflow">return</span> 0;</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="comment">// Trivial case: row length 1:</span></div>
<div class="line"><a name="l00215"></a><span class="lineno"> 215</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A == 1)</div>
<div class="line"><a name="l00216"></a><span class="lineno"> 216</span>  {</div>
<div class="line"><a name="l00217"></a><span class="lineno"> 217</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00218"></a><span class="lineno"> 218</span>  <span class="keywordflow">return</span> B_row_buffer[A_col + 1] - B_row_buffer[A_col];</div>
<div class="line"><a name="l00219"></a><span class="lineno"> 219</span>  }</div>
<div class="line"><a name="l00220"></a><span class="lineno"> 220</span> </div>
<div class="line"><a name="l00221"></a><span class="lineno"> 221</span>  <span class="comment">// Optimizations for row length 2:</span></div>
<div class="line"><a name="l00222"></a><span class="lineno"> 222</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_C_len = 0;</div>
<div class="line"><a name="l00223"></a><span class="lineno"> 223</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A == 2)</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>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00226"></a><span class="lineno"> 226</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</div>
<div class="line"><a name="l00227"></a><span class="lineno"> 227</span>  <span class="keywordflow">return</span> row_C_scan_symbolic_vector_1<spgemm_output_write_disabled>(B_col_buffer + B_row_buffer[A_col_1], B_col_buffer + B_row_buffer[A_col_1 + 1],</div>
<div class="line"><a name="l00228"></a><span class="lineno"> 228</span>  B_col_buffer + B_row_buffer[A_col_2], B_col_buffer + B_row_buffer[A_col_2 + 1],</div>
<div class="line"><a name="l00229"></a><span class="lineno"> 229</span>  B_size2,</div>
<div class="line"><a name="l00230"></a><span class="lineno"> 230</span>  row_C_vector_1);</div>
<div class="line"><a name="l00231"></a><span class="lineno"> 231</span>  }</div>
<div class="line"><a name="l00232"></a><span class="lineno"> 232</span>  <span class="keywordflow">else</span> <span class="comment">// for more than two rows we can safely merge the first two:</span></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="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00235"></a><span class="lineno"> 235</span>  row_C_len = row_C_scan_symbolic_vector_AVX2((<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_start_A), (<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_end_A),</div>
<div class="line"><a name="l00236"></a><span class="lineno"> 236</span>  (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_row_buffer, (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_col_buffer, <span class="keywordtype">int</span>(B_size2),</div>
<div class="line"><a name="l00237"></a><span class="lineno"> 237</span>  (<span class="keywordtype">int</span>*)row_C_vector_1);</div>
<div class="line"><a name="l00238"></a><span class="lineno"> 238</span>  row_start_A += 8;</div>
<div class="line"><a name="l00239"></a><span class="lineno"> 239</span> <span class="preprocessor">#else</span></div>
<div class="line"><a name="l00240"></a><span class="lineno"> 240</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00241"></a><span class="lineno"> 241</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</div>
<div class="line"><a name="l00242"></a><span class="lineno"> 242</span>  row_C_len = row_C_scan_symbolic_vector_1<spgemm_output_write_enabled>(B_col_buffer + B_row_buffer[A_col_1], B_col_buffer + B_row_buffer[A_col_1 + 1],</div>
<div class="line"><a name="l00243"></a><span class="lineno"> 243</span>  B_col_buffer + B_row_buffer[A_col_2], B_col_buffer + B_row_buffer[A_col_2 + 1],</div>
<div class="line"><a name="l00244"></a><span class="lineno"> 244</span>  B_size2,</div>
<div class="line"><a name="l00245"></a><span class="lineno"> 245</span>  row_C_vector_1);</div>
<div class="line"><a name="l00246"></a><span class="lineno"> 246</span>  row_start_A += 2;</div>
<div class="line"><a name="l00247"></a><span class="lineno"> 247</span> <span class="preprocessor">#endif</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="comment">// all other row lengths:</span></div>
<div class="line"><a name="l00251"></a><span class="lineno"> 251</span>  <span class="keywordflow">while</span> (row_end_A > row_start_A)</div>
<div class="line"><a name="l00252"></a><span class="lineno"> 252</span>  {</div>
<div class="line"><a name="l00253"></a><span class="lineno"> 253</span> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00254"></a><span class="lineno"> 254</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A > 2) <span class="comment">// we deal with one or two remaining rows more efficiently below:</span></div>
<div class="line"><a name="l00255"></a><span class="lineno"> 255</span>  {</div>
<div class="line"><a name="l00256"></a><span class="lineno"> 256</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> merged_len = row_C_scan_symbolic_vector_AVX2((<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_start_A), (<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_end_A),</div>
<div class="line"><a name="l00257"></a><span class="lineno"> 257</span>  (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_row_buffer, (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_col_buffer, <span class="keywordtype">int</span>(B_size2),</div>
<div class="line"><a name="l00258"></a><span class="lineno"> 258</span>  (<span class="keywordtype">int</span>*)row_C_vector_3);</div>
<div class="line"><a name="l00259"></a><span class="lineno"> 259</span>  <span class="keywordflow">if</span> (row_start_A + 8 >= row_end_A)</div>
<div class="line"><a name="l00260"></a><span class="lineno"> 260</span>  row_C_len = row_C_scan_symbolic_vector_1<spgemm_output_write_disabled>(row_C_vector_3, row_C_vector_3 + merged_len,</div>
<div class="line"><a name="l00261"></a><span class="lineno"> 261</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00262"></a><span class="lineno"> 262</span>  B_size2,</div>
<div class="line"><a name="l00263"></a><span class="lineno"> 263</span>  row_C_vector_2);</div>
<div class="line"><a name="l00264"></a><span class="lineno"> 264</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00265"></a><span class="lineno"> 265</span>  row_C_len = row_C_scan_symbolic_vector_1<spgemm_output_write_enabled>(row_C_vector_3, row_C_vector_3 + merged_len,</div>
<div class="line"><a name="l00266"></a><span class="lineno"> 266</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00267"></a><span class="lineno"> 267</span>  B_size2,</div>
<div class="line"><a name="l00268"></a><span class="lineno"> 268</span>  row_C_vector_2);</div>
<div class="line"><a name="l00269"></a><span class="lineno"> 269</span>  row_start_A += 8;</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="keywordflow">else</span></div>
<div class="line"><a name="l00272"></a><span class="lineno"> 272</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00273"></a><span class="lineno"> 273</span>  <span class="keywordflow">if</span> (row_start_A == row_end_A - 1) <span class="comment">// last merge operation. No need to write output</span></div>
<div class="line"><a name="l00274"></a><span class="lineno"> 274</span>  {</div>
<div class="line"><a name="l00275"></a><span class="lineno"> 275</span>  <span class="comment">// process last row</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_index_B = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00277"></a><span class="lineno"> 277</span>  <span class="keywordflow">return</span> row_C_scan_symbolic_vector_1<spgemm_output_write_disabled>(B_col_buffer + B_row_buffer[row_index_B], B_col_buffer + B_row_buffer[row_index_B + 1],</div>
<div class="line"><a name="l00278"></a><span class="lineno"> 278</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00279"></a><span class="lineno"> 279</span>  B_size2,</div>
<div class="line"><a name="l00280"></a><span class="lineno"> 280</span>  row_C_vector_2);</div>
<div class="line"><a name="l00281"></a><span class="lineno"> 281</span>  }</div>
<div class="line"><a name="l00282"></a><span class="lineno"> 282</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (row_start_A + 1 < row_end_A)<span class="comment">// at least two more rows left, so merge them</span></div>
<div class="line"><a name="l00283"></a><span class="lineno"> 283</span>  {</div>
<div class="line"><a name="l00284"></a><span class="lineno"> 284</span>  <span class="comment">// process single row:</span></div>
<div class="line"><a name="l00285"></a><span class="lineno"> 285</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00286"></a><span class="lineno"> 286</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</div>
<div class="line"><a name="l00287"></a><span class="lineno"> 287</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> merged_len = row_C_scan_symbolic_vector_1<spgemm_output_write_enabled>(B_col_buffer + B_row_buffer[A_col_1], B_col_buffer + B_row_buffer[A_col_1 + 1],</div>
<div class="line"><a name="l00288"></a><span class="lineno"> 288</span>  B_col_buffer + B_row_buffer[A_col_2], B_col_buffer + B_row_buffer[A_col_2 + 1],</div>
<div class="line"><a name="l00289"></a><span class="lineno"> 289</span>  B_size2,</div>
<div class="line"><a name="l00290"></a><span class="lineno"> 290</span>  row_C_vector_3);</div>
<div class="line"><a name="l00291"></a><span class="lineno"> 291</span>  <span class="keywordflow">if</span> (row_start_A + 2 == row_end_A) <span class="comment">// last merge does not need a write:</span></div>
<div class="line"><a name="l00292"></a><span class="lineno"> 292</span>  <span class="keywordflow">return</span> row_C_scan_symbolic_vector_1<spgemm_output_write_disabled>(row_C_vector_3, row_C_vector_3 + merged_len,</div>
<div class="line"><a name="l00293"></a><span class="lineno"> 293</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00294"></a><span class="lineno"> 294</span>  B_size2,</div>
<div class="line"><a name="l00295"></a><span class="lineno"> 295</span>  row_C_vector_2);</div>
<div class="line"><a name="l00296"></a><span class="lineno"> 296</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00297"></a><span class="lineno"> 297</span>  row_C_len = row_C_scan_symbolic_vector_1<spgemm_output_write_enabled>(row_C_vector_3, row_C_vector_3 + merged_len,</div>
<div class="line"><a name="l00298"></a><span class="lineno"> 298</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00299"></a><span class="lineno"> 299</span>  B_size2,</div>
<div class="line"><a name="l00300"></a><span class="lineno"> 300</span>  row_C_vector_2);</div>
<div class="line"><a name="l00301"></a><span class="lineno"> 301</span>  row_start_A += 2;</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="keywordflow">else</span> <span class="comment">// at least two more rows left</span></div>
<div class="line"><a name="l00304"></a><span class="lineno"> 304</span>  {</div>
<div class="line"><a name="l00305"></a><span class="lineno"> 305</span>  <span class="comment">// process single row:</span></div>
<div class="line"><a name="l00306"></a><span class="lineno"> 306</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_index_B = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00307"></a><span class="lineno"> 307</span>  row_C_len = row_C_scan_symbolic_vector_1<spgemm_output_write_enabled>(B_col_buffer + B_row_buffer[row_index_B], B_col_buffer + B_row_buffer[row_index_B + 1],</div>
<div class="line"><a name="l00308"></a><span class="lineno"> 308</span>  row_C_vector_1, row_C_vector_1 + row_C_len,</div>
<div class="line"><a name="l00309"></a><span class="lineno"> 309</span>  B_size2,</div>
<div class="line"><a name="l00310"></a><span class="lineno"> 310</span>  row_C_vector_2);</div>
<div class="line"><a name="l00311"></a><span class="lineno"> 311</span>  ++row_start_A;</div>
<div class="line"><a name="l00312"></a><span class="lineno"> 312</span>  }</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>  <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a46cc5446cd00d687be1828a5bd661542">std::swap</a>(row_C_vector_1, row_C_vector_2);</div>
<div class="line"><a name="l00315"></a><span class="lineno"> 315</span>  }</div>
<div class="line"><a name="l00316"></a><span class="lineno"> 316</span> </div>
<div class="line"><a name="l00317"></a><span class="lineno"> 317</span>  <span class="keywordflow">return</span> row_C_len;</div>
<div class="line"><a name="l00318"></a><span class="lineno"> 318</span> }</div>
<div class="line"><a name="l00319"></a><span class="lineno"> 319</span> </div>
<div class="line"><a name="l00321"></a><span class="lineno"> 321</span> </div>
<div class="line"><a name="l00326"></a><span class="lineno"> 326</span> <span class="keyword">template</span><<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> IndexNum, <span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00327"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a7762639b80e4de0a154c36c281f5bdee"> 327</a></span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a7762639b80e4de0a154c36c281f5bdee">row_C_scan_numeric_vector_N</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *val_A,</div>
<div class="line"><a name="l00328"></a><span class="lineno"> 328</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *B_elements, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00329"></a><span class="lineno"> 329</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_C_vector_input, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *row_C_vector_input_end, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *row_C_vector_input_values,</div>
<div class="line"><a name="l00330"></a><span class="lineno"> 330</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_output, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *row_C_vector_output_values)</div>
<div class="line"><a name="l00331"></a><span class="lineno"> 331</span> {</div>
<div class="line"><a name="l00332"></a><span class="lineno"> 332</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index_front[IndexNum+1];</div>
<div class="line"><a name="l00333"></a><span class="lineno"> 333</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *index_front_start[IndexNum+1];</div>
<div class="line"><a name="l00334"></a><span class="lineno"> 334</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *index_front_end[IndexNum+1];</div>
<div class="line"><a name="l00335"></a><span class="lineno"> 335</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> * value_front_start[IndexNum+1];</div>
<div class="line"><a name="l00336"></a><span class="lineno"> 336</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> values_A[IndexNum+1];</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="comment">// Set up pointers for loading the indices:</span></div>
<div class="line"><a name="l00339"></a><span class="lineno"> 339</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<IndexNum; ++i, ++row_indices_B)</div>
<div class="line"><a name="l00340"></a><span class="lineno"> 340</span>  {</div>
<div class="line"><a name="l00341"></a><span class="lineno"> 341</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_B = *row_indices_B;</div>
<div class="line"><a name="l00342"></a><span class="lineno"> 342</span> </div>
<div class="line"><a name="l00343"></a><span class="lineno"> 343</span>  index_front_start[i] = B_col_buffer + B_row_buffer[row_B];</div>
<div class="line"><a name="l00344"></a><span class="lineno"> 344</span>  index_front_end[i] = B_col_buffer + B_row_buffer[row_B + 1];</div>
<div class="line"><a name="l00345"></a><span class="lineno"> 345</span>  value_front_start[i] = B_elements + B_row_buffer[row_B];</div>
<div class="line"><a name="l00346"></a><span class="lineno"> 346</span>  values_A[i] = val_A[i];</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>  index_front_start[IndexNum] = row_C_vector_input;</div>
<div class="line"><a name="l00349"></a><span class="lineno"> 349</span>  index_front_end[IndexNum] = row_C_vector_input_end;</div>
<div class="line"><a name="l00350"></a><span class="lineno"> 350</span>  value_front_start[IndexNum] = row_C_vector_input_values;</div>
<div class="line"><a name="l00351"></a><span class="lineno"> 351</span>  values_A[IndexNum] = <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1);</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>  <span class="comment">// load indices:</span></div>
<div class="line"><a name="l00354"></a><span class="lineno"> 354</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</div>
<div class="line"><a name="l00355"></a><span class="lineno"> 355</span>  index_front[i] = (index_front_start[i] < index_front_end[i]) ? *index_front_start[i] : B_size2;</div>
<div class="line"><a name="l00356"></a><span class="lineno"> 356</span> </div>
<div class="line"><a name="l00357"></a><span class="lineno"> 357</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_ptr = row_C_vector_output;</div>
<div class="line"><a name="l00358"></a><span class="lineno"> 358</span> </div>
<div class="line"><a name="l00359"></a><span class="lineno"> 359</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00360"></a><span class="lineno"> 360</span>  {</div>
<div class="line"><a name="l00361"></a><span class="lineno"> 361</span>  <span class="comment">// get minimum index in current front:</span></div>
<div class="line"><a name="l00362"></a><span class="lineno"> 362</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index_in_front = B_size2;</div>
<div class="line"><a name="l00363"></a><span class="lineno"> 363</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</div>
<div class="line"><a name="l00364"></a><span class="lineno"> 364</span>  min_index_in_front = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#aef015779a92be597e546c3491cebf6c1">std::min</a>(min_index_in_front, index_front[i]);</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>  <span class="keywordflow">if</span> (min_index_in_front == B_size2) <span class="comment">// we're done</span></div>
<div class="line"><a name="l00367"></a><span class="lineno"> 367</span>  <span class="keywordflow">break</span>;</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>  <span class="comment">// advance index front where equal to minimum index:</span></div>
<div class="line"><a name="l00370"></a><span class="lineno"> 370</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> row_C_value = 0;</div>
<div class="line"><a name="l00371"></a><span class="lineno"> 371</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<=IndexNum; ++i)</div>
<div class="line"><a name="l00372"></a><span class="lineno"> 372</span>  {</div>
<div class="line"><a name="l00373"></a><span class="lineno"> 373</span>  <span class="keywordflow">if</span> (index_front[i] == min_index_in_front)</div>
<div class="line"><a name="l00374"></a><span class="lineno"> 374</span>  {</div>
<div class="line"><a name="l00375"></a><span class="lineno"> 375</span>  index_front_start[i] += 1;</div>
<div class="line"><a name="l00376"></a><span class="lineno"> 376</span>  index_front[i] = (index_front_start[i] < index_front_end[i]) ? *index_front_start[i] : B_size2;</div>
<div class="line"><a name="l00377"></a><span class="lineno"> 377</span> </div>
<div class="line"><a name="l00378"></a><span class="lineno"> 378</span>  row_C_value += values_A[i] * *value_front_start[i];</div>
<div class="line"><a name="l00379"></a><span class="lineno"> 379</span>  value_front_start[i] += 1;</div>
<div class="line"><a name="l00380"></a><span class="lineno"> 380</span>  }</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> </div>
<div class="line"><a name="l00383"></a><span class="lineno"> 383</span>  <span class="comment">// write current entry:</span></div>
<div class="line"><a name="l00384"></a><span class="lineno"> 384</span>  *output_ptr = min_index_in_front;</div>
<div class="line"><a name="l00385"></a><span class="lineno"> 385</span>  ++output_ptr;</div>
<div class="line"><a name="l00386"></a><span class="lineno"> 386</span>  *row_C_vector_output_values = row_C_value;</div>
<div class="line"><a name="l00387"></a><span class="lineno"> 387</span>  ++row_C_vector_output_values;</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> </div>
<div class="line"><a name="l00390"></a><span class="lineno"> 390</span>  <span class="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - row_C_vector_output);</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> </div>
<div class="line"><a name="l00395"></a><span class="lineno"> 395</span> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00396"></a><span class="lineno"> 396</span> <span class="keyword">inline</span></div>
<div class="line"><a name="l00397"></a><span class="lineno"> 397</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_C_scan_numeric_vector_AVX2(<span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B_begin, <span class="keywordtype">int</span> <span class="keyword">const</span> *row_indices_B_end, <span class="keywordtype">double</span> <span class="keyword">const</span> *values_A,</div>
<div class="line"><a name="l00398"></a><span class="lineno"> 398</span>  <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <span class="keywordtype">double</span> <span class="keyword">const</span> *B_elements,</div>
<div class="line"><a name="l00399"></a><span class="lineno"> 399</span>  <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00400"></a><span class="lineno"> 400</span>  <span class="keywordtype">int</span> *row_C_vector_output, <span class="keywordtype">double</span> *row_C_vector_output_values)</div>
<div class="line"><a name="l00401"></a><span class="lineno"> 401</span> {</div>
<div class="line"><a name="l00402"></a><span class="lineno"> 402</span>  __m256i avx_all_ones = _mm256_set_epi32(1, 1, 1, 1, 1, 1, 1, 1);</div>
<div class="line"><a name="l00403"></a><span class="lineno"> 403</span>  __m256i avx_all_bsize2 = _mm256_set_epi32(B_size2, B_size2, B_size2, B_size2, B_size2, B_size2, B_size2, B_size2);</div>
<div class="line"><a name="l00404"></a><span class="lineno"> 404</span> </div>
<div class="line"><a name="l00405"></a><span class="lineno"> 405</span>  __m256i avx_row_indices_offsets = _mm256_set_epi32(7, 6, 5, 4, 3, 2, 1, 0);</div>
<div class="line"><a name="l00406"></a><span class="lineno"> 406</span>  __m256i avx_load_mask = _mm256_sub_epi32(avx_row_indices_offsets, _mm256_set1_epi32(row_indices_B_end - row_indices_B_begin));</div>
<div class="line"><a name="l00407"></a><span class="lineno"> 407</span>  __m256i avx_load_mask2 = avx_load_mask;</div>
<div class="line"><a name="l00408"></a><span class="lineno"> 408</span> </div>
<div class="line"><a name="l00409"></a><span class="lineno"> 409</span>  __m256i avx_row_indices = _mm256_set1_epi32(0);</div>
<div class="line"><a name="l00410"></a><span class="lineno"> 410</span>  avx_row_indices = _mm256_mask_i32gather_epi32(avx_row_indices, row_indices_B_begin, avx_row_indices_offsets, avx_load_mask, 4);</div>
<div class="line"><a name="l00411"></a><span class="lineno"> 411</span> </div>
<div class="line"><a name="l00412"></a><span class="lineno"> 412</span>  <span class="comment">// load values from A:</span></div>
<div class="line"><a name="l00413"></a><span class="lineno"> 413</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00414"></a><span class="lineno"> 414</span>  __m256d avx_value_A_low = _mm256_mask_i32gather_pd(_mm256_set_pd(0, 0, 0, 0), <span class="comment">//src</span></div>
<div class="line"><a name="l00415"></a><span class="lineno"> 415</span>  values_A, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00416"></a><span class="lineno"> 416</span>  _mm256_extractf128_si256(avx_row_indices_offsets, 0), <span class="comment">//indices</span></div>
<div class="line"><a name="l00417"></a><span class="lineno"> 417</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(3, 7, 2, 6, 1, 5, 0, 4)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00418"></a><span class="lineno"> 418</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00419"></a><span class="lineno"> 419</span>  __m256d avx_value_A_high = _mm256_mask_i32gather_pd(_mm256_set_pd(0, 0, 0, 0), <span class="comment">//src</span></div>
<div class="line"><a name="l00420"></a><span class="lineno"> 420</span>  values_A, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00421"></a><span class="lineno"> 421</span>  _mm256_extractf128_si256(avx_row_indices_offsets, 1), <span class="comment">//indices</span></div>
<div class="line"><a name="l00422"></a><span class="lineno"> 422</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00423"></a><span class="lineno"> 423</span> </div>
<div class="line"><a name="l00424"></a><span class="lineno"> 424</span> </div>
<div class="line"><a name="l00425"></a><span class="lineno"> 425</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00426"></a><span class="lineno"> 426</span>  __m256i avx_row_start = _mm256_mask_i32gather_epi32(avx_all_ones, B_row_buffer, avx_row_indices, avx_load_mask, 4);</div>
<div class="line"><a name="l00427"></a><span class="lineno"> 427</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00428"></a><span class="lineno"> 428</span>  __m256i avx_row_end = _mm256_mask_i32gather_epi32(avx_all_ones, B_row_buffer+1, avx_row_indices, avx_load_mask, 4);</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>  avx_load_mask = _mm256_cmpgt_epi32(avx_row_end, avx_row_start);</div>
<div class="line"><a name="l00431"></a><span class="lineno"> 431</span>  avx_load_mask2 = avx_load_mask;</div>
<div class="line"><a name="l00432"></a><span class="lineno"> 432</span>  __m256i avx_index_front = avx_all_bsize2;</div>
<div class="line"><a name="l00433"></a><span class="lineno"> 433</span>  avx_index_front = _mm256_mask_i32gather_epi32(avx_index_front, B_col_buffer, avx_row_start, avx_load_mask, 4);</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>  <span class="comment">// load front values from B:</span></div>
<div class="line"><a name="l00436"></a><span class="lineno"> 436</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00437"></a><span class="lineno"> 437</span>  __m256d avx_value_front_low = _mm256_mask_i32gather_pd(_mm256_set_pd(0, 0, 0, 0), <span class="comment">//src</span></div>
<div class="line"><a name="l00438"></a><span class="lineno"> 438</span>  B_elements, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00439"></a><span class="lineno"> 439</span>  _mm256_extractf128_si256(avx_row_start, 0), <span class="comment">//indices</span></div>
<div class="line"><a name="l00440"></a><span class="lineno"> 440</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(3, 7, 2, 6, 1, 5, 0, 4)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00441"></a><span class="lineno"> 441</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00442"></a><span class="lineno"> 442</span>  __m256d avx_value_front_high = _mm256_mask_i32gather_pd(_mm256_set_pd(0, 0, 0, 0), <span class="comment">//src</span></div>
<div class="line"><a name="l00443"></a><span class="lineno"> 443</span>  B_elements, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00444"></a><span class="lineno"> 444</span>  _mm256_extractf128_si256(avx_row_start, 1), <span class="comment">//indices</span></div>
<div class="line"><a name="l00445"></a><span class="lineno"> 445</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00446"></a><span class="lineno"> 446</span> </div>
<div class="line"><a name="l00447"></a><span class="lineno"> 447</span>  <span class="keywordtype">int</span> *output_ptr = row_C_vector_output;</div>
<div class="line"><a name="l00448"></a><span class="lineno"> 448</span> </div>
<div class="line"><a name="l00449"></a><span class="lineno"> 449</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00450"></a><span class="lineno"> 450</span>  {</div>
<div class="line"><a name="l00451"></a><span class="lineno"> 451</span>  <span class="comment">// get minimum index in current front:</span></div>
<div class="line"><a name="l00452"></a><span class="lineno"> 452</span>  __m256i avx_index_min1 = avx_index_front;</div>
<div class="line"><a name="l00453"></a><span class="lineno"> 453</span>  __m256i avx_temp = _mm256_permutevar8x32_epi32(avx_index_min1, _mm256_set_epi32(3, 2, 1, 0, 7, 6, 5, 4));</div>
<div class="line"><a name="l00454"></a><span class="lineno"> 454</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// first four elements compared against last four elements</span></div>
<div class="line"><a name="l00455"></a><span class="lineno"> 455</span> </div>
<div class="line"><a name="l00456"></a><span class="lineno"> 456</span>  avx_temp = _mm256_shuffle_epi32(avx_index_min1, <span class="keywordtype">int</span>(78)); <span class="comment">// 0b01001110 = 78, using shuffle instead of permutevar here because of lower latency</span></div>
<div class="line"><a name="l00457"></a><span class="lineno"> 457</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// first two elements compared against elements three and four (same for upper half of register)</span></div>
<div class="line"><a name="l00458"></a><span class="lineno"> 458</span> </div>
<div class="line"><a name="l00459"></a><span class="lineno"> 459</span>  avx_temp = _mm256_shuffle_epi32(avx_index_min1, <span class="keywordtype">int</span>(177)); <span class="comment">// 0b10110001 = 177, using shuffle instead of permutevar here because of lower latency</span></div>
<div class="line"><a name="l00460"></a><span class="lineno"> 460</span>  avx_index_min1 = _mm256_min_epi32(avx_index_min1, avx_temp); <span class="comment">// now all entries of avx_index_min1 hold the minimum</span></div>
<div class="line"><a name="l00461"></a><span class="lineno"> 461</span> </div>
<div class="line"><a name="l00462"></a><span class="lineno"> 462</span>  <span class="keywordtype">int</span> min_index_in_front = ((<span class="keywordtype">int</span>*)&avx_index_min1)[0];</div>
<div class="line"><a name="l00463"></a><span class="lineno"> 463</span>  <span class="comment">// check for end of merge operation:</span></div>
<div class="line"><a name="l00464"></a><span class="lineno"> 464</span>  <span class="keywordflow">if</span> (min_index_in_front == B_size2)</div>
<div class="line"><a name="l00465"></a><span class="lineno"> 465</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00466"></a><span class="lineno"> 466</span> </div>
<div class="line"><a name="l00467"></a><span class="lineno"> 467</span>  <span class="comment">// accumulate value (can certainly be done more elegantly...)</span></div>
<div class="line"><a name="l00468"></a><span class="lineno"> 468</span>  <span class="keywordtype">double</span> value = 0;</div>
<div class="line"><a name="l00469"></a><span class="lineno"> 469</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[0]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_low)[0] * ((<span class="keywordtype">double</span>*)&avx_value_A_low)[0] : 0;</div>
<div class="line"><a name="l00470"></a><span class="lineno"> 470</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[1]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_low)[1] * ((<span class="keywordtype">double</span>*)&avx_value_A_low)[1] : 0;</div>
<div class="line"><a name="l00471"></a><span class="lineno"> 471</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[2]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_low)[2] * ((<span class="keywordtype">double</span>*)&avx_value_A_low)[2] : 0;</div>
<div class="line"><a name="l00472"></a><span class="lineno"> 472</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[3]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_low)[3] * ((<span class="keywordtype">double</span>*)&avx_value_A_low)[3] : 0;</div>
<div class="line"><a name="l00473"></a><span class="lineno"> 473</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[4]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_high)[0] * ((<span class="keywordtype">double</span>*)&avx_value_A_high)[0] : 0;</div>
<div class="line"><a name="l00474"></a><span class="lineno"> 474</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[5]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_high)[1] * ((<span class="keywordtype">double</span>*)&avx_value_A_high)[1] : 0;</div>
<div class="line"><a name="l00475"></a><span class="lineno"> 475</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[6]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_high)[2] * ((<span class="keywordtype">double</span>*)&avx_value_A_high)[2] : 0;</div>
<div class="line"><a name="l00476"></a><span class="lineno"> 476</span>  value += (min_index_in_front == ((<span class="keywordtype">int</span>*)&avx_index_front)[7]) ? ((<span class="keywordtype">double</span>*)&avx_value_front_high)[3] * ((<span class="keywordtype">double</span>*)&avx_value_A_high)[3] : 0;</div>
<div class="line"><a name="l00477"></a><span class="lineno"> 477</span>  *row_C_vector_output_values = value;</div>
<div class="line"><a name="l00478"></a><span class="lineno"> 478</span>  ++row_C_vector_output_values;</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>  <span class="comment">// write current entry:</span></div>
<div class="line"><a name="l00481"></a><span class="lineno"> 481</span>  *output_ptr = min_index_in_front;</div>
<div class="line"><a name="l00482"></a><span class="lineno"> 482</span>  ++output_ptr;</div>
<div class="line"><a name="l00483"></a><span class="lineno"> 483</span> </div>
<div class="line"><a name="l00484"></a><span class="lineno"> 484</span>  <span class="comment">// advance index front where equal to minimum index:</span></div>
<div class="line"><a name="l00485"></a><span class="lineno"> 485</span>  avx_load_mask = _mm256_cmpeq_epi32(avx_index_front, avx_index_min1);</div>
<div class="line"><a name="l00486"></a><span class="lineno"> 486</span>  <span class="comment">// first part: set index to B_size2 if equal to minimum index:</span></div>
<div class="line"><a name="l00487"></a><span class="lineno"> 487</span>  avx_temp = _mm256_and_si256(avx_all_bsize2, avx_load_mask);</div>
<div class="line"><a name="l00488"></a><span class="lineno"> 488</span>  avx_index_front = _mm256_max_epi32(avx_index_front, avx_temp);</div>
<div class="line"><a name="l00489"></a><span class="lineno"> 489</span>  <span class="comment">// second part: increment row_start registers where minimum found:</span></div>
<div class="line"><a name="l00490"></a><span class="lineno"> 490</span>  avx_temp = _mm256_and_si256(avx_all_ones, avx_load_mask); <span class="comment">//ones only where the minimum was found</span></div>
<div class="line"><a name="l00491"></a><span class="lineno"> 491</span>  avx_row_start = _mm256_add_epi32(avx_row_start, avx_temp);</div>
<div class="line"><a name="l00492"></a><span class="lineno"> 492</span>  <span class="comment">// third part part: load new data where more entries available:</span></div>
<div class="line"><a name="l00493"></a><span class="lineno"> 493</span>  avx_load_mask = _mm256_cmpgt_epi32(avx_row_end, avx_row_start);</div>
<div class="line"><a name="l00494"></a><span class="lineno"> 494</span>  avx_load_mask2 = avx_load_mask;</div>
<div class="line"><a name="l00495"></a><span class="lineno"> 495</span>  avx_index_front = _mm256_mask_i32gather_epi32(avx_index_front, B_col_buffer, avx_row_start, avx_load_mask, 4);</div>
<div class="line"><a name="l00496"></a><span class="lineno"> 496</span> </div>
<div class="line"><a name="l00497"></a><span class="lineno"> 497</span>  <span class="comment">// load new values where necessary:</span></div>
<div class="line"><a name="l00498"></a><span class="lineno"> 498</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00499"></a><span class="lineno"> 499</span>  avx_value_front_low = _mm256_mask_i32gather_pd(avx_value_front_low, <span class="comment">//src</span></div>
<div class="line"><a name="l00500"></a><span class="lineno"> 500</span>  B_elements, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00501"></a><span class="lineno"> 501</span>  _mm256_extractf128_si256(avx_row_start, 0), <span class="comment">//indices</span></div>
<div class="line"><a name="l00502"></a><span class="lineno"> 502</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(3, 7, 2, 6, 1, 5, 0, 4)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00503"></a><span class="lineno"> 503</span> </div>
<div class="line"><a name="l00504"></a><span class="lineno"> 504</span>  avx_load_mask = avx_load_mask2; <span class="comment">// reload mask (destroyed by gather)</span></div>
<div class="line"><a name="l00505"></a><span class="lineno"> 505</span>  avx_value_front_high = _mm256_mask_i32gather_pd(avx_value_front_high, <span class="comment">//src</span></div>
<div class="line"><a name="l00506"></a><span class="lineno"> 506</span>  B_elements, <span class="comment">//base ptr</span></div>
<div class="line"><a name="l00507"></a><span class="lineno"> 507</span>  _mm256_extractf128_si256(avx_row_start, 1), <span class="comment">//indices</span></div>
<div class="line"><a name="l00508"></a><span class="lineno"> 508</span>  _mm256_permutevar8x32_epi32(avx_load_mask, _mm256_set_epi32(7, 3, 6, 2, 5, 1, 4, 0)), 8); <span class="comment">// mask</span></div>
<div class="line"><a name="l00509"></a><span class="lineno"> 509</span> </div>
<div class="line"><a name="l00510"></a><span class="lineno"> 510</span>  <span class="comment">//multiply new entries:</span></div>
<div class="line"><a name="l00511"></a><span class="lineno"> 511</span> </div>
<div class="line"><a name="l00512"></a><span class="lineno"> 512</span>  }</div>
<div class="line"><a name="l00513"></a><span class="lineno"> 513</span> </div>
<div class="line"><a name="l00514"></a><span class="lineno"> 514</span>  <span class="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - row_C_vector_output);</div>
<div class="line"><a name="l00515"></a><span class="lineno"> 515</span> }</div>
<div class="line"><a name="l00516"></a><span class="lineno"> 516</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00517"></a><span class="lineno"> 517</span> </div>
<div class="line"><a name="l00518"></a><span class="lineno"> 518</span> </div>
<div class="line"><a name="l00519"></a><span class="lineno"> 519</span> <span class="keyword">template</span><<span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00520"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2"> 520</a></span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input1_index_begin, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input1_index_end, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *input1_values_begin, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> factor1,</div>
<div class="line"><a name="l00521"></a><span class="lineno"> 521</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input2_index_begin, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *input2_index_end, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *input2_values_begin, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> factor2,</div>
<div class="line"><a name="l00522"></a><span class="lineno"> 522</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> termination_index,</div>
<div class="line"><a name="l00523"></a><span class="lineno"> 523</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_index_begin, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *output_values_begin)</div>
<div class="line"><a name="l00524"></a><span class="lineno"> 524</span> {</div>
<div class="line"><a name="l00525"></a><span class="lineno"> 525</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *output_ptr = output_index_begin;</div>
<div class="line"><a name="l00526"></a><span class="lineno"> 526</span> </div>
<div class="line"><a name="l00527"></a><span class="lineno"> 527</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index1 = (input1_index_begin < input1_index_end) ? *input1_index_begin : termination_index;</div>
<div class="line"><a name="l00528"></a><span class="lineno"> 528</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> index2 = (input2_index_begin < input2_index_end) ? *input2_index_begin : termination_index;</div>
<div class="line"><a name="l00529"></a><span class="lineno"> 529</span> </div>
<div class="line"><a name="l00530"></a><span class="lineno"> 530</span>  <span class="keywordflow">while</span> (1)</div>
<div class="line"><a name="l00531"></a><span class="lineno"> 531</span>  {</div>
<div class="line"><a name="l00532"></a><span class="lineno"> 532</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> min_index = <a class="code" href="namespaceviennacl_1_1linalg_1_1detail.html#aef015779a92be597e546c3491cebf6c1">std::min</a>(index1, index2);</div>
<div class="line"><a name="l00533"></a><span class="lineno"> 533</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> value = 0;</div>
<div class="line"><a name="l00534"></a><span class="lineno"> 534</span> </div>
<div class="line"><a name="l00535"></a><span class="lineno"> 535</span>  <span class="keywordflow">if</span> (min_index == termination_index)</div>
<div class="line"><a name="l00536"></a><span class="lineno"> 536</span>  <span class="keywordflow">break</span>;</div>
<div class="line"><a name="l00537"></a><span class="lineno"> 537</span> </div>
<div class="line"><a name="l00538"></a><span class="lineno"> 538</span>  <span class="keywordflow">if</span> (min_index == index1)</div>
<div class="line"><a name="l00539"></a><span class="lineno"> 539</span>  {</div>
<div class="line"><a name="l00540"></a><span class="lineno"> 540</span>  ++input1_index_begin;</div>
<div class="line"><a name="l00541"></a><span class="lineno"> 541</span>  index1 = (input1_index_begin < input1_index_end) ? *input1_index_begin : termination_index;</div>
<div class="line"><a name="l00542"></a><span class="lineno"> 542</span> </div>
<div class="line"><a name="l00543"></a><span class="lineno"> 543</span>  value += factor1 * *input1_values_begin;</div>
<div class="line"><a name="l00544"></a><span class="lineno"> 544</span>  ++input1_values_begin;</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="keywordflow">if</span> (min_index == index2)</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>  ++input2_index_begin;</div>
<div class="line"><a name="l00550"></a><span class="lineno"> 550</span>  index2 = (input2_index_begin < input2_index_end) ? *input2_index_begin : termination_index;</div>
<div class="line"><a name="l00551"></a><span class="lineno"> 551</span> </div>
<div class="line"><a name="l00552"></a><span class="lineno"> 552</span>  value += factor2 * *input2_values_begin;</div>
<div class="line"><a name="l00553"></a><span class="lineno"> 553</span>  ++input2_values_begin;</div>
<div class="line"><a name="l00554"></a><span class="lineno"> 554</span>  }</div>
<div class="line"><a name="l00555"></a><span class="lineno"> 555</span> </div>
<div class="line"><a name="l00556"></a><span class="lineno"> 556</span>  <span class="comment">// write current entry:</span></div>
<div class="line"><a name="l00557"></a><span class="lineno"> 557</span>  *output_ptr = min_index;</div>
<div class="line"><a name="l00558"></a><span class="lineno"> 558</span>  ++output_ptr;</div>
<div class="line"><a name="l00559"></a><span class="lineno"> 559</span>  *output_values_begin = value;</div>
<div class="line"><a name="l00560"></a><span class="lineno"> 560</span>  ++output_values_begin;</div>
<div class="line"><a name="l00561"></a><span class="lineno"> 561</span>  }</div>
<div class="line"><a name="l00562"></a><span class="lineno"> 562</span> </div>
<div class="line"><a name="l00563"></a><span class="lineno"> 563</span>  <span class="keywordflow">return</span> <span class="keyword">static_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span><span class="keyword">></span>(output_ptr - output_index_begin);</div>
<div class="line"><a name="l00564"></a><span class="lineno"> 564</span> }</div>
<div class="line"><a name="l00565"></a><span class="lineno"> 565</span> </div>
<div class="line"><a name="l00566"></a><span class="lineno"> 566</span> <span class="keyword">template</span><<span class="keyword">typename</span> NumericT></div>
<div class="line"><a name="l00567"></a><span class="lineno"><a class="line" href="namespaceviennacl_1_1linalg_1_1host__based.html#a4ce33d1460ff2f80bdc7026392cc13aa"> 567</a></span> <span class="keywordtype">void</span> <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a4ce33d1460ff2f80bdc7026392cc13aa">row_C_scan_numeric_vector</a>(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_start_A, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_end_A, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *A_col_buffer, <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="l00568"></a><span class="lineno"> 568</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_row_buffer, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">const</span> *B_col_buffer, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> <span class="keyword">const</span> *B_elements, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_size2,</div>
<div class="line"><a name="l00569"></a><span class="lineno"> 569</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_start_C, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_end_C, <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *C_col_buffer, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *C_elements,</div>
<div class="line"><a name="l00570"></a><span class="lineno"> 570</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_1, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *row_C_vector_1_values,</div>
<div class="line"><a name="l00571"></a><span class="lineno"> 571</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_2, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *row_C_vector_2_values,</div>
<div class="line"><a name="l00572"></a><span class="lineno"> 572</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *row_C_vector_3, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> *row_C_vector_3_values)</div>
<div class="line"><a name="l00573"></a><span class="lineno"> 573</span> {</div>
<div class="line"><a name="l00574"></a><span class="lineno"> 574</span>  (void)row_end_C;</div>
<div class="line"><a name="l00575"></a><span class="lineno"> 575</span> </div>
<div class="line"><a name="l00576"></a><span class="lineno"> 576</span>  <span class="comment">// Trivial case: row length 0:</span></div>
<div class="line"><a name="l00577"></a><span class="lineno"> 577</span>  <span class="keywordflow">if</span> (row_start_A == row_end_A)</div>
<div class="line"><a name="l00578"></a><span class="lineno"> 578</span>  <span class="keywordflow">return</span>;</div>
<div class="line"><a name="l00579"></a><span class="lineno"> 579</span> </div>
<div class="line"><a name="l00580"></a><span class="lineno"> 580</span>  <span class="comment">// Trivial case: row length 1:</span></div>
<div class="line"><a name="l00581"></a><span class="lineno"> 581</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A == 1)</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>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00584"></a><span class="lineno"> 584</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_end = B_row_buffer[A_col + 1];</div>
<div class="line"><a name="l00585"></a><span class="lineno"> 585</span>  <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a> A_value = A_elements[row_start_A];</div>
<div class="line"><a name="l00586"></a><span class="lineno"> 586</span>  C_col_buffer += row_start_C;</div>
<div class="line"><a name="l00587"></a><span class="lineno"> 587</span>  C_elements += row_start_C;</div>
<div class="line"><a name="l00588"></a><span class="lineno"> 588</span>  <span class="keywordflow">for</span> (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> j = B_row_buffer[A_col]; j < B_end; ++j, ++C_col_buffer, ++C_elements)</div>
<div class="line"><a name="l00589"></a><span class="lineno"> 589</span>  {</div>
<div class="line"><a name="l00590"></a><span class="lineno"> 590</span>  *C_col_buffer = B_col_buffer[j];</div>
<div class="line"><a name="l00591"></a><span class="lineno"> 591</span>  *C_elements = A_value * B_elements[j];</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>  <span class="keywordflow">return</span>;</div>
<div class="line"><a name="l00594"></a><span class="lineno"> 594</span>  }</div>
<div class="line"><a name="l00595"></a><span class="lineno"> 595</span> </div>
<div class="line"><a name="l00596"></a><span class="lineno"> 596</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> row_C_len = 0;</div>
<div class="line"><a name="l00597"></a><span class="lineno"> 597</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A == 2) <span class="comment">// directly merge to C:</span></div>
<div class="line"><a name="l00598"></a><span class="lineno"> 598</span>  {</div>
<div class="line"><a name="l00599"></a><span class="lineno"> 599</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00600"></a><span class="lineno"> 600</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</div>
<div class="line"><a name="l00601"></a><span class="lineno"> 601</span> </div>
<div class="line"><a name="l00602"></a><span class="lineno"> 602</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_1 = B_row_buffer[A_col_1];</div>
<div class="line"><a name="l00603"></a><span class="lineno"> 603</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_2 = B_row_buffer[A_col_2];</div>
<div class="line"><a name="l00604"></a><span class="lineno"> 604</span> </div>
<div class="line"><a name="l00605"></a><span class="lineno"> 605</span>  <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(B_col_buffer + B_offset_1, B_col_buffer + B_row_buffer[A_col_1+1], B_elements + B_offset_1, A_elements[row_start_A],</div>
<div class="line"><a name="l00606"></a><span class="lineno"> 606</span>  B_col_buffer + B_offset_2, B_col_buffer + B_row_buffer[A_col_2+1], B_elements + B_offset_2, A_elements[row_start_A + 1],</div>
<div class="line"><a name="l00607"></a><span class="lineno"> 607</span>  B_size2,</div>
<div class="line"><a name="l00608"></a><span class="lineno"> 608</span>  C_col_buffer + row_start_C, C_elements + row_start_C);</div>
<div class="line"><a name="l00609"></a><span class="lineno"> 609</span>  <span class="keywordflow">return</span>;</div>
<div class="line"><a name="l00610"></a><span class="lineno"> 610</span>  }</div>
<div class="line"><a name="l00611"></a><span class="lineno"> 611</span> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00612"></a><span class="lineno"> 612</span>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (row_end_A - row_start_A > 10) <span class="comment">// safely merge eight rows into temporary buffer:</span></div>
<div class="line"><a name="l00613"></a><span class="lineno"> 613</span>  {</div>
<div class="line"><a name="l00614"></a><span class="lineno"> 614</span>  row_C_len = row_C_scan_numeric_vector_AVX2((<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_start_A), (<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_end_A), A_elements + row_start_A,</div>
<div class="line"><a name="l00615"></a><span class="lineno"> 615</span>  (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_row_buffer, (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_col_buffer, B_elements, <span class="keywordtype">int</span>(B_size2),</div>
<div class="line"><a name="l00616"></a><span class="lineno"> 616</span>  (<span class="keywordtype">int</span>*)row_C_vector_1, row_C_vector_1_values);</div>
<div class="line"><a name="l00617"></a><span class="lineno"> 617</span>  row_start_A += 8;</div>
<div class="line"><a name="l00618"></a><span class="lineno"> 618</span>  }</div>
<div class="line"><a name="l00619"></a><span class="lineno"> 619</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00620"></a><span class="lineno"> 620</span>  <span class="keywordflow">else</span> <span class="comment">// safely merge two rows into temporary buffer:</span></div>
<div class="line"><a name="l00621"></a><span class="lineno"> 621</span>  {</div>
<div class="line"><a name="l00622"></a><span class="lineno"> 622</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00623"></a><span class="lineno"> 623</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</div>
<div class="line"><a name="l00624"></a><span class="lineno"> 624</span> </div>
<div class="line"><a name="l00625"></a><span class="lineno"> 625</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_1 = B_row_buffer[A_col_1];</div>
<div class="line"><a name="l00626"></a><span class="lineno"> 626</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_2 = B_row_buffer[A_col_2];</div>
<div class="line"><a name="l00627"></a><span class="lineno"> 627</span> </div>
<div class="line"><a name="l00628"></a><span class="lineno"> 628</span>  row_C_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(B_col_buffer + B_offset_1, B_col_buffer + B_row_buffer[A_col_1+1], B_elements + B_offset_1, A_elements[row_start_A],</div>
<div class="line"><a name="l00629"></a><span class="lineno"> 629</span>  B_col_buffer + B_offset_2, B_col_buffer + B_row_buffer[A_col_2+1], B_elements + B_offset_2, A_elements[row_start_A + 1],</div>
<div class="line"><a name="l00630"></a><span class="lineno"> 630</span>  B_size2,</div>
<div class="line"><a name="l00631"></a><span class="lineno"> 631</span>  row_C_vector_1, row_C_vector_1_values);</div>
<div class="line"><a name="l00632"></a><span class="lineno"> 632</span>  row_start_A += 2;</div>
<div class="line"><a name="l00633"></a><span class="lineno"> 633</span>  }</div>
<div class="line"><a name="l00634"></a><span class="lineno"> 634</span> </div>
<div class="line"><a name="l00635"></a><span class="lineno"> 635</span>  <span class="comment">// process remaining rows:</span></div>
<div class="line"><a name="l00636"></a><span class="lineno"> 636</span>  <span class="keywordflow">while</span> (row_end_A > row_start_A)</div>
<div class="line"><a name="l00637"></a><span class="lineno"> 637</span>  {</div>
<div class="line"><a name="l00638"></a><span class="lineno"> 638</span> <span class="preprocessor">#ifdef VIENNACL_WITH_AVX2</span></div>
<div class="line"><a name="l00639"></a><span class="lineno"> 639</span>  <span class="keywordflow">if</span> (row_end_A - row_start_A > 9) <span class="comment">// code in other if-conditionals ensures that values get written to C</span></div>
<div class="line"><a name="l00640"></a><span class="lineno"> 640</span>  {</div>
<div class="line"><a name="l00641"></a><span class="lineno"> 641</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> merged_len = row_C_scan_numeric_vector_AVX2((<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_start_A), (<span class="keyword">const</span> <span class="keywordtype">int</span>*)(A_col_buffer + row_end_A), A_elements + row_start_A,</div>
<div class="line"><a name="l00642"></a><span class="lineno"> 642</span>  (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_row_buffer, (<span class="keyword">const</span> <span class="keywordtype">int</span>*)B_col_buffer, B_elements, <span class="keywordtype">int</span>(B_size2),</div>
<div class="line"><a name="l00643"></a><span class="lineno"> 643</span>  (<span class="keywordtype">int</span>*)row_C_vector_3, row_C_vector_3_values);</div>
<div class="line"><a name="l00644"></a><span class="lineno"> 644</span>  row_C_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(row_C_vector_3, row_C_vector_3 + merged_len, row_C_vector_3_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00645"></a><span class="lineno"> 645</span>  row_C_vector_1, row_C_vector_1 + row_C_len, row_C_vector_1_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00646"></a><span class="lineno"> 646</span>  B_size2,</div>
<div class="line"><a name="l00647"></a><span class="lineno"> 647</span>  row_C_vector_2, row_C_vector_2_values);</div>
<div class="line"><a name="l00648"></a><span class="lineno"> 648</span>  row_start_A += 8;</div>
<div class="line"><a name="l00649"></a><span class="lineno"> 649</span>  }</div>
<div class="line"><a name="l00650"></a><span class="lineno"> 650</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00651"></a><span class="lineno"> 651</span> <span class="preprocessor">#endif</span></div>
<div class="line"><a name="l00652"></a><span class="lineno"> 652</span>  <span class="keywordflow">if</span> (row_start_A + 1 == row_end_A) <span class="comment">// last row to merge, write directly to C:</span></div>
<div class="line"><a name="l00653"></a><span class="lineno"> 653</span>  {</div>
<div class="line"><a name="l00654"></a><span class="lineno"> 654</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00655"></a><span class="lineno"> 655</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset = B_row_buffer[A_col];</div>
<div class="line"><a name="l00656"></a><span class="lineno"> 656</span> </div>
<div class="line"><a name="l00657"></a><span class="lineno"> 657</span>  row_C_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(B_col_buffer + B_offset, B_col_buffer + B_row_buffer[A_col+1], B_elements + B_offset, A_elements[row_start_A],</div>
<div class="line"><a name="l00658"></a><span class="lineno"> 658</span>  row_C_vector_1, row_C_vector_1 + row_C_len, row_C_vector_1_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00659"></a><span class="lineno"> 659</span>  B_size2,</div>
<div class="line"><a name="l00660"></a><span class="lineno"> 660</span>  C_col_buffer + row_start_C, C_elements + row_start_C);</div>
<div class="line"><a name="l00661"></a><span class="lineno"> 661</span>  <span class="keywordflow">return</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>  <span class="keywordflow">else</span> <span class="keywordflow">if</span> (row_start_A + 2 < row_end_A)<span class="comment">// at least three more rows left, so merge two</span></div>
<div class="line"><a name="l00664"></a><span class="lineno"> 664</span>  {</div>
<div class="line"><a name="l00665"></a><span class="lineno"> 665</span>  <span class="comment">// process single row:</span></div>
<div class="line"><a name="l00666"></a><span class="lineno"> 666</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_1 = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00667"></a><span class="lineno"> 667</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col_2 = A_col_buffer[row_start_A + 1];</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="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_1 = B_row_buffer[A_col_1];</div>
<div class="line"><a name="l00670"></a><span class="lineno"> 670</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset_2 = B_row_buffer[A_col_2];</div>
<div class="line"><a name="l00671"></a><span class="lineno"> 671</span> </div>
<div class="line"><a name="l00672"></a><span class="lineno"> 672</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> merged_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(B_col_buffer + B_offset_1, B_col_buffer + B_row_buffer[A_col_1+1], B_elements + B_offset_1, A_elements[row_start_A],</div>
<div class="line"><a name="l00673"></a><span class="lineno"> 673</span>  B_col_buffer + B_offset_2, B_col_buffer + B_row_buffer[A_col_2+1], B_elements + B_offset_2, A_elements[row_start_A + 1],</div>
<div class="line"><a name="l00674"></a><span class="lineno"> 674</span>  B_size2,</div>
<div class="line"><a name="l00675"></a><span class="lineno"> 675</span>  row_C_vector_3, row_C_vector_3_values);</div>
<div class="line"><a name="l00676"></a><span class="lineno"> 676</span>  row_C_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(row_C_vector_3, row_C_vector_3 + merged_len, row_C_vector_3_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00677"></a><span class="lineno"> 677</span>  row_C_vector_1, row_C_vector_1 + row_C_len, row_C_vector_1_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00678"></a><span class="lineno"> 678</span>  B_size2,</div>
<div class="line"><a name="l00679"></a><span class="lineno"> 679</span>  row_C_vector_2, row_C_vector_2_values);</div>
<div class="line"><a name="l00680"></a><span class="lineno"> 680</span>  row_start_A += 2;</div>
<div class="line"><a name="l00681"></a><span class="lineno"> 681</span>  }</div>
<div class="line"><a name="l00682"></a><span class="lineno"> 682</span>  <span class="keywordflow">else</span></div>
<div class="line"><a name="l00683"></a><span class="lineno"> 683</span>  {</div>
<div class="line"><a name="l00684"></a><span class="lineno"> 684</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> A_col = A_col_buffer[row_start_A];</div>
<div class="line"><a name="l00685"></a><span class="lineno"> 685</span>  <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> B_offset = B_row_buffer[A_col];</div>
<div class="line"><a name="l00686"></a><span class="lineno"> 686</span> </div>
<div class="line"><a name="l00687"></a><span class="lineno"> 687</span>  row_C_len = <a class="code" href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">row_C_scan_numeric_vector_1</a>(B_col_buffer + B_offset, B_col_buffer + B_row_buffer[A_col+1], B_elements + B_offset, A_elements[row_start_A],</div>
<div class="line"><a name="l00688"></a><span class="lineno"> 688</span>  row_C_vector_1, row_C_vector_1 + row_C_len, row_C_vector_1_values, <a class="code" href="tests_2src_2bisect_8cpp.html#a52b5d30a2d7b064678644a3bf49b7f6c">NumericT</a>(1.0),</div>
<div class="line"><a name="l00689"></a><span class="lineno"> 689</span>  B_size2,</div>
<div class="line"><a name="l00690"></a><span class="lineno"> 690</span>  row_C_vector_2, row_C_vector_2_values);</div>
<div class="line"><a name="l00691"></a><span class="lineno"> 691</span>  ++row_start_A;</div>
<div class="line"><a name="l00692"></a><span class="lineno"> 692</span>  }</div>
<div class="line"><a name="l00693"></a><span class="lineno"> 693</span> </div>
<div class="line"><a name="l00694"></a><span class="lineno"> 694</span>  <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a46cc5446cd00d687be1828a5bd661542">std::swap</a>(row_C_vector_1, row_C_vector_2);</div>
<div class="line"><a name="l00695"></a><span class="lineno"> 695</span>  <a class="code" href="namespaceviennacl_1_1linalg_1_1cuda.html#a46cc5446cd00d687be1828a5bd661542">std::swap</a>(row_C_vector_1_values, row_C_vector_2_values);</div>
<div class="line"><a name="l00696"></a><span class="lineno"> 696</span>  }</div>
<div class="line"><a name="l00697"></a><span class="lineno"> 697</span> }</div>
<div class="line"><a name="l00698"></a><span class="lineno"> 698</span> </div>
<div class="line"><a name="l00699"></a><span class="lineno"> 699</span> </div>
<div class="line"><a name="l00700"></a><span class="lineno"> 700</span> } <span class="comment">// namespace host_based</span></div>
<div class="line"><a name="l00701"></a><span class="lineno"> 701</span> } <span class="comment">//namespace linalg</span></div>
<div class="line"><a name="l00702"></a><span class="lineno"> 702</span> } <span class="comment">//namespace viennacl</span></div>
<div class="line"><a name="l00703"></a><span class="lineno"> 703</span> </div>
<div class="line"><a name="l00704"></a><span class="lineno"> 704</span> </div>
<div class="line"><a name="l00705"></a><span class="lineno"> 705</span> <span class="preprocessor">#endif</span></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1host__based_html_a9e8dab5478aee178b364387de085efb2"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a9e8dab5478aee178b364387de085efb2">viennacl::linalg::host_based::row_C_scan_numeric_vector_1</a></div><div class="ttdeci">unsigned int row_C_scan_numeric_vector_1(unsigned int const *input1_index_begin, unsigned int const *input1_index_end, NumericT const *input1_values_begin, NumericT factor1, unsigned int const *input2_index_begin, unsigned int const *input2_index_end, NumericT const *input2_values_begin, NumericT factor2, unsigned int termination_index, unsigned int *output_index_begin, NumericT *output_values_begin)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00520">spgemm_vector.hpp:520</a></div></div>
<div class="ttc" id="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled_html"><div class="ttname"><a href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled.html">viennacl::linalg::host_based::spgemm_output_write_enabled</a></div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00165">spgemm_vector.hpp:165</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="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled_html"><div class="ttname"><a href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled.html">viennacl::linalg::host_based::spgemm_output_write_disabled</a></div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00166">spgemm_vector.hpp:166</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="namespaceviennacl_1_1linalg_1_1host__based_html_a16889bdb3e98178a23e6848bacae8fb1"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a16889bdb3e98178a23e6848bacae8fb1">viennacl::linalg::host_based::row_C_scan_symbolic_vector_1</a></div><div class="ttdeci">unsigned int row_C_scan_symbolic_vector_1(unsigned int const *input1_begin, unsigned int const *input1_end, unsigned int const *input2_begin, unsigned int const *input2_end, unsigned int termination_index, unsigned int *output_begin)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00169">spgemm_vector.hpp:169</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1host__based_html_a4ce33d1460ff2f80bdc7026392cc13aa"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a4ce33d1460ff2f80bdc7026392cc13aa">viennacl::linalg::host_based::row_C_scan_numeric_vector</a></div><div class="ttdeci">void row_C_scan_numeric_vector(unsigned int row_start_A, unsigned int row_end_A, unsigned int const *A_col_buffer, NumericT const *A_elements, unsigned int const *B_row_buffer, unsigned int const *B_col_buffer, NumericT const *B_elements, unsigned int B_size2, unsigned int row_start_C, unsigned int row_end_C, unsigned int *C_col_buffer, NumericT *C_elements, unsigned int *row_C_vector_1, NumericT *row_C_vector_1_values, unsigned int *row_C_vector_2, NumericT *row_C_vector_2_values, unsigned int *row_C_vector_3, NumericT *row_C_vector_3_values)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00567">spgemm_vector.hpp:567</a></div></div>
<div class="ttc" id="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled_html_a4fdb2b5a1b4907488e6fc629e7a6195f"><div class="ttname"><a href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__disabled.html#a4fdb2b5a1b4907488e6fc629e7a6195f">viennacl::linalg::host_based::spgemm_output_write_disabled::apply</a></div><div class="ttdeci">static void apply(unsigned int *, unsigned int)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00166">spgemm_vector.hpp:166</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1host__based_html_a7762639b80e4de0a154c36c281f5bdee"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a7762639b80e4de0a154c36c281f5bdee">viennacl::linalg::host_based::row_C_scan_numeric_vector_N</a></div><div class="ttdeci">unsigned int row_C_scan_numeric_vector_N(unsigned int const *row_indices_B, NumericT const *val_A, unsigned int const *B_row_buffer, unsigned int const *B_col_buffer, NumericT const *B_elements, unsigned int B_size2, unsigned int const *row_C_vector_input, unsigned int const *row_C_vector_input_end, NumericT *row_C_vector_input_values, unsigned int *row_C_vector_output, NumericT *row_C_vector_output_values)</div><div class="ttdoc">Merges up to IndexNum rows from B into the result buffer. </div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00327">spgemm_vector.hpp:327</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1host__based_html_a258b78fbff5101b5c6703d1e8db07847"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a258b78fbff5101b5c6703d1e8db07847">viennacl::linalg::host_based::row_C_scan_symbolic_vector</a></div><div class="ttdeci">unsigned int row_C_scan_symbolic_vector(unsigned int row_start_A, unsigned int row_end_A, unsigned int const *A_col_buffer, unsigned int const *B_row_buffer, unsigned int const *B_col_buffer, unsigned int B_size2, unsigned int *row_C_vector_1, unsigned int *row_C_vector_2, unsigned int *row_C_vector_3)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00206">spgemm_vector.hpp:206</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1host__based_html_a52dc2d661efa35953871e706c2fe0cba"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1host__based.html#a52dc2d661efa35953871e706c2fe0cba">viennacl::linalg::host_based::row_C_scan_symbolic_vector_N</a></div><div class="ttdeci">unsigned int row_C_scan_symbolic_vector_N(unsigned int const *row_indices_B, unsigned int const *B_row_buffer, unsigned int const *B_col_buffer, unsigned int B_size2, unsigned int const *row_C_vector_input, unsigned int const *row_C_vector_input_end, unsigned int *row_C_vector_output)</div><div class="ttdoc">Merges up to IndexNum rows from B into the result buffer. </div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00113">spgemm_vector.hpp:113</a></div></div>
<div class="ttc" id="linalg_2host__based_2common_8hpp_html"><div class="ttname"><a href="linalg_2host__based_2common_8hpp.html">common.hpp</a></div><div class="ttdoc">Common routines for single-threaded or OpenMP-enabled execution on CPU. </div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1cuda_html_a46cc5446cd00d687be1828a5bd661542"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1cuda.html#a46cc5446cd00d687be1828a5bd661542">viennacl::linalg::cuda::swap</a></div><div class="ttdeci">viennacl::enable_if< viennacl::is_scalar< ScalarT1 >::value &&viennacl::is_scalar< ScalarT2 >::value >::type swap(ScalarT1 &s1, ScalarT2 &s2)</div><div class="ttdoc">Swaps the contents of two scalars, data is copied. </div><div class="ttdef"><b>Definition:</b> <a href="cuda_2scalar__operations_8hpp_source.html#l00361">scalar_operations.hpp:361</a></div></div>
<div class="ttc" id="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled_html_a4d2edbd6f51ca00712c5b3ea2bc64fa4"><div class="ttname"><a href="structviennacl_1_1linalg_1_1host__based_1_1spgemm__output__write__enabled.html#a4d2edbd6f51ca00712c5b3ea2bc64fa4">viennacl::linalg::host_based::spgemm_output_write_enabled::apply</a></div><div class="ttdeci">static void apply(unsigned int *ptr, unsigned int value)</div><div class="ttdef"><b>Definition:</b> <a href="spgemm__vector_8hpp_source.html#l00165">spgemm_vector.hpp:165</a></div></div>
<div class="ttc" id="namespaceviennacl_1_1linalg_1_1detail_html_aef015779a92be597e546c3491cebf6c1"><div class="ttname"><a href="namespaceviennacl_1_1linalg_1_1detail.html#aef015779a92be597e546c3491cebf6c1">viennacl::linalg::detail::min</a></div><div class="ttdeci">T min(const T &lhs, const T &rhs)</div><div class="ttdoc">Minimum. </div><div class="ttdef"><b>Definition:</b> <a href="linalg_2detail_2bisect_2util_8hpp_source.html#l00045">util.hpp:45</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_8d734bf48796b8afc2ea3e50ddaea273.html">host_based</a></li><li class="navelem"><a class="el" href="spgemm__vector_8hpp.html">spgemm_vector.hpp</a></li>
<li class="footer">Generated on Wed Jan 20 2016 22:32:42 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>
|
