forked from AnonymousRepo123/AlphaSparse
-
Notifications
You must be signed in to change notification settings - Fork 0
/
parameter_set_strategy.cc
executable file
·1342 lines (1059 loc) · 64.8 KB
/
parameter_set_strategy.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#include "parameter_set_strategy.hpp"
dense_begin_memory_cache_input_file_direct_param_strategy_t get_begin_memory_cache_input_file_direct_param_strategy_from_coo_file(string file_name, data_type type)
{
assert(type == DOUBLE || type == FLOAT);
vector<float> float_val_vec;
vector<double> double_val_vec;
unsigned long max_col_index;
unsigned long max_row_index;
vector<unsigned long> col_index_vec;
vector<unsigned long> row_index_vec;
get_matrix_index_and_val_from_file(file_name, row_index_vec, col_index_vec, float_val_vec, double_val_vec, type, max_row_index, max_col_index);
dense_begin_memory_cache_input_file_direct_param_strategy_t param;
param.col_index_cache = col_index_vec;
param.row_index_cache = row_index_vec;
param.float_val_cache = float_val_vec;
param.double_val_cache = double_val_vec;
param.col_index_max = max_col_index;
param.row_index_max = max_row_index;
param.val_data_type = type;
// 将结果直接返回
return param;
}
param_strategy_of_sub_graph_t val_copy_from_old_param_strategy_of_sub_graph(param_strategy_of_sub_graph_t old_param_strategy_of_compressed_sub_graph)
{
// 检查输入是不是正确
assert(old_param_strategy_of_compressed_sub_graph.param_strategy_vec.size() > 0);
for (unsigned long i = 0; i < old_param_strategy_of_compressed_sub_graph.param_strategy_vec.size(); i++)
{
assert(old_param_strategy_of_compressed_sub_graph.param_strategy_vec[i].param_strategy != NULL);
assert(old_param_strategy_of_compressed_sub_graph.param_strategy_vec[i].param != NULL);
}
// 遍历所有的策略，新建策略参数，然后拷贝出来，最后返回
param_strategy_of_sub_graph_t return_param_strategy_of_compressed_sub_graph;
for (unsigned long i = 0; i < old_param_strategy_of_compressed_sub_graph.param_strategy_vec.size(); i++)
{
// 申请新的节点
param_strategy_node_t new_node;
param_strategy_node_t old_node = old_param_strategy_of_compressed_sub_graph.param_strategy_vec[i];
// 执行一系列的拷贝，其中优化路径的参数在这个函数外面可能需要重新绑定
new_node.node_type = old_node.node_type;
new_node.param = old_node.param;
new_node.strategy_type = old_node.strategy_type;
if (new_node.strategy_type == COMPRESSED_ROW_PADDING_DIRECT_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_row_padding_direct_param_strategy_t();
compressed_row_padding_direct_param_strategy_t* new_param_strategy_ptr = (compressed_row_padding_direct_param_strategy_t *)new_node.param_strategy;
compressed_row_padding_direct_param_strategy_t* old_param_strategy_ptr = (compressed_row_padding_direct_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->multiply = old_param_strategy_ptr->multiply;
new_param_strategy_ptr->padding_row_length = old_param_strategy_ptr->padding_row_length;
}
else if (new_node.strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_tblock_level_row_div_evenly_param_strategy_t();
compressed_tblock_level_row_div_evenly_param_strategy_t* new_param_strategy_ptr = (compressed_tblock_level_row_div_evenly_param_strategy_t *)new_node.param_strategy;
compressed_tblock_level_row_div_evenly_param_strategy_t* old_param_strategy_ptr = (compressed_tblock_level_row_div_evenly_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->block_row_num = old_param_strategy_ptr->block_row_num;
}
else if (new_node.strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_ACC_TO_LEAST_NNZ_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t();
compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t* new_param_strategy_ptr = (compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t *)new_node.param_strategy;
compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t* old_param_strategy_ptr = (compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->nnz_low_bound = old_param_strategy_ptr->nnz_low_bound;
}
else if (new_node.strategy_type == COMPRESSED_TBLOCK_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_tblock_level_col_div_fixed_param_strategy_t();
compressed_tblock_level_col_div_fixed_param_strategy_t* new_param_strategy_ptr = (compressed_tblock_level_col_div_fixed_param_strategy_t *)new_node.param_strategy;
compressed_tblock_level_col_div_fixed_param_strategy_t* old_param_strategy_ptr = (compressed_tblock_level_col_div_fixed_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->col_block_nnz_num = old_param_strategy_ptr->col_block_nnz_num;
}
else if (new_node.strategy_type == COMPRESSED_WARP_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_warp_level_row_div_evenly_param_strategy_t();
compressed_warp_level_row_div_evenly_param_strategy_t* new_param_strategy_ptr = (compressed_warp_level_row_div_evenly_param_strategy_t *)new_node.param_strategy;
compressed_warp_level_row_div_evenly_param_strategy_t* old_param_strategy_ptr = (compressed_warp_level_row_div_evenly_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->warp_row_num_of_each_BLB = old_param_strategy_ptr->warp_row_num_of_each_BLB;
}
else if (new_node.strategy_type == COMPRESSED_WARP_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_warp_level_col_div_fixed_param_strategy_t();
compressed_warp_level_col_div_fixed_param_strategy_t* new_param_strategy_ptr = (compressed_warp_level_col_div_fixed_param_strategy_t *)new_node.param_strategy;
compressed_warp_level_col_div_fixed_param_strategy_t* old_param_strategy_ptr = (compressed_warp_level_col_div_fixed_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->col_block_nnz_num = old_param_strategy_ptr->col_block_nnz_num;
}
else if (new_node.strategy_type == COMPRESSED_THREAD_LEVEL_ROW_DIV_NONE_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_thread_level_row_div_none_param_strategy_t();
compressed_thread_level_row_div_none_param_strategy_t* new_param_strategy_ptr = (compressed_thread_level_row_div_none_param_strategy_t *)new_node.param_strategy;
compressed_thread_level_row_div_none_param_strategy_t* old_param_strategy_ptr = (compressed_thread_level_row_div_none_param_strategy_t *)old_node.param_strategy;
}
else if (new_node.strategy_type == COMPRESSED_THREAD_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_thread_level_col_div_fixed_param_strategy_t();
compressed_thread_level_col_div_fixed_param_strategy_t* new_param_strategy_ptr = (compressed_thread_level_col_div_fixed_param_strategy_t *)new_node.param_strategy;
compressed_thread_level_col_div_fixed_param_strategy_t* old_param_strategy_ptr = (compressed_thread_level_col_div_fixed_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->col_block_nnz_num = old_param_strategy_ptr->col_block_nnz_num;
}
else if (new_node.strategy_type == COMPRESSED_THREAD_LEVEL_NNZ_DIV_DIRECT_PARAM_STRATEGY)
{
new_node.param_strategy = new compressed_thread_level_nnz_div_direct_param_strategy_t();
compressed_thread_level_nnz_div_direct_param_strategy_t* new_param_strategy_ptr = (compressed_thread_level_nnz_div_direct_param_strategy_t *)new_node.param_strategy;
compressed_thread_level_nnz_div_direct_param_strategy_t* old_param_strategy_ptr = (compressed_thread_level_nnz_div_direct_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->block_nnz_num = old_param_strategy_ptr->block_nnz_num;
}
else if (new_node.strategy_type == DENSE_ROW_COARSE_SORT_FIXED_PARAM_STRATEGY)
{
new_node.param_strategy = new dense_row_coarse_sort_fixed_param_strategy_t();
dense_row_coarse_sort_fixed_param_strategy_t* new_param_strategy_ptr = (dense_row_coarse_sort_fixed_param_strategy_t*)new_node.param_strategy;
dense_row_coarse_sort_fixed_param_strategy_t* old_param_strategy_ptr = (dense_row_coarse_sort_fixed_param_strategy_t*)old_node.param_strategy;
new_param_strategy_ptr->row_nnz_low_bound_step_size = old_param_strategy_ptr->row_nnz_low_bound_step_size;
}
else if (new_node.strategy_type == DENSE_BEGIN_MEMORY_CACHE_INPUT_FILE_DIRECT_PARAM_STRATEGY)
{
new_node.param_strategy = new dense_begin_memory_cache_input_file_direct_param_strategy_t();
dense_begin_memory_cache_input_file_direct_param_strategy_t* new_param_strategy_ptr = (dense_begin_memory_cache_input_file_direct_param_strategy_t *)new_node.param_strategy;
dense_begin_memory_cache_input_file_direct_param_strategy_t* old_param_strategy_ptr = (dense_begin_memory_cache_input_file_direct_param_strategy_t *)old_node.param_strategy;
new_param_strategy_ptr->col_index_cache = old_param_strategy_ptr->col_index_cache;
new_param_strategy_ptr->col_index_max = old_param_strategy_ptr->col_index_max;
new_param_strategy_ptr->double_val_cache = old_param_strategy_ptr->double_val_cache;
new_param_strategy_ptr->float_val_cache = old_param_strategy_ptr->float_val_cache;
new_param_strategy_ptr->row_index_cache = old_param_strategy_ptr->row_index_cache;
new_param_strategy_ptr->row_index_max = old_param_strategy_ptr->row_index_max;
new_param_strategy_ptr->val_data_type = old_param_strategy_ptr->val_data_type;
}
else if (new_node.strategy_type == COMPRESS_NONE_PARAM_STRATEGY)
{
new_node.param_strategy = new compress_none_param_strategy_t();
// 但是实际上啥都不用做
compress_none_param_strategy_t* new_param_strategy_ptr = (compress_none_param_strategy_t*)new_node.param_strategy;
compress_none_param_strategy_t* old_param_strategy_ptr = (compress_none_param_strategy_t*)old_node.param_strategy;
}
else if (new_node.strategy_type == DENSE_ROW_DIV_ACC_TO_EXPONENTIAL_INCREASE_ROW_NNZ_PARAM_STRATEGY)
{
new_node.param_strategy = new dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t();
// 新的和旧的参数
dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t* new_param_strategy_ptr = (dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t*)new_node.param_strategy;
dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t* old_param_strategy_ptr = (dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t*)old_node.param_strategy;
new_param_strategy_ptr->expansion_rate = old_param_strategy_ptr->expansion_rate;
new_param_strategy_ptr->lowest_nnz_bound_of_row = old_param_strategy_ptr->lowest_nnz_bound_of_row;
new_param_strategy_ptr->highest_nnz_bound_of_row = old_param_strategy_ptr->highest_nnz_bound_of_row;
new_param_strategy_ptr->sub_dense_block_id = old_param_strategy_ptr->sub_dense_block_id;
}
else
{
// 当前策略不被支持
cout << "val_copy_from_old_param_strategy_of_sub_graph: param strategy type is not supported" << endl;
assert(false);
}
// 将新的策略节点放到子图的策略骨架中
return_param_strategy_of_compressed_sub_graph.param_strategy_vec.push_back(new_node);
}
return return_param_strategy_of_compressed_sub_graph;
}
void execute_compressed_row_padding_direct_param_strategy(compressed_row_padding_direct_param_strategy_t* param_strategy, exe_compress_row_padding_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
// 当前应该仅仅完成了压缩
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() == 2);
// cout << "param_strategy->multiply:" << param_strategy->multiply << ",param_strategy->padding_row_length:" << param_strategy->padding_row_length << endl;
assert(param_strategy->multiply > 0 && param_strategy->padding_row_length > 0);
// 仅仅完成了压缩
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
assert(param_strategy != NULL && param != NULL);
// 给对应的节点赋值
param->multiply = param_strategy->multiply;
param->padding_row_length = param_strategy->padding_row_length;
}
// 执行等长的行条带分块
void execute_compressed_tblock_level_row_div_evenly_param_strategy(compressed_tblock_level_row_div_evenly_param_strategy_t* param_strategy, exe_compress_tblock_level_row_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() == 2);
// cout << "param_strategy->block_row_num:" << param_strategy->block_row_num << ",param->row_num_of_each_BLB.size():" << param->row_num_of_each_BLB.size() << endl;
assert(param_strategy != NULL && param_strategy->block_row_num > 0 && param->row_num_of_each_BLB.size() == 0);
// 仅仅完成了压缩
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
assert(param_strategy != NULL && param != NULL);
// 计算当前压缩子块的行数量
assert(compressed_block_ptr->read_index[0]->max_row_index >= compressed_block_ptr->read_index[0]->min_row_index);
unsigned long sub_matrix_row_num = compressed_block_ptr->read_index[0]->max_row_index - compressed_block_ptr->read_index[0]->min_row_index + 1;
vector<unsigned long> block_row_num_vec = row_block_size_of_a_sub_matrix_by_fixed_div(sub_matrix_row_num, param_strategy->block_row_num);
// 将参数放到对应的节点中
for (auto row_num : block_row_num_vec)
{
param->row_num_of_each_BLB.push_back(row_num);
}
}
void execute_compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy(compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t* param_strategy, exe_compress_tblock_level_row_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() == 2);
assert(param_strategy != NULL && param_strategy->nnz_low_bound > 0 && param->row_num_of_each_BLB.size() == 0);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
// 获取当前子块的非零元数量
vector<unsigned long> row_nnz_vec = get_nnz_of_each_row_in_compressed_sub_matrix(compressed_block_ptr);
// 获取分块结果
vector<unsigned long> block_row_num_vec = row_block_size_of_a_sub_matrix_by_nnz_low_bound(row_nnz_vec, param_strategy->nnz_low_bound);
// 将分块结果赋值给对应的节点
for (auto row_num : block_row_num_vec)
{
param->row_num_of_each_BLB.push_back(row_num);
}
}
// 按照纵切分
void execute_compressed_tblock_level_col_div_fixed_param_strategy(compressed_tblock_level_col_div_fixed_param_strategy_t* param_strategy, exe_compress_tblock_level_col_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() == 2);
assert(param_strategy != NULL && param_strategy->col_block_nnz_num > 0 && param->col_block_nnz_num_of_each_BLB.size() == 0);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
// 获取当前子块的非零元数量
vector<unsigned long> row_nnz_vec = get_nnz_of_each_row_in_compressed_sub_matrix(compressed_block_ptr);
// 获得每一行的纵分块的非零元数量
vector<vector<unsigned int>> col_block_size_vec = col_block_size_of_each_row(row_nnz_vec, param_strategy->col_block_nnz_num);
assert(row_nnz_vec.size() >= col_block_size_vec.size());
param->col_block_nnz_num_of_each_BLB = col_block_size_vec;
}
void execute_compressed_warp_level_row_div_evenly_param_strategy(compressed_warp_level_row_div_evenly_param_strategy_t* param_strategy, exe_compress_warp_level_row_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(param_strategy != NULL && param_strategy->warp_row_num_of_each_BLB > 0 && param->row_num_of_each_WLB_in_BLB.size() == 0);
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() <= 3);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
// 如果没有行分块就执行一个默认的行分块
if (compressed_block_ptr->read_index.size() == 2)
{
// 执行默认的行分块
default_sep_tblock_level_row_csr(compressed_block_ptr);
}
assert(matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr->read_index.size() == 3);
index_of_compress_block_t* block_level_index = compressed_block_ptr->read_index[2];
// 获取每一个BLB子块的行数量，BLB块中的行在索引上可能并不是连续的
vector<unsigned long> row_num_of_each_BLB;
for (unsigned long BLB_id = 0; BLB_id < block_level_index->block_num; BLB_id++)
{
row_num_of_each_BLB.push_back(read_from_array_with_data_type(block_level_index->row_number_of_block_arr, block_level_index->data_type_of_row_number_of_block_arr, BLB_id));
}
assert(row_num_of_each_BLB.size() > 0);
// 获取进一步分块的数组
vector<vector<unsigned int>> row_num_of_each_sub_block = row_block_size_of_each_sub_block_by_fixed_div(row_num_of_each_BLB, param_strategy->warp_row_num_of_each_BLB);
assert(row_num_of_each_sub_block.size() == row_num_of_each_BLB.size());
param->row_num_of_each_WLB_in_BLB = row_num_of_each_sub_block;
}
void execute_compressed_warp_level_col_div_fixed_param_strategy(compressed_warp_level_col_div_fixed_param_strategy_t* param_strategy, exe_compress_warp_level_col_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(param_strategy != NULL && param != NULL && param_strategy->col_block_nnz_num > 0 && param->col_num_of_WLB_in_each_parent_row_block_or_BLB.size() == 0);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
assert(compressed_block_ptr->read_index.size() <= 3);
// 如果没有执行过BLB切分，那就执行一个默认的BLB分块
if (compressed_block_ptr->read_index.size() == 2)
{
// 默认的block分块
default_sep_tblock_level_row_csr(compressed_block_ptr);
}
// 仅仅经过了BLB切分
assert(compressed_block_ptr->read_index.size() == 3);
index_of_compress_block_t* block_level_index = compressed_block_ptr->read_index[2];
// 做一个检查，禁止在BLB层次进行纵分块，查看BLB的首行行号，
for (unsigned long i = 0; i < block_level_index->block_num - 1; i++)
{
unsigned long cur_BLB_first_row_index = read_from_array_with_data_type(block_level_index->index_of_the_first_row_arr, block_level_index->data_type_of_index_of_the_first_row_arr, i);
unsigned long next_BLB_first_row_index = read_from_array_with_data_type(block_level_index->index_of_the_first_row_arr, block_level_index->data_type_of_index_of_the_first_row_arr, i + 1);
// BLB的首行行号不能相等，并且递增
assert(cur_BLB_first_row_index < next_BLB_first_row_index);
}
// 查看整个子块行非零元的数量
vector<unsigned long> row_nnz_vec = get_nnz_of_each_row_in_compressed_sub_matrix(compressed_block_ptr);
// 为每一个非空行执行列分块
vector<vector<unsigned int>> col_block_size_vec = col_block_size_of_each_row_without_block_merge(row_nnz_vec, param_strategy->col_block_nnz_num);
param->col_num_of_WLB_in_each_parent_row_block_or_BLB = col_block_size_vec;
}
void execute_compressed_thread_level_row_div_none_param_strategy(compressed_thread_level_row_div_none_param_strategy_t* param_strategy, exe_compress_thread_level_row_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
// 必须执行完对应的
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(param != NULL && param_strategy != NULL);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
// 最多只能执行到行分块，
assert(compressed_block_ptr->read_index.size() <= 4);
// 什么都不用做
}
void execute_compressed_thread_level_col_div_fixed_param_strategy(compressed_thread_level_col_div_fixed_param_strategy_t* param_strategy, exe_compress_thread_level_col_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(matrix != NULL);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
assert(param != NULL && param_strategy != NULL);
assert(param->col_num_of_TLB_in_each_parent_block.size() == 0 && param_strategy->col_block_nnz_num > 0);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
assert(compressed_block_ptr->read_index.size() <= 4);
// 因为需要获得WLB的块数量，如果缺少对应的分块就需要执行默认分块
if (compressed_block_ptr->read_index.size() == 2)
{
// 执行BLB的默认分块
default_sep_tblock_level_row_csr(compressed_block_ptr);
assert(compressed_block_ptr->read_index.size() == 3);
}
// 默认的WLB分块
if (compressed_block_ptr->read_index.size() == 3)
{
default_sep_warp_level_row_csr(compressed_block_ptr);
}
assert(compressed_block_ptr->read_index.size() == 4);
unsigned long WLB_num = compressed_block_ptr->read_index[3]->block_num;
assert(WLB_num > 0);
// 设定TLB纵分块的参数
vector<unsigned long> col_block_size_vec = col_div_of_TLB_global_fixed_col_size(WLB_num, param_strategy->col_block_nnz_num);
assert(col_block_size_vec.size() == WLB_num);
param->col_num_of_TLB_in_each_parent_block = col_block_size_vec;
}
void execute_compressed_thread_level_nnz_div_direct_param_strategy(compressed_thread_level_nnz_div_direct_param_strategy_t* param_strategy, exe_compress_thread_level_nnz_div_param_t* param, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
// 之前不能执行任何的分块
assert(param_strategy != NULL && param != NULL && matrix != NULL);
assert(param_strategy->block_nnz_num > 0);
assert(matrix->block_coor_table.item_arr.size() > sub_matrix_id);
assert(matrix->block_coor_table.item_arr[sub_matrix_id] != NULL && matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr != NULL);
compressed_block_t* compressed_block_ptr = matrix->block_coor_table.item_arr[sub_matrix_id]->compressed_block_ptr;
assert(compressed_block_ptr->read_index.size() == 2);
// 执行对应的参数设定
param->TLB_nnz_num = param_strategy->block_nnz_num;
}
void execute_dense_row_coarse_sort_fixed_param_strategy(dense_row_coarse_sort_fixed_param_strategy_t* param_strategy, exe_dense_row_coarse_sort_param_t* param, sparse_struct_t *matrix)
{
assert(param_strategy != NULL && param != NULL && matrix != NULL);
// 之前不能排序，不能分块
assert(matrix->is_sorted == false && matrix->is_blocked == false && matrix->is_compressed == false);
assert(matrix->sorted_row_index == NULL && matrix->coo_value_cache != NULL && matrix->coo_col_index_cache != NULL && matrix->coo_row_index_cache != NULL);
// 根据矩阵的最长非零元来处理
// 最大非零元数量
vector<unsigned long> row_nnz_vec = get_nnz_of_each_row_in_spec_range(matrix->coo_row_index_cache, UNSIGNED_LONG, 0, matrix->dense_row_number - 1, 0, matrix->nnz - 1);
assert(row_nnz_vec.size() == matrix->dense_row_number);
assert(row_nnz_vec.size() > 0);
// 查看最大的行非零元数量
unsigned long max_row_nnz = row_nnz_vec[0];
// 查看最小的行非零元数量
unsigned long min_row_nnz = row_nnz_vec[0];
for (unsigned long i = 0; i < row_nnz_vec.size(); i++)
{
if (max_row_nnz < row_nnz_vec[i])
{
max_row_nnz = row_nnz_vec[i];
}
if (min_row_nnz > row_nnz_vec[i])
{
min_row_nnz = row_nnz_vec[i];
}
}
// 要执行的粗粒度排序的行非零元下界
vector<unsigned long> bin_nnz_range;
// cout << "param_strategy->row_nnz_low_bound_step_size:" << param_strategy->row_nnz_low_bound_step_size << endl;
// 如果可以容纳的行非零元宽度为仅仅为1，这个操作就退化成了排序操作，引入排序相关的操作，提高性能
if (param_strategy->row_nnz_low_bound_step_size == 1)
{
bin_nnz_range = bin_row_nnz_low_bound_of_fixed_granularity_coar_sort(row_nnz_vec, 1);
}
else
{
for (unsigned long i = min_row_nnz; i <= max_row_nnz; i = i + param_strategy->row_nnz_low_bound_step_size)
{
if (i == min_row_nnz)
{
bin_nnz_range.push_back(0);
}
else
{
bin_nnz_range.push_back(i);
}
}
}
// 将对应的排序算好，给对应的图节点赋值
param->bin_row_nnz_low_bound = bin_nnz_range;
}
void execute_dense_begin_memory_cache_input_file_direct_param_strategy(dense_begin_memory_cache_input_file_direct_param_strategy_t* param_strategy, exe_begin_memory_cache_input_file_param_t* param, sparse_struct_t* matrix)
{
// 保证参数都是有的，但是矩阵是没的
assert(param_strategy != NULL && param != NULL && matrix == NULL);
// 检查一下参数，将策略的参数拷贝到节点的参数中
assert(param_strategy->row_index_cache.size() > 0 && param_strategy->col_index_max > 0 && param_strategy->row_index_max > 0);
assert(param_strategy->row_index_cache.size() == param_strategy->col_index_cache.size());
assert(param_strategy->val_data_type == DOUBLE || param_strategy->val_data_type == FLOAT);
// 检查值数组的缓存
if (param_strategy->val_data_type == DOUBLE)
{
assert(param_strategy->double_val_cache.size() == param_strategy->row_index_cache.size());
assert(param_strategy->float_val_cache.size() == 0);
}
if (param_strategy->val_data_type == FLOAT)
{
assert(param_strategy->float_val_cache.size() == param_strategy->row_index_cache.size());
assert(param_strategy->double_val_cache.size() == 0);
}
// 将参数直接拷贝到节点中
param->col_index_cache = param_strategy->col_index_cache;
param->col_index_max = param_strategy->col_index_max;
param->double_val_cache = param_strategy->double_val_cache;
param->float_val_cache = param_strategy->float_val_cache;
param->row_index_cache = param_strategy->row_index_cache;
param->row_index_max = param_strategy->row_index_max;
param->val_data_type = param_strategy->val_data_type;
}
void execute_compress_none_param_strategy(compress_none_param_strategy_t* param_strategy, exe_compress_param_t* param, sparse_struct_t* matrix)
{
assert(param_strategy != NULL && param != NULL && matrix != NULL);
}
void execute_dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy(dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t* param_strategy, exe_dense_row_div_param_t* param, sparse_struct_t* matrix)
{
assert(param_strategy != NULL && param != NULL && matrix != NULL);
// 查看子块的编号是不是满足范围
if (matrix->block_coor_table.item_arr.size() == 0)
{
assert(param_strategy->sub_dense_block_id == matrix->block_coor_table.item_arr.size());
}
if (matrix->block_coor_table.item_arr.size() > 0)
{
assert(param_strategy->sub_dense_block_id < matrix->block_coor_table.item_arr.size());
}
// 检查参数是否正确
assert(param_strategy->expansion_rate > 0 && param_strategy->lowest_nnz_bound_of_row > 0);
// 将参数覆盖到优化路径的节点
param->dense_sub_block_id = param_strategy->sub_dense_block_id;
// 子块的行非零元数量
vector<unsigned long> row_nnz_vec;
if (matrix->block_coor_table.item_arr.size() == 0)
{
// 整个矩阵的行非零元数量
// vector<unsigned long> get_nnz_of_each_row_in_spec_range(void *row_index_arr, data_type data_type_of_row_index_arr, unsigned long begin_row_bound, unsigned long end_row_bound, unsigned long global_coo_start, unsigned long global_coo_end)
row_nnz_vec = get_nnz_of_each_row_in_spec_range(matrix->coo_row_index_cache, UNSIGNED_LONG, 0, matrix->dense_row_number - 1, 0, matrix->nnz - 1);
assert(row_nnz_vec.size() == matrix->dense_row_number);
}
else
{
// 获取一个子块的索引范围和非零元范围
unsigned long row_begin_index = matrix->block_coor_table.item_arr[param->dense_sub_block_id]->min_dense_row_index;
unsigned long row_end_index = matrix->block_coor_table.item_arr[param->dense_sub_block_id]->max_dense_row_index;
unsigned long coo_begin_index = matrix->block_coor_table.item_arr[param->dense_sub_block_id]->begin_coo_index;
unsigned long coo_end_index = matrix->block_coor_table.item_arr[param->dense_sub_block_id]->end_coo_index;
assert(coo_end_index <= matrix->nnz - 1);
assert(row_end_index <= matrix->dense_row_number - 1);
row_nnz_vec = get_nnz_of_each_row_in_spec_range(matrix->coo_row_index_cache, UNSIGNED_LONG, row_begin_index, row_end_index, coo_begin_index, coo_end_index);
assert(row_nnz_vec.size() == row_end_index - row_begin_index + 1);
}
// 行非零元数量的区间是不包含上界的但是包含下界
vector<unsigned long> row_div_position = row_div_position_acc_to_exponential_increase_row_nnz_range(row_nnz_vec, param_strategy->lowest_nnz_bound_of_row, param_strategy->highest_nnz_bound_of_row, param_strategy->expansion_rate);
param->row_div_position = row_div_position;
}
void execute_param_strategy_node_of_dense_matrix(param_strategy_node_t* node, sparse_struct_t* matrix)
{
assert(node != NULL);
assert(node->param != NULL && node->param_strategy != NULL);
if (node->strategy_type == DENSE_ROW_COARSE_SORT_FIXED_PARAM_STRATEGY)
{
assert(node->node_type == DENSE_ROW_COARSE_SORT);
// 这里要求matrix必须存在
assert(matrix != NULL);
execute_dense_row_coarse_sort_fixed_param_strategy((dense_row_coarse_sort_fixed_param_strategy_t *)node->param_strategy, (exe_dense_row_coarse_sort_param_t *)node->param, matrix);
assert(((exe_dense_row_coarse_sort_param_t *)node->param)->bin_row_nnz_low_bound.size() > 0);
return;
}
// 执行对应的策略
if (node->strategy_type == DENSE_BEGIN_MEMORY_CACHE_INPUT_FILE_DIRECT_PARAM_STRATEGY)
{
assert(node->node_type == BEGIN_MEMORY_CACHE_INPUT_FILE);
assert(matrix == NULL);
execute_dense_begin_memory_cache_input_file_direct_param_strategy((dense_begin_memory_cache_input_file_direct_param_strategy *)node->param_strategy, (exe_begin_memory_cache_input_file_param_t *)node->param, matrix);
assert(((exe_begin_memory_cache_input_file_param_t *)node->param)->col_index_cache.size() > 0);
assert(((exe_begin_memory_cache_input_file_param_t *)node->param)->row_index_cache.size() > 0);
return;
}
// 执行对应的调参策略
if (node->strategy_type == COMPRESS_NONE_PARAM_STRATEGY)
{
assert(node->node_type == COMPRESS);
assert(matrix != NULL);
execute_compress_none_param_strategy((compress_none_param_strategy_t *)node->param_strategy, (exe_compress_param_t*)node->param, matrix);
return;
}
// 执行对应的调参策略
if (node->strategy_type == DENSE_ROW_DIV_ACC_TO_EXPONENTIAL_INCREASE_ROW_NNZ_PARAM_STRATEGY)
{
assert(node->node_type == DENSE_ROW_DIV);
assert(matrix != NULL);
execute_dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy((dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t *)node->param_strategy, (exe_dense_row_div_param_t *)node->param, matrix);
return;
}
// 不支持参数设定
cout << "execute_param_strategy_node_of_dense_matrix: strategy is not supported" << endl;
assert(false);
}
void execute_param_strategy_node_of_sub_compressed_matrix(param_strategy_node_t* node, sparse_struct_t* matrix, unsigned long sub_matrix_id)
{
assert(node != NULL && matrix != NULL);
assert(node->param != NULL && node->param_strategy != NULL);
assert(sub_matrix_id < matrix->block_coor_table.item_arr.size());
// 执行所有的的参数设定
if (node->strategy_type == COMPRESSED_ROW_PADDING_DIRECT_PARAM_STRATEGY)
{
assert(node->node_type == COMPRESSED_ROW_PADDING);
execute_compressed_row_padding_direct_param_strategy((compressed_row_padding_direct_param_strategy_t *)node->param_strategy, (exe_compress_row_padding_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_row_padding_param_t *)node->param)->multiply > 0);
assert(((exe_compress_row_padding_param_t *)node->param)->padding_row_length > 0);
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
// 均匀的BLB行分块
assert(node->node_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV);
execute_compressed_tblock_level_row_div_evenly_param_strategy((compressed_tblock_level_row_div_evenly_param_strategy_t *)node->param_strategy, (exe_compress_tblock_level_row_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_tblock_level_row_div_param_t *)node->param)->row_num_of_each_BLB.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_ACC_TO_LEAST_NNZ_PARAM_STRATEGY)
{
// 按照非零元数量的BLB行分块
assert(node->node_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV);
execute_compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy((compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t *)node->param_strategy, (exe_compress_tblock_level_row_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_tblock_level_row_div_param_t *)node->param)->row_num_of_each_BLB.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
// 执行固定长度的BLB分块
assert(node->node_type == COMPRESSED_TBLOCK_LEVEL_COL_DIV);
execute_compressed_tblock_level_col_div_fixed_param_strategy((compressed_tblock_level_col_div_fixed_param_strategy_t *)node->param_strategy, (exe_compress_tblock_level_col_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_tblock_level_col_div_param_t *)node->param)->col_block_nnz_num_of_each_BLB.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_WARP_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
// 执行固定长度WLB行分块
assert(node->node_type == COMPRESSED_WARP_LEVEL_ROW_DIV);
execute_compressed_warp_level_row_div_evenly_param_strategy((compressed_warp_level_row_div_evenly_param_strategy_t *)node->param_strategy, (exe_compress_warp_level_row_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_warp_level_row_div_param_t *)node->param)->row_num_of_each_WLB_in_BLB.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_WARP_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
// 固定长度的WLB列分块
assert(node->node_type == COMPRESSED_WARP_LEVEL_COL_DIV);
execute_compressed_warp_level_col_div_fixed_param_strategy((compressed_warp_level_col_div_fixed_param_strategy_t *)node->param_strategy, (exe_compress_warp_level_col_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_warp_level_col_div_param_t *)node->param)->col_num_of_WLB_in_each_parent_row_block_or_BLB.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_ROW_DIV_NONE_PARAM_STRATEGY)
{
// 一行一个TLB分块
assert(node->node_type == COMPRESSED_THREAD_LEVEL_ROW_DIV);
execute_compressed_thread_level_row_div_none_param_strategy((compressed_thread_level_row_div_none_param_strategy_t *)node->param_strategy, (exe_compress_thread_level_row_div_param_t *)node->param, matrix, sub_matrix_id);
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
// TLB列分块
assert(node->node_type == COMPRESSED_THREAD_LEVEL_COL_DIV);
execute_compressed_thread_level_col_div_fixed_param_strategy((compressed_thread_level_col_div_fixed_param_strategy_t *)node->param_strategy, (exe_compress_thread_level_col_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_thread_level_col_div_param_t *)node->param)->col_num_of_TLB_in_each_parent_block.size() > 0);
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_NNZ_DIV_DIRECT_PARAM_STRATEGY)
{
// TLB的非零元数量分块
assert(node->node_type == COMPRESSED_THREAD_LEVEL_NNZ_DIV);
execute_compressed_thread_level_nnz_div_direct_param_strategy((compressed_thread_level_nnz_div_direct_param_strategy_t *)node->param_strategy, (exe_compress_thread_level_nnz_div_param_t *)node->param, matrix, sub_matrix_id);
assert(((exe_compress_thread_level_nnz_div_param_t *)node->param)->TLB_nnz_num > 0);
return;
}
// 不支持参数设定
cout << "execute_param_strategy_node_of_sub_compressed_matrix: strategy is not supported" << endl;
assert(false);
}
void del_strategy_of_param_strategy_node(param_strategy_node_t* node)
{
assert(node->param_strategy != NULL);
// 根据类型执行对应的析构操作
if (node->strategy_type == COMPRESSED_ROW_PADDING_DIRECT_PARAM_STRATEGY)
{
delete (compressed_row_padding_direct_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
delete (compressed_tblock_level_row_div_evenly_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_ROW_DIV_ACC_TO_LEAST_NNZ_PARAM_STRATEGY)
{
delete (compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_TBLOCK_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
delete (compressed_tblock_level_col_div_fixed_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_WARP_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY)
{
delete (compressed_warp_level_row_div_evenly_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_WARP_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
delete (compressed_warp_level_col_div_fixed_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_ROW_DIV_NONE_PARAM_STRATEGY)
{
delete (compressed_thread_level_row_div_none_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY)
{
delete (compressed_thread_level_col_div_fixed_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESSED_THREAD_LEVEL_NNZ_DIV_DIRECT_PARAM_STRATEGY)
{
delete (compressed_thread_level_nnz_div_direct_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == DENSE_ROW_COARSE_SORT_FIXED_PARAM_STRATEGY)
{
delete (dense_row_coarse_sort_fixed_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == DENSE_BEGIN_MEMORY_CACHE_INPUT_FILE_DIRECT_PARAM_STRATEGY)
{
delete (dense_begin_memory_cache_input_file_direct_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == COMPRESS_NONE_PARAM_STRATEGY)
{
delete (compress_none_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
if (node->strategy_type == DENSE_ROW_DIV_ACC_TO_EXPONENTIAL_INCREASE_ROW_NNZ_PARAM_STRATEGY)
{
delete (dense_row_div_acc_to_exponential_increase_row_nnz_param_strategy_t *)node->param_strategy;
node->param_strategy = NULL;
return;
}
cout << "del_strategy_of_param_strategy_node: strategy is not supported" << endl;
assert(false);
}
param_strategy_node_t init_compressed_row_padding_direct_param_strategy(compressed_row_padding_direct_param_strategy_t param_strategy, exe_compress_row_padding_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_ROW_PADDING;
return_node.strategy_type = COMPRESSED_ROW_PADDING_DIRECT_PARAM_STRATEGY;
return_node.param = param;
compressed_row_padding_direct_param_strategy_t* param_strategy_ptr = new compressed_row_padding_direct_param_strategy_t();
param_strategy_ptr->multiply = param_strategy.multiply;
param_strategy_ptr->padding_row_length = param_strategy.padding_row_length;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_tblock_level_row_div_evenly_param_strategy(compressed_tblock_level_row_div_evenly_param_strategy_t param_strategy, exe_compress_tblock_level_row_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_TBLOCK_LEVEL_ROW_DIV;
return_node.strategy_type = COMPRESSED_TBLOCK_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY;
return_node.param = param;
compressed_tblock_level_row_div_evenly_param_strategy_t* param_strategy_ptr = new compressed_tblock_level_row_div_evenly_param_strategy_t();
param_strategy_ptr->block_row_num = param_strategy.block_row_num;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy(compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t param_strategy, exe_compress_tblock_level_row_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_TBLOCK_LEVEL_ROW_DIV;
return_node.strategy_type = COMPRESSED_TBLOCK_LEVEL_ROW_DIV_ACC_TO_LEAST_NNZ_PARAM_STRATEGY;
return_node.param = param;
compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t* param_strategy_ptr = new compressed_tblock_level_row_div_acc_to_least_nnz_param_strategy_t();
param_strategy_ptr->nnz_low_bound = param_strategy.nnz_low_bound;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_tblock_level_col_div_fixed_param_strategy(compressed_tblock_level_col_div_fixed_param_strategy_t param_strategy, exe_compress_tblock_level_col_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_TBLOCK_LEVEL_COL_DIV;
return_node.strategy_type = COMPRESSED_TBLOCK_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY;
return_node.param = param;
compressed_tblock_level_col_div_fixed_param_strategy_t* param_strategy_ptr = new compressed_tblock_level_col_div_fixed_param_strategy_t();
param_strategy_ptr->col_block_nnz_num = param_strategy.col_block_nnz_num;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_warp_level_row_div_evenly_param_strategy(compressed_warp_level_row_div_evenly_param_strategy_t param_strategy, exe_compress_warp_level_row_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_WARP_LEVEL_ROW_DIV;
return_node.strategy_type = COMPRESSED_WARP_LEVEL_ROW_DIV_EVENLY_PARAM_STRATEGY;
return_node.param = param;
compressed_warp_level_row_div_evenly_param_strategy_t* param_strategy_ptr = new compressed_warp_level_row_div_evenly_param_strategy_t();
param_strategy_ptr->warp_row_num_of_each_BLB = param_strategy.warp_row_num_of_each_BLB;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_warp_level_col_div_fixed_param_strategy(compressed_warp_level_col_div_fixed_param_strategy_t param_strategy, exe_compress_warp_level_col_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_WARP_LEVEL_COL_DIV;
return_node.strategy_type = COMPRESSED_WARP_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY;
return_node.param = param;
compressed_warp_level_col_div_fixed_param_strategy_t* param_strategy_ptr = new compressed_warp_level_col_div_fixed_param_strategy_t();
param_strategy_ptr->col_block_nnz_num = param_strategy.col_block_nnz_num;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_thread_level_row_div_none_param_strategy(compressed_thread_level_row_div_none_param_strategy_t param_strategy, exe_compress_thread_level_row_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_THREAD_LEVEL_ROW_DIV;
return_node.strategy_type = COMPRESSED_THREAD_LEVEL_ROW_DIV_NONE_PARAM_STRATEGY;
return_node.param = param;
compressed_thread_level_row_div_none_param_strategy_t* param_strategy_ptr = new compressed_thread_level_row_div_none_param_strategy_t();
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_thread_level_col_div_fixed_param_strategy(compressed_thread_level_col_div_fixed_param_strategy_t param_strategy, exe_compress_thread_level_col_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_THREAD_LEVEL_COL_DIV;
return_node.strategy_type = COMPRESSED_THREAD_LEVEL_COL_DIV_FIXED_PARAM_STRATEGY;
return_node.param = param;
compressed_thread_level_col_div_fixed_param_strategy_t* param_strategy_ptr = new compressed_thread_level_col_div_fixed_param_strategy_t();
param_strategy_ptr->col_block_nnz_num = param_strategy.col_block_nnz_num;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_compressed_thread_level_nnz_div_direct_param_strategy(compressed_thread_level_nnz_div_direct_param_strategy_t param_strategy, exe_compress_thread_level_nnz_div_param_t* param)
{
assert(param != NULL);
param_strategy_node_t return_node;
return_node.node_type = COMPRESSED_THREAD_LEVEL_NNZ_DIV;
return_node.strategy_type = COMPRESSED_THREAD_LEVEL_NNZ_DIV_DIRECT_PARAM_STRATEGY;
return_node.param = param;
compressed_thread_level_nnz_div_direct_param_strategy_t* param_strategy_ptr = new compressed_thread_level_nnz_div_direct_param_strategy_t();
param_strategy_ptr->block_nnz_num = param_strategy.block_nnz_num;
return_node.param_strategy = param_strategy_ptr;
return return_node;
}
param_strategy_node_t init_dense_row_coarse_sort_fixed_param_strategy(dense_row_coarse_sort_fixed_param_strategy_t param_strategy, exe_dense_row_coarse_sort_param_t* param)
{
assert(param != NULL);