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
|
/*
* Copyright (c) 2017-2018 ARM Limited.
*
* SPDX-License-Identifier: MIT
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in all
* copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include "helpers.h"
#include "helpers_asymm.h"
#if defined(COLS_B)
/** This OpenCL kernel computes the matrix multiplication between matrix A (src0) and matrix B (src1)
* Matrix A and matrix B must be reshaped respectively with @ref gemm_interleave4x4_8bit and @ref gemm_transpose1x16 before running the matrix multiplication
*
* @attention The number of matrix B columns needs to be passed at compile time using -DCOLS_B
*
* @param[in] src0_ptr Pointer to the source matrix. Supported data type: QASYMM8
* @param[in] src0_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src0_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src0_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src0_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src0_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[in] src1_ptr Pointer to the source matrix. Supported data type: same as @p src0_ptr
* @param[in] src1_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src1_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src1_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src1_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src1_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[out] dst_ptr Pointer to the destination matrix Supported data type: S32
* @param[in] dst_stride_x Stride of the destination matrix in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination matrix in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination matrix
*/
__kernel void gemmlowp_mm_interleaved_transposed(IMAGE_DECLARATION(src0),
IMAGE_DECLARATION(src1),
IMAGE_DECLARATION(dst))
{
// src_addr.s0 = address of matrix A
// src_addr.s1 = address of matrix B
// Compute address for matrix A and B
int2 src_addr = (int2)(get_global_id(1), get_global_id(0)) * (int2)((src0_stride_y),
(src1_stride_y));
// Add offset_first_element_in_bytes
src_addr = src_addr + ((int2)(src0_offset_first_element_in_bytes, src1_offset_first_element_in_bytes));
// Compute end row address for matrix B
int end_row_mtx_b = src_addr.s1 + COLS_B;
// Reset accumulators
int16 c00 = 0;
int16 c10 = 0;
int16 c20 = 0;
int16 c30 = 0;
for(; src_addr.s1 <= (end_row_mtx_b - 32); src_addr += (int2)(8, 32))
{
// Load values from matrix A (interleaved) and matrix B (transposed)
int8 a0 = convert_int8(vload8(0, ((__global uchar *)src0_ptr) + src_addr.s0));
int16 b0 = convert_int16(vload16(0, ((__global uchar *)src1_ptr) + src_addr.s1));
c00 += (int16)a0.s0 * b0;
c10 += (int16)a0.s1 * b0;
c20 += (int16)a0.s2 * b0;
c30 += (int16)a0.s3 * b0;
int16 b1 = convert_int16(vload16(0, ((__global uchar *)src1_ptr) + src_addr.s1 + 16));
c00 += (int16)a0.s4 * b1;
c10 += (int16)a0.s5 * b1;
c20 += (int16)a0.s6 * b1;
c30 += (int16)a0.s7 * b1;
}
for(; src_addr.s1 < end_row_mtx_b; src_addr += (int2)(4, 16))
{
// Load values from matrix A (interleaved) and matrix B (transposed)
int4 a0 = convert_int4(vload4(0, ((__global uchar *)src0_ptr) + src_addr.s0));
int16 b0 = convert_int16(vload16(0, ((__global uchar *)src1_ptr) + src_addr.s1));
c00 += (int16)a0.s0 * b0;
c10 += (int16)a0.s1 * b0;
c20 += (int16)a0.s2 * b0;
c30 += (int16)a0.s3 * b0;
}
// Compute destination address
Image dst = CONVERT_TO_IMAGE_STRUCT(dst);
// Store 4x16 block
vstore16(c00, 0, (__global int *)(offset(&dst, 0, 0)));
vstore16(c10, 0, (__global int *)(offset(&dst, 0, 1)));
vstore16(c20, 0, (__global int *)(offset(&dst, 0, 2)));
vstore16(c30, 0, (__global int *)(offset(&dst, 0, 3)));
}
#endif // defined(COLS_B)
#if defined(NUM_ELEMS_PROCESSED_PER_THREAD_X) && defined(NUM_ELEMS_PROCESSED_PER_THREAD_Y) && defined(COLS_A)
#define VECTOR_UCHAR VEC_DATA_TYPE(uchar, NUM_ELEMS_PROCESSED_PER_THREAD_X)
#define VECTOR_UINT VEC_DATA_TYPE(uint, NUM_ELEMS_PROCESSED_PER_THREAD_X)
#define VECTOR_INT VEC_DATA_TYPE(int, NUM_ELEMS_PROCESSED_PER_THREAD_X)
/** This OpenCL kernel computes the matrix multiplication between matrix A (src0) and matrix B (src1) in case both matrices have not beed reshaped
*
* @attention The number of matrix A columns needs to be passed at compile time using -DCOLS_A
*
* @param[in] src0_ptr Pointer to the source matrix. Supported data type: QASYMM8
* @param[in] src0_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src0_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src0_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src0_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src0_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[in] src1_ptr Pointer to the source matrix. Supported data type: same as @p src0_ptr
* @param[in] src1_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src1_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src1_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src1_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src1_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[out] dst_ptr Pointer to the destination matrix Supported data type: S32
* @param[in] dst_stride_x Stride of the destination matrix in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination matrix in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination matrix
*/
__kernel void gemmlowp_mm_midgard(IMAGE_DECLARATION(src0),
IMAGE_DECLARATION(src1),
IMAGE_DECLARATION(dst))
{
int idx = get_global_id(0) * NUM_ELEMS_PROCESSED_PER_THREAD_X;
// Compute starting address for matrix A and Matrix B
int2 src_addr = ((int2)(src0_offset_first_element_in_bytes, src1_offset_first_element_in_bytes));
// Update address for the matrix A
src_addr.s0 += get_global_id(1) * src0_stride_y * NUM_ELEMS_PROCESSED_PER_THREAD_Y;
// Update address for the matrix B
src_addr.s1 += idx;
int end_row_vec_a = src_addr.s0 + COLS_A;
VECTOR_UINT acc0 = 0;
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
VECTOR_UINT acc1 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
VECTOR_UINT acc2 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
VECTOR_UINT acc3 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
VECTOR_UINT acc4 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
for(; src_addr.s0 <= (end_row_vec_a - 2); src_addr += (int2)(2, 2 * src1_stride_y))
{
// Load values from matrix A
uchar2 a0 = vload2(0, src0_ptr + src_addr.s0 + 0 * src0_stride_y);
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
uchar2 a1 = vload2(0, src0_ptr + src_addr.s0 + 1 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
uchar2 a2 = vload2(0, src0_ptr + src_addr.s0 + 2 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
uchar2 a3 = vload2(0, src0_ptr + src_addr.s0 + 3 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
uchar2 a4 = vload2(0, src0_ptr + src_addr.s0 + 4 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
// Load values from matrix B
VECTOR_UCHAR b0 = VLOAD(NUM_ELEMS_PROCESSED_PER_THREAD_X)(0, src1_ptr + src_addr.s1);
VECTOR_UCHAR b1 = VLOAD(NUM_ELEMS_PROCESSED_PER_THREAD_X)(0, src1_ptr + src_addr.s1 + src1_stride_y);
// Accumulate
acc0 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a0.s0;
acc0 += CONVERT(b1, VECTOR_UINT) * (VECTOR_UINT)a0.s1;
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
acc1 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a1.s0;
acc1 += CONVERT(b1, VECTOR_UINT) * (VECTOR_UINT)a1.s1;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
acc2 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a2.s0;
acc2 += CONVERT(b1, VECTOR_UINT) * (VECTOR_UINT)a2.s1;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
acc3 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a3.s0;
acc3 += CONVERT(b1, VECTOR_UINT) * (VECTOR_UINT)a3.s1;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
acc4 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a4.s0;
acc4 += CONVERT(b1, VECTOR_UINT) * (VECTOR_UINT)a4.s1;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
for(; src_addr.s0 < end_row_vec_a; src_addr += (int2)(1, src1_stride_y))
{
// Load values from matrix A
uchar a0 = *(src0_ptr + src_addr.s0 + 0 * src0_stride_y);
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
uchar a1 = *(src0_ptr + src_addr.s0 + 1 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
uchar a2 = *(src0_ptr + src_addr.s0 + 2 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
uchar a3 = *(src0_ptr + src_addr.s0 + 3 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
uchar a4 = *(src0_ptr + src_addr.s0 + 4 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
// Load values from matrix B
VECTOR_UCHAR b0 = VLOAD(NUM_ELEMS_PROCESSED_PER_THREAD_X)(0, src1_ptr + src_addr.s1);
// Accumulate
acc0 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a0;
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
acc1 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a1;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
acc2 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a2;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
acc3 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a3;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
acc4 += CONVERT(b0, VECTOR_UINT) * (VECTOR_UINT)a4;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
// Compute destination address
Image dst = CONVERT_TO_IMAGE_STRUCT(dst);
// Store the result
VSTORE(NUM_ELEMS_PROCESSED_PER_THREAD_X)
(CONVERT(acc0, VECTOR_INT), 0, (__global int *)(offset(&dst, 0, 0)));
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
VSTORE(NUM_ELEMS_PROCESSED_PER_THREAD_X)
(CONVERT(acc1, VECTOR_INT), 0, (__global int *)(offset(&dst, 0, 1)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
VSTORE(NUM_ELEMS_PROCESSED_PER_THREAD_X)
(CONVERT(acc2, VECTOR_INT), 0, (__global int *)(offset(&dst, 0, 2)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
VSTORE(NUM_ELEMS_PROCESSED_PER_THREAD_X)
(CONVERT(acc3, VECTOR_INT), 0, (__global int *)(offset(&dst, 0, 3)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
VSTORE(NUM_ELEMS_PROCESSED_PER_THREAD_X)
(CONVERT(acc4, VECTOR_INT), 0, (__global int *)(offset(&dst, 0, 4)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
/** OpenCL kernel optimized for Bifrost architectures that computes the matrix multiplication between matrix A (src0) and matrix B (src1) in case both matrices have not beed reshaped
*
* @attention The number of matrix A columns needs to be passed at compile time using -DCOLS_A
*
* @param[in] src0_ptr Pointer to the source matrix. Supported data type: QASYMM8
* @param[in] src0_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src0_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src0_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src0_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src0_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[in] src1_ptr Pointer to the source matrix. Supported data type: same as @p src0_ptr
* @param[in] src1_stride_x Stride of the source matrix in X dimension (in bytes)
* @param[in] src1_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src1_stride_y Stride of the source matrix in Y dimension (in bytes)
* @param[in] src1_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src1_offset_first_element_in_bytes The offset of the first element in the source matrix
* @param[out] dst_ptr Pointer to the destination matrix Supported data type: S32
* @param[in] dst_stride_x Stride of the destination matrix in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination matrix in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination matrix
*/
__kernel void gemmlowp_mm_bifrost(IMAGE_DECLARATION(src0),
IMAGE_DECLARATION(src1),
IMAGE_DECLARATION(dst))
{
int idx = get_global_id(0) * NUM_ELEMS_PROCESSED_PER_THREAD_X;
// Compute starting address for matrix A and Matrix B
int2 src_addr = ((int2)(src0_offset_first_element_in_bytes, src1_offset_first_element_in_bytes));
// Update address for the matrix A
src_addr.s0 += get_global_id(1) * src0_stride_y * NUM_ELEMS_PROCESSED_PER_THREAD_Y;
// Update address for the matrix B
src_addr.s1 += idx;
int end_row_vec_a = src_addr.s0 + COLS_A;
uint acc00 = 0;
uint acc01 = 0;
uint acc02 = 0;
uint acc03 = 0;
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
uint acc10 = 0;
uint acc11 = 0;
uint acc12 = 0;
uint acc13 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
uint acc20 = 0;
uint acc21 = 0;
uint acc22 = 0;
uint acc23 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
uint acc30 = 0;
uint acc31 = 0;
uint acc32 = 0;
uint acc33 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
uint acc40 = 0;
uint acc41 = 0;
uint acc42 = 0;
uint acc43 = 0;
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
for(; src_addr.s0 <= (end_row_vec_a - 4); src_addr += (int2)(4, 4 * src1_stride_y))
{
// Load values from matrix A
uchar4 a0 = vload4(0, src0_ptr + src_addr.s0 + 0 * src0_stride_y);
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
uchar4 a1 = vload4(0, src0_ptr + src_addr.s0 + 1 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
uchar4 a2 = vload4(0, src0_ptr + src_addr.s0 + 2 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
uchar4 a3 = vload4(0, src0_ptr + src_addr.s0 + 3 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
uchar4 a4 = vload4(0, src0_ptr + src_addr.s0 + 4 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
// Load values from matrix B
uchar4 b0 = vload4(0, src1_ptr + src_addr.s1 + 0 * src1_stride_y);
uchar4 b1 = vload4(0, src1_ptr + src_addr.s1 + 1 * src1_stride_y);
uchar4 b2 = vload4(0, src1_ptr + src_addr.s1 + 2 * src1_stride_y);
uchar4 b3 = vload4(0, src1_ptr + src_addr.s1 + 3 * src1_stride_y);
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a0.s0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a0.s0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a0.s0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a0.s0;
ushort tmp4 = (ushort)b1.s0 * (ushort)a0.s1;
ushort tmp5 = (ushort)b1.s1 * (ushort)a0.s1;
ushort tmp6 = (ushort)b1.s2 * (ushort)a0.s1;
ushort tmp7 = (ushort)b1.s3 * (ushort)a0.s1;
ushort tmp8 = (ushort)b2.s0 * (ushort)a0.s2;
ushort tmp9 = (ushort)b2.s1 * (ushort)a0.s2;
ushort tmpA = (ushort)b2.s2 * (ushort)a0.s2;
ushort tmpB = (ushort)b2.s3 * (ushort)a0.s2;
ushort tmpC = (ushort)b3.s0 * (ushort)a0.s3;
ushort tmpD = (ushort)b3.s1 * (ushort)a0.s3;
ushort tmpE = (ushort)b3.s2 * (ushort)a0.s3;
ushort tmpF = (ushort)b3.s3 * (ushort)a0.s3;
acc00 += ((uint)tmp0 + (uint)tmp4 + (uint)tmp8 + (uint)tmpC);
acc01 += ((uint)tmp1 + (uint)tmp5 + (uint)tmp9 + (uint)tmpD);
acc02 += ((uint)tmp2 + (uint)tmp6 + (uint)tmpA + (uint)tmpE);
acc03 += ((uint)tmp3 + (uint)tmp7 + (uint)tmpB + (uint)tmpF);
}
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a1.s0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a1.s0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a1.s0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a1.s0;
ushort tmp4 = (ushort)b1.s0 * (ushort)a1.s1;
ushort tmp5 = (ushort)b1.s1 * (ushort)a1.s1;
ushort tmp6 = (ushort)b1.s2 * (ushort)a1.s1;
ushort tmp7 = (ushort)b1.s3 * (ushort)a1.s1;
ushort tmp8 = (ushort)b2.s0 * (ushort)a1.s2;
ushort tmp9 = (ushort)b2.s1 * (ushort)a1.s2;
ushort tmpA = (ushort)b2.s2 * (ushort)a1.s2;
ushort tmpB = (ushort)b2.s3 * (ushort)a1.s2;
ushort tmpC = (ushort)b3.s0 * (ushort)a1.s3;
ushort tmpD = (ushort)b3.s1 * (ushort)a1.s3;
ushort tmpE = (ushort)b3.s2 * (ushort)a1.s3;
ushort tmpF = (ushort)b3.s3 * (ushort)a1.s3;
acc10 += ((uint)tmp0 + (uint)tmp4 + (uint)tmp8 + (uint)tmpC);
acc11 += ((uint)tmp1 + (uint)tmp5 + (uint)tmp9 + (uint)tmpD);
acc12 += ((uint)tmp2 + (uint)tmp6 + (uint)tmpA + (uint)tmpE);
acc13 += ((uint)tmp3 + (uint)tmp7 + (uint)tmpB + (uint)tmpF);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a2.s0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a2.s0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a2.s0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a2.s0;
ushort tmp4 = (ushort)b1.s0 * (ushort)a2.s1;
ushort tmp5 = (ushort)b1.s1 * (ushort)a2.s1;
ushort tmp6 = (ushort)b1.s2 * (ushort)a2.s1;
ushort tmp7 = (ushort)b1.s3 * (ushort)a2.s1;
ushort tmp8 = (ushort)b2.s0 * (ushort)a2.s2;
ushort tmp9 = (ushort)b2.s1 * (ushort)a2.s2;
ushort tmpA = (ushort)b2.s2 * (ushort)a2.s2;
ushort tmpB = (ushort)b2.s3 * (ushort)a2.s2;
ushort tmpC = (ushort)b3.s0 * (ushort)a2.s3;
ushort tmpD = (ushort)b3.s1 * (ushort)a2.s3;
ushort tmpE = (ushort)b3.s2 * (ushort)a2.s3;
ushort tmpF = (ushort)b3.s3 * (ushort)a2.s3;
acc20 += ((uint)tmp0 + (uint)tmp4 + (uint)tmp8 + (uint)tmpC);
acc21 += ((uint)tmp1 + (uint)tmp5 + (uint)tmp9 + (uint)tmpD);
acc22 += ((uint)tmp2 + (uint)tmp6 + (uint)tmpA + (uint)tmpE);
acc23 += ((uint)tmp3 + (uint)tmp7 + (uint)tmpB + (uint)tmpF);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a3.s0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a3.s0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a3.s0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a3.s0;
ushort tmp4 = (ushort)b1.s0 * (ushort)a3.s1;
ushort tmp5 = (ushort)b1.s1 * (ushort)a3.s1;
ushort tmp6 = (ushort)b1.s2 * (ushort)a3.s1;
ushort tmp7 = (ushort)b1.s3 * (ushort)a3.s1;
ushort tmp8 = (ushort)b2.s0 * (ushort)a3.s2;
ushort tmp9 = (ushort)b2.s1 * (ushort)a3.s2;
ushort tmpA = (ushort)b2.s2 * (ushort)a3.s2;
ushort tmpB = (ushort)b2.s3 * (ushort)a3.s2;
ushort tmpC = (ushort)b3.s0 * (ushort)a3.s3;
ushort tmpD = (ushort)b3.s1 * (ushort)a3.s3;
ushort tmpE = (ushort)b3.s2 * (ushort)a3.s3;
ushort tmpF = (ushort)b3.s3 * (ushort)a3.s3;
acc30 += ((uint)tmp0 + (uint)tmp4 + (uint)tmp8 + (uint)tmpC);
acc31 += ((uint)tmp1 + (uint)tmp5 + (uint)tmp9 + (uint)tmpD);
acc32 += ((uint)tmp2 + (uint)tmp6 + (uint)tmpA + (uint)tmpE);
acc33 += ((uint)tmp3 + (uint)tmp7 + (uint)tmpB + (uint)tmpF);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a4.s0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a4.s0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a4.s0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a4.s0;
ushort tmp4 = (ushort)b1.s0 * (ushort)a4.s1;
ushort tmp5 = (ushort)b1.s1 * (ushort)a4.s1;
ushort tmp6 = (ushort)b1.s2 * (ushort)a4.s1;
ushort tmp7 = (ushort)b1.s3 * (ushort)a4.s1;
ushort tmp8 = (ushort)b2.s0 * (ushort)a4.s2;
ushort tmp9 = (ushort)b2.s1 * (ushort)a4.s2;
ushort tmpA = (ushort)b2.s2 * (ushort)a4.s2;
ushort tmpB = (ushort)b2.s3 * (ushort)a4.s2;
ushort tmpC = (ushort)b3.s0 * (ushort)a4.s3;
ushort tmpD = (ushort)b3.s1 * (ushort)a4.s3;
ushort tmpE = (ushort)b3.s2 * (ushort)a4.s3;
ushort tmpF = (ushort)b3.s3 * (ushort)a4.s3;
acc40 += ((uint)tmp0 + (uint)tmp4 + (uint)tmp8 + (uint)tmpC);
acc41 += ((uint)tmp1 + (uint)tmp5 + (uint)tmp9 + (uint)tmpD);
acc42 += ((uint)tmp2 + (uint)tmp6 + (uint)tmpA + (uint)tmpE);
acc43 += ((uint)tmp3 + (uint)tmp7 + (uint)tmpB + (uint)tmpF);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
for(; src_addr.s0 < end_row_vec_a; src_addr += (int2)(1, src1_stride_y))
{
// Load values from matrix A
uchar a0 = *(src0_ptr + src_addr.s0 + 0 * src0_stride_y);
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
uchar a1 = *(src0_ptr + src_addr.s0 + 1 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
uchar a2 = *(src0_ptr + src_addr.s0 + 2 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
uchar a3 = *(src0_ptr + src_addr.s0 + 3 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
uchar a4 = *(src0_ptr + src_addr.s0 + 4 * src0_stride_y);
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
// Load values from matrix B
uchar4 b0 = vload4(0, src1_ptr + src_addr.s1);
// Accumulate
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a0;
ushort tmp1 = (ushort)b0.s1 * (ushort)a0;
ushort tmp2 = (ushort)b0.s2 * (ushort)a0;
ushort tmp3 = (ushort)b0.s3 * (ushort)a0;
acc00 += ((uint)tmp0);
acc01 += ((uint)tmp1);
acc02 += ((uint)tmp2);
acc03 += ((uint)tmp3);
}
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a1;
ushort tmp1 = (ushort)b0.s1 * (ushort)a1;
ushort tmp2 = (ushort)b0.s2 * (ushort)a1;
ushort tmp3 = (ushort)b0.s3 * (ushort)a1;
acc10 += ((uint)tmp0);
acc11 += ((uint)tmp1);
acc12 += ((uint)tmp2);
acc13 += ((uint)tmp3);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a2;
ushort tmp1 = (ushort)b0.s1 * (ushort)a2;
ushort tmp2 = (ushort)b0.s2 * (ushort)a2;
ushort tmp3 = (ushort)b0.s3 * (ushort)a2;
acc20 += ((uint)tmp0);
acc21 += ((uint)tmp1);
acc22 += ((uint)tmp2);
acc23 += ((uint)tmp3);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a3;
ushort tmp1 = (ushort)b0.s1 * (ushort)a3;
ushort tmp2 = (ushort)b0.s2 * (ushort)a3;
ushort tmp3 = (ushort)b0.s3 * (ushort)a3;
acc30 += ((uint)tmp0);
acc31 += ((uint)tmp1);
acc32 += ((uint)tmp2);
acc33 += ((uint)tmp3);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
{
// Accumulate
ushort tmp0 = (ushort)b0.s0 * (ushort)a4;
ushort tmp1 = (ushort)b0.s1 * (ushort)a4;
ushort tmp2 = (ushort)b0.s2 * (ushort)a4;
ushort tmp3 = (ushort)b0.s3 * (ushort)a4;
acc40 += ((uint)tmp0);
acc41 += ((uint)tmp1);
acc42 += ((uint)tmp2);
acc43 += ((uint)tmp3);
}
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
// Compute destination address
Image dst = CONVERT_TO_IMAGE_STRUCT(dst);
// Store the result
vstore4((int4)(acc00, acc01, acc02, acc03), 0, (__global int *)(offset(&dst, 0, 0)));
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
vstore4((int4)(acc10, acc11, acc12, acc13), 0, (__global int *)(offset(&dst, 0, 1)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 1
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
vstore4((int4)(acc20, acc21, acc22, acc23), 0, (__global int *)(offset(&dst, 0, 2)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 2
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
vstore4((int4)(acc30, acc31, acc32, acc33), 0, (__global int *)(offset(&dst, 0, 3)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 3
#if NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
vstore4((int4)(acc40, acc41, acc42, acc43), 0, (__global int *)(offset(&dst, 0, 4)));
#endif // NUM_ELEMS_PROCESSED_PER_THREAD_Y > 4
}
#endif // defined(NUM_ELEMS_PROCESSED_PER_THREAD_X) && defined(NUM_ELEMS_PROCESSED_PER_THREAD_Y) && defined(COLS_A)
#if defined(COLS_A)
/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each row of Matrix A.
*
* @note This stage is needed to handle the offset of matrix product
* https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
*
* @attention The number of matrix A columns needs to be passed at compile time using -DCOLS_A
*
* @param[in] src_ptr Pointer to the source tensor. Supported data type: QASYMM8
* @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] src_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[out] dst_ptr Pointer to the destination tensor Supported data type: S32
* @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination tensor in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor
*/
__kernel void gemmlowp_matrix_a_reduction(TENSOR3D_DECLARATION(src),
IMAGE_DECLARATION(dst))
{
// Compute source and destination addresses
Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);
Image dst = CONVERT_TO_IMAGE_STRUCT(dst);
uint4 sum_row_u32 = (uint4)0;
uint sum_row = 0;
__global const uchar *matrix_a = (__global const uchar *)(src.ptr + get_global_id(0) * src_stride_y + get_global_id(1) * src_stride_z);
int i = 0;
// This for loop performs 16 accumulations
for(; i <= ((int)COLS_A - 16); i += 16)
{
const uchar16 a0_u8 = vload16(0, matrix_a + i);
sum_row_u32 += convert_uint4(a0_u8.s0123) + convert_uint4(a0_u8.s4567) + convert_uint4(a0_u8.s89AB) + convert_uint4(a0_u8.sCDEF);
}
// This for loop performs the leftover accumulations
for(; i < COLS_A; ++i)
{
sum_row += matrix_a[i];
}
sum_row += sum_row_u32.s0 + sum_row_u32.s1 + sum_row_u32.s2 + sum_row_u32.s3;
*((__global int *)dst.ptr) = (int)sum_row;
}
#endif // defined(COLS_A)
#if defined(COLS_B) && defined(ROWS_B)
/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each column of Matrix B.
*
* @note This stage is needed to handle the offset of matrix product
* https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
*
* @attention The number of matrix B columns and rows needs to be passed at compile time using -DCOLS_B and -DROWS_B
*
* @param[in] src_ptr Pointer to the source tensor. Supported data type: QASYMM8
* @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] src_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[out] dst_ptr Pointer to the destination tensor Supported data type: S32
* @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination tensor in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor
*/
__kernel void gemmlowp_matrix_b_reduction(TENSOR3D_DECLARATION(src),
IMAGE_DECLARATION(dst))
{
// Compute source and destination addresses
Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);
Image dst = CONVERT_TO_IMAGE_STRUCT(dst);
uint16 sum_col_u32 = (uint16)0;
__global const uchar *matrix_b = (__global const uchar *)(src.ptr + get_global_id(1) * src_stride_z);
int i = 0;
// This for loop performs 4 accumulations
for(; i <= ((int)ROWS_B - 4); i += 4)
{
const uchar16 b0_u8 = vload16(0, matrix_b + 0 * src_stride_y);
const uchar16 b1_u8 = vload16(0, matrix_b + 1 * src_stride_y);
const uchar16 b2_u8 = vload16(0, matrix_b + 2 * src_stride_y);
const uchar16 b3_u8 = vload16(0, matrix_b + 3 * src_stride_y);
sum_col_u32 += convert_uint16(b0_u8) + convert_uint16(b1_u8) + convert_uint16(b2_u8) + convert_uint16(b3_u8);
matrix_b += 4 * src_stride_y;
}
// This for loop perfoms the leftover accumulations
for(; i < (int)ROWS_B; ++i)
{
const uchar16 b0_u8 = vload16(0, matrix_b);
sum_col_u32 += convert_uint16(b0_u8);
matrix_b += src_stride_y;
}
vstore16(convert_int16(sum_col_u32), 0, (__global int *)dst.ptr);
}
#endif // defined(COLS_B) && defined(ROWS_B)
#if defined(K_OFFSET)
/* OpenCL kernel used to add the offset contribution after @ref CLGEMMLowpMatrixMultiplyKernel. The computation is performed in-place
*
* This kernel takes a final int32 accumulator value (the output of @CLGEMMLowpMatrixMultiplyKernel),
* and adds to it the offset contribution of matrix A and matrix B in-place.
*
* @attention The k_offset = a_offset * b_offset * k (where k is the number of matrix A columns) needs to be passed at compile time using -DK_OFFSET (i.e. -DK_OFFSET=1200)
* @note In case the offset contribution due to a_offset is required, a_offset needs to be passed at compile time using -DA_OFFSET (i.e. -DA_OFFSET=1)
* @note In case the offset contribution due to b_offset is required, b_offset needs to be passed at compile time using -DB_OFFSET (i.e. -DB_OFFSET=6)
* @note In case sum_col has batches, -DSUM_COL_HAS_BATCHES must be passed at compile time. Usually if gemmlowp is used to accelerate convolution layer, sum_col will not have batches
*
* The final result is:
*
* mm_result[i][k] = mm_result[i][k] +
* (sum_col[k] * A_OFFSET) +
* (sum_row[i] * B_OFFSET) +
* (K_OFFSET)
*
* @param[in] mm_result_ptr Pointer to the source tensor. Supported data type: S32
* @param[in] mm_result_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] mm_result_step_x mm_result_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] mm_result_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] mm_result_step_y mm_result_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] mm_result_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] mm_result_step_z mm_result_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] mm_result_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[in] sum_col_result_ptr Pointer to the source tensor. Supported data type: same as @p mm_result_ptr
* @param[in] sum_col_result_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] sum_col_result_step_x sum_col_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] sum_col_result_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] sum_col_result_step_y sum_col_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] sum_col_result_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[in] sum_row_result_ptr Pointer to the source tensor. Supported data type: same as @p mm_result_ptr
* @param[in] sum_row_result_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] sum_row_result_step_x sum_row_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] sum_row_result_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] sum_row_result_step_y sum_row_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] sum_row_result_offset_first_element_in_bytes The offset of the first element in the source tensor
*/
__kernel void gemmlowp_offset_contribution(TENSOR3D_DECLARATION(mm_result)
#if defined(A_OFFSET)
,
IMAGE_DECLARATION(sum_col)
#endif // defined(A_OFFSET)
#if defined(B_OFFSET)
,
IMAGE_DECLARATION(sum_row)
#endif // defined(B_OFFSET)
)
{
Tensor3D mm_result = CONVERT_TO_TENSOR3D_STRUCT(mm_result);
int16 a_offset_s32 = (int16)0;
int16 b_offset_s32 = (int16)0;
#if defined(A_OFFSET)
Image sum_col = CONVERT_TO_IMAGE_STRUCT(sum_col);
// Compute the offset contribution due to A_OFFSET
#if defined(SUM_COL_HAS_BATCHES)
a_offset_s32 = vload16(0, (__global int *)(sum_col.ptr + get_global_id(2) * sum_col_stride_y));
#else // defined(MATRIX_B_HAS_BATCHES)
a_offset_s32 = vload16(0, (__global int *)(sum_col.ptr));
#endif // defined(MATRIX_B_HAS_BATCHES)
a_offset_s32 *= (int16)A_OFFSET;
#endif // defined(A_OFFSET)
#if defined(B_OFFSET)
Image sum_row = CONVERT_TO_IMAGE_STRUCT(sum_row);
// Compute the offset contribution due to B_OFFSET
b_offset_s32 = (int16) * (((__global int *)(sum_row.ptr + get_global_id(2) * sum_row_stride_y)) + get_global_id(1));
b_offset_s32 *= (int16)B_OFFSET;
#endif // defined(B_OFFSET)
const int16 offset_term_s32 = (int16)K_OFFSET + a_offset_s32 + b_offset_s32;
int16 in_s32 = vload16(0, (__global int *)mm_result.ptr);
// Add the offset terms to GEMM's result
in_s32 += offset_term_s32;
// Store the result with the offset contribution
vstore16(in_s32, 0, (__global int *)mm_result.ptr);
}
#endif // defined(K_OFFSET)
#if defined(RESULT_OFFSET) && defined(RESULT_MULT_INT) && defined(RESULT_SHIFT)
/** This OpenCL kernel is used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8
*
* This kernel takes a final int32 accumulator value and processes it to obtain the final QASYMM8 value.
* The following computations will be performed by the kernel:
*
* -# Add offset terms to final result
* -# Multiply each entry of result by result_mult_int
* -# Add bias to final result (if -DADD_BIAS is passed at compile time)
* -# Shift the int32 accumulator by result_shift
* -# Clamp the value between the specified min and max bounds (if -DMIN_BOUND and/or -DMAX_BOUND are passed at compile time)
* -# Clamp the resulting int32 values to the [0..255] range and cast to QASYMM8.
*
* @attention The offset, scalar scale factor and number of bits to shift right of output tensor must be passed at compile time using -DRESULT_OFFSET, -RESULT_MULT_INT and -DRESULT_SHIFT
*
* @note In case the addition of int32 biases is required, -DADD_BIAS should be passed at compile time
* @note In case the clamping of the result is required, the min and max bounds can be passed at compile time using -DMIN_BOUND and -DMAX_BOUND.
* These values can be used to implement "rectified linear unit" activation functions
*
* @param[in] src_ptr Pointer to the source tensor. Supported data type: S32
* @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] src_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[in] biases_ptr Pointer to the biases tensor. Supported data type: same as @p src_ptr
* @param[in] biases_stride_x Stride of the biases tensor in X dimension (in bytes)
* @param[in] biases_step_x biases_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] biases_offset_first_element_in_bytes The offset of the first element in the biases tensor
* @param[out] dst_ptr Pointer to the destination tensor Supported data type: QASYMM8
* @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination tensor in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] dst_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor
*/
__kernel void gemmlowp_output_stage_quantize_down(TENSOR3D_DECLARATION(src),
#if defined(ADD_BIAS)
VECTOR_DECLARATION(biases),
#endif // defined(ADD_BIAS)
TENSOR3D_DECLARATION(dst))
{
// Compute source and destination addresses
Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);
Tensor3D dst = CONVERT_TO_TENSOR3D_STRUCT(dst);
#if defined(ADD_BIAS)
Vector biases = CONVERT_TO_VECTOR_STRUCT(biases);
#endif // defined(ADD_BIAS)
int16 input_values = vload16(0, (__global int *)src.ptr);
// Add the offset terms to GEMM's result
input_values += (int16)RESULT_OFFSET;
#if defined(ADD_BIAS)
// Add bias
const int16 biases_values = vload16(0, (__global int *)biases.ptr);
input_values += (int16)biases_values;
#endif // defined(ADD_BIAS)
// Multiply by result_mult_int and shift
input_values *= RESULT_MULT_INT;
input_values >>= RESULT_SHIFT;
uchar16 res = convert_uchar16_sat(input_values);
#if defined(MIN_BOUND)
res = max(res, (uchar16)MIN_BOUND);
#endif // defined(MIN_BOUND)
#if defined(MAX_BOUND)
res = min(res, (uchar16)MAX_BOUND);
#endif // defined(MAX_BOUND)
// Store the result
vstore16(res, 0, dst.ptr);
}
#endif // defined(RESULT_OFFSET) && defined(RESULT_MULT_INT) && defined(RESULT_SHIFT)
#if defined(RESULT_OFFSET_AFTER_SHIFT) && defined(RESULT_FIXEDPOINT_MULTIPLIER) && defined(RESULT_SHIFT)
/** This OpenCL kernel is used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8
*
* This kernel takes a final int32 accumulator value (the output of @ref CLGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8 value.
* The following computations will be performed by the kernel:
*
* -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier
* -# Add bias to final result if bias tensor is not a nullptr
* -# Round to nearest division by a power-of-two using result_shift
* -# Add offset to each result
* -# Clamp the value between the specified min and max bounds
* -# Clamp the resulting int32 values to the [0..255] range and cast to QASYMM8.
*
* @attention The offset, scalar scale factor and number of bits to shift right of output tensor must be passed at compile time using -DRESULT_OFFSET, -RESULT_MULT_INT and -DRESULT_SHIFT
*
* @note In case the addition of int32 biases is required, -DADD_BIAS should be passed at compile time
* @note In case the clamping of the result is required, the min and max bounds can be passed at compile time using -DMIN_BOUND and -DMAX_BOUND.
* These values can be used to implement "rectified linear unit" activation functions
*
* @param[in] src_ptr Pointer to the source tensor. Supported data type: S32
* @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)
* @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)
* @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] src_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor
* @param[in] biases_ptr Pointer to the biases tensor. Supported data type: same as @p src_ptr
* @param[in] biases_stride_x Stride of the biases tensor in X dimension (in bytes)
* @param[in] biases_step_x biases_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] biases_offset_first_element_in_bytes The offset of the first element in the biases tensor
* @param[out] dst_ptr Pointer to the destination tensor Supported data type: QASYMM8
* @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)
* @param[in] dst_step_x dst_gx_stride_x * number of elements along X processed per workitem(in bytes)
* @param[in] dst_stride_y Stride of the destination tensor in Y dimension (in bytes)
* @param[in] dst_step_y dst_gx_stride_y * number of elements along Y processed per workitem(in bytes)
* @param[in] dst_stride_z Stride of the source tensor in Z dimension (in bytes)
* @param[in] dst_step_z src_stride_z * number of elements along Z processed per workitem(in bytes)
* @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor
*/
__kernel void gemmlowp_output_stage_quantize_down_fixedpoint(TENSOR3D_DECLARATION(src),
#if defined(ADD_BIAS)
VECTOR_DECLARATION(biases),
#endif // defined(ADD_BIAS)
TENSOR3D_DECLARATION(dst))
{
// Compute source and destination addresses
Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);
Tensor3D dst = CONVERT_TO_TENSOR3D_STRUCT(dst);
#if defined(ADD_BIAS)
Vector biases = CONVERT_TO_VECTOR_STRUCT(biases);
#endif // defined(ADD_BIAS)
int16 input_values = vload16(0, (__global int *)src.ptr);
#if defined(ADD_BIAS)
// Add bias
const int16 biases_values = vload16(0, (__global int *)biases.ptr);
input_values += (int16)biases_values;
#endif // defined(ADD_BIAS)
// Multiply by result_mult_int and shift
input_values = ASYMM_MULT_BY_QUANT_MULTIPLIER_LESS_THAN_ONE(input_values, RESULT_FIXEDPOINT_MULTIPLIER, RESULT_SHIFT, 16);
// Add the offset terms to GEMM's result
input_values += (int16)RESULT_OFFSET_AFTER_SHIFT;
uchar16 res = convert_uchar16_sat(input_values);
#if defined(MIN_BOUND)
res = max(res, (uchar16)MIN_BOUND);
#endif // defined(MIN_BOUND)
#if defined(MAX_BOUND)
res = min(res, (uchar16)MAX_BOUND);
#endif // defined(MAX_BOUND)
// Store the result
vstore16(res, 0, dst.ptr);
}
#endif // defined(RESULT_OFFSET_AFTER_SHIFT) && defined(RESULT_FIXEDPOINT_MULTIPLIER) && defined(RESULT_SHIFT)
|
