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
|
/*========================== begin_copyright_notice ============================
Copyright (C) 2017-2021 Intel Corporation
SPDX-License-Identifier: MIT
============================= end_copyright_notice ===========================*/
//===- IBiF_Sub_Groups.cl -===============================================//
//
// This file defines OpenCL sub groups functions.
// Many of these functions are part of the KHR cl_khr_subgroups extension.
// Some of these functions are part of the Intel cl_intel_subgroups Extension.
//
//===----------------------------------------------------------------------===//
// Private helper functions:
uint __intel_get_local_size( void );
uint __intel_get_enqueued_local_size( void );
uint __intel_get_local_linear_id( void );
int OVERLOADABLE sub_group_reduce_add( int );
INLINE uint OVERLOADABLE get_max_sub_group_size( void )
{
return __builtin_IB_get_simd_size();
}
INLINE uint OVERLOADABLE get_num_sub_groups( void )
{
uint totalWorkGroupSize =
__intel_get_local_size() +
get_max_sub_group_size() - 1;
return totalWorkGroupSize / get_max_sub_group_size();
}
INLINE uint OVERLOADABLE get_sub_group_id( void )
{
return __intel_get_local_linear_id() / get_max_sub_group_size();
}
INLINE uint OVERLOADABLE get_sub_group_size( void )
{
uint remainder =
__intel_get_local_size() & ( get_max_sub_group_size() - 1 );
bool fullSubGroup =
( remainder == 0 ) ||
( get_sub_group_id() < get_num_sub_groups() - 1 );
return fullSubGroup ? get_max_sub_group_size() : remainder;
}
INLINE uint OVERLOADABLE get_enqueued_num_sub_groups( void )
{
uint totalEnqueuedWorkGroupSize =
__intel_get_enqueued_local_size() +
get_max_sub_group_size() - 1;
return totalEnqueuedWorkGroupSize / get_max_sub_group_size();
}
INLINE uint OVERLOADABLE get_sub_group_local_id( void )
{
return __builtin_IB_get_simd_id();
}
INLINE void OVERLOADABLE sub_group_barrier( cl_mem_fence_flags flags )
{
SPIRV_BUILTIN(ControlBarrier, _i32_i32_i32, )( Subgroup, 0, AcquireRelease );
}
#if(__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
INLINE void OVERLOADABLE sub_group_barrier( cl_mem_fence_flags flags, memory_scope scope )
{
SPIRV_BUILTIN(ControlBarrier, _i32_i32_i32, )( Subgroup, 0, AcquireRelease );
}
#endif // __OPENCL_C_VERSION__ >= CL_VERSION_2_0
// gentype intel_sub_group_shuffle(gentype data, uint sub_group_local_id )
//
// Allows data to be arbitrarily transferred between work items in a subgroup.
// The data that is returned for this work item is the value of data for the work item
// identified by sub_group_local_id.
#define DEFN_INTEL_SUB_GROUP_SHUFFLE(TYPE, SPIRV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE intel_sub_group_shuffle( TYPE x, uint c ) \
{ \
return SPIRV_BUILTIN(SubgroupShuffleINTEL, _##TYPE_ABBR##_i32, )( as_##SPIRV_TYPE(x), c ); \
}
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE(char, char, i8)
DEFN_INTEL_SUB_GROUP_SHUFFLE(uchar, char, i8)
#endif // cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE(short, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE(ushort, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE(int, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE(uint, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE(long, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE(ulong, long, i64)
#ifdef cl_khr_fp16
DEFN_INTEL_SUB_GROUP_SHUFFLE(half, half, f16)
#endif // cl_khr_fp16
DEFN_INTEL_SUB_GROUP_SHUFFLE(float, float, f32)
#ifdef cl_khr_fp64
DEFN_INTEL_SUB_GROUP_SHUFFLE(double, double, f64)
#endif // cl_khr_fp64
#ifdef cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, char, char, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, uchar, uchar, uint )
#endif // cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, short, short, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, ushort, ushort, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, int, int, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, uint, uint, uint )
#ifdef cl_khr_fp16
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, half, half, uint )
#endif // cl_khr_fp16
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, float, float, uint )
// gentype intel_sub_group_shuffle_down( gentype current, gentype next, uint delta )
//
// Allows data to be transferred from a work item in the subgroup with a higher sub_group_local_id
// down to a work item in the subgroup with a lower sub_group_local_id.
#define DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(TYPE, SPIRV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE intel_sub_group_shuffle_down(TYPE cur, TYPE next, uint c) \
{ \
return SPIRV_BUILTIN(SubgroupShuffleDownINTEL, _##TYPE_ABBR##_##TYPE_ABBR##_i32, )(as_##SPIRV_TYPE(cur), as_##SPIRV_TYPE(next), c); \
}
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(char, char, i8)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(uchar, char, i8)
#endif // cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(short, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(ushort, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(int, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(uint, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(long, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(ulong, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(float, float, f32)
#ifdef cl_khr_fp64
DEFN_INTEL_SUB_GROUP_SHUFFLE_DOWN(double, double, f64)
#endif // cl_khr_fp64
#ifdef cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, char, char, uint)
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, uchar, uchar, uint)
#endif // cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, short, short, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, ushort, ushort, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, int, int, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, uint, uint, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_down, float, float, uint )
// gentype intel_sub_group_shuffle_up(gentype previous, gentype current, uint delta )
//
// Allows data to be transferred from a work item in the subgroup with a lower sub_group_local_id
// up to a work item in the subgroup with a higher sub_group_local_id.
#define DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(TYPE, SPIRV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE intel_sub_group_shuffle_up( TYPE prev, TYPE cur, uint c ) \
{ \
return SPIRV_BUILTIN(SubgroupShuffleUpINTEL, _##TYPE_ABBR##_##TYPE_ABBR##_i32, )( as_##SPIRV_TYPE(prev), as_##SPIRV_TYPE(cur), c); \
}
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(char, char, i8)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(uchar, char, i8)
#endif // cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(short, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(ushort, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(int, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(uint, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(long, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(ulong, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(float, float, f32)
#ifdef cl_khr_fp64
DEFN_INTEL_SUB_GROUP_SHUFFLE_UP(double, double, f64)
#endif // cl_khr_fp64
#ifdef cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, char, char, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, uchar, uchar, uint )
#endif // cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, short, short, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, ushort, ushort, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, int, int, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, uint, uint, uint )
GENERATE_VECTOR_FUNCTIONS_3ARGS_VVS( intel_sub_group_shuffle_up, float, float, uint )
// gentype intel_sub_group_shuffle_xor( gentype data, uint value )
//
// The data that is returned for this work item is the value of data for the work item with
// sub_group_local_id equal to this work item's sub_group_local_id XOR'd with the specified value.
#define DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(TYPE, SPIRV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE intel_sub_group_shuffle_xor( TYPE x, uint c ) \
{ \
return SPIRV_BUILTIN(SubgroupShuffleXorINTEL, _##TYPE_ABBR##_i32, )( as_##SPIRV_TYPE(x), c ); \
}
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(char, char, i8)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(uchar, char, i8)
#endif // cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(short, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(ushort, short, i16)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(int, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(uint, int, i32)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(long, long, i64)
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(ulong, long, i64)
#ifdef cl_khr_fp16
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(half, half, f16)
#endif // cl_khr_fp16
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(float, float, f32)
#ifdef cl_khr_fp64
DEFN_INTEL_SUB_GROUP_SHUFFLE_XOR(double, double, f64)
#endif // cl_khr_fp64
#ifdef cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, char, char, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, uchar, uchar, uint )
#endif // cl_intel_subgroups_char
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, short, short, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, ushort, ushort, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, int, int, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, uint, uint, uint )
#ifdef cl_khr_fp16
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, half, half, uint)
#endif // cl_khr_fp16
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle_xor, float, float, uint )
#ifdef cl_khr_fp16
#define SHUFFLE_AS_INTEGER_2_1(FNAME)\
INLINE half OVERLOADABLE FNAME( half x, half y, uint c )\
{\
return as_half( FNAME( as_ushort(x), as_ushort(y), c ) );\
}
#define SHUFFLE_AS_INTEGER_2_2(FNAME)\
INLINE half2 OVERLOADABLE FNAME( half2 x, half2 y, uint c )\
{\
return as_half2( FNAME( as_uint(x), as_uint(y), c ) );\
}
#define SHUFFLE_AS_INTEGER_2_4(FNAME)\
INLINE half4 OVERLOADABLE FNAME( half4 x, half4 y, uint c )\
{\
return as_half4( FNAME( as_uint2(x), as_uint2(y), c ) );\
}
#define SHUFFLE_AS_INTEGER_2_8(FNAME)\
INLINE half8 OVERLOADABLE FNAME( half8 x, half8 y, uint c )\
{\
return as_half8( FNAME( as_uint4(x), as_uint4(y), c ) );\
}
#define SHUFFLE_AS_INTEGER_2_16(FNAME)\
INLINE half16 OVERLOADABLE FNAME( half16 x, half16 y, uint c )\
{\
return as_half16( FNAME( as_uint8(x), as_uint8(y), c ) );\
}
#define OVERLOAD_AS_INTEGERS_2(FNAME)\
SHUFFLE_AS_INTEGER_2_1(FNAME)\
SHUFFLE_AS_INTEGER_2_2(FNAME)\
SHUFFLE_AS_INTEGER_2_4(FNAME)\
SHUFFLE_AS_INTEGER_2_8(FNAME)\
SHUFFLE_AS_INTEGER_2_16(FNAME)
OVERLOAD_AS_INTEGERS_2(intel_sub_group_shuffle_down)
OVERLOAD_AS_INTEGERS_2(intel_sub_group_shuffle_up)
#undef OVERLOAD_AS_INTEGERS_2
#undef SHUFFLE_AS_INTEGER_2_1
#undef SHUFFLE_AS_INTEGER_2_2
#undef SHUFFLE_AS_INTEGER_2_4
#undef SHUFFLE_AS_INTEGER_2_8
#undef SHUFFLE_AS_INTEGER_2_16
#endif // cl_khr_fp16
#if defined(_metal_simdgroups)
// 8-bit types implementation for shuffle
INLINE uchar OVERLOADABLE intel_sub_group_shuffle( uchar x, uint c )
{
return (uchar)(intel_sub_group_shuffle((ushort)(x),c));
}
INLINE char OVERLOADABLE intel_sub_group_shuffle( char x, uint c )
{
return as_char(intel_sub_group_shuffle(as_uchar(x),c));
}
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, char, char, uint )
GENERATE_VECTOR_FUNCTIONS_2ARGS_VS( intel_sub_group_shuffle, uchar, uchar, uint )
#endif
#define DEFN_SUB_GROUP_BROADCAST(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE sub_group_broadcast( TYPE x, uint sub_group_local_id ) \
{ \
int3 local_id = (int3)(sub_group_local_id,0,0); \
return as_##TYPE(SPIRV_BUILTIN(GroupBroadcast, _i32_##TYPE_ABBR##_v3i32, )(Subgroup,as_##SPV_TYPE(x),local_id)); \
}
#define DEFN_INTEL_SUB_GROUP_BROADCAST(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE intel_sub_group_broadcast( TYPE x, uint sub_group_local_id ) \
{ \
int3 local_id = (int3)(sub_group_local_id,0,0); \
return SPIRV_BUILTIN(GroupBroadcast, _i32_##TYPE_ABBR##_v3i32, )(Subgroup,as_##SPV_TYPE(x),local_id); \
}
DEFN_SUB_GROUP_BROADCAST(int, int, i32)
DEFN_SUB_GROUP_BROADCAST(uint, int, i32)
DEFN_SUB_GROUP_BROADCAST(long, long, i64)
DEFN_SUB_GROUP_BROADCAST(ulong, long, i64)
DEFN_SUB_GROUP_BROADCAST(float, float, f32)
#if defined(cl_khr_fp16)
DEFN_SUB_GROUP_BROADCAST(half, half, f16)
#endif // cl_khr_fp16
#if defined(cl_khr_fp64)
DEFN_SUB_GROUP_BROADCAST(double, double, f64)
#endif // cl_khr_fp64
#if defined(cl_khr_subgroup_extended_types)
DEFN_SUB_GROUP_BROADCAST(char, char, i8)
DEFN_SUB_GROUP_BROADCAST(uchar, char, i8)
DEFN_SUB_GROUP_BROADCAST(short, short, i16)
DEFN_SUB_GROUP_BROADCAST(ushort, short, i16)
#define DEFN_SUB_GROUP_BROADCAST_VEC(TYPE, SPV_TYPE, TYPE_ABBR) \
DEFN_SUB_GROUP_BROADCAST(TYPE##2, SPV_TYPE##2, v2##TYPE_ABBR) \
DEFN_SUB_GROUP_BROADCAST(TYPE##3, SPV_TYPE##3, v3##TYPE_ABBR) \
DEFN_SUB_GROUP_BROADCAST(TYPE##4, SPV_TYPE##4, v4##TYPE_ABBR) \
DEFN_SUB_GROUP_BROADCAST(TYPE##8, SPV_TYPE##8, v8##TYPE_ABBR) \
DEFN_SUB_GROUP_BROADCAST(TYPE##16, SPV_TYPE##16, v16##TYPE_ABBR)
DEFN_SUB_GROUP_BROADCAST_VEC(char, char, i8)
DEFN_SUB_GROUP_BROADCAST_VEC(uchar, char, i8)
DEFN_SUB_GROUP_BROADCAST_VEC(short, short, i16)
DEFN_SUB_GROUP_BROADCAST_VEC(ushort, short, i16)
DEFN_SUB_GROUP_BROADCAST_VEC(int, int, i32)
DEFN_SUB_GROUP_BROADCAST_VEC(uint, int, i32)
DEFN_SUB_GROUP_BROADCAST_VEC(long, long, i64)
DEFN_SUB_GROUP_BROADCAST_VEC(ulong, long, i64)
DEFN_SUB_GROUP_BROADCAST_VEC(float, float, f32)
#if defined(cl_khr_fp16)
DEFN_SUB_GROUP_BROADCAST_VEC(half, half, f16)
#endif // defined(cl_khr_fp16)
#if defined(cl_khr_fp64)
DEFN_SUB_GROUP_BROADCAST_VEC(double, double, f64)
#endif // defined(cl_khr_fp64)
#endif
#if defined(cl_intel_subgroups_short)
DEFN_INTEL_SUB_GROUP_BROADCAST(short, short, i16)
DEFN_INTEL_SUB_GROUP_BROADCAST(ushort, short, i16)
#endif // cl_intel_subgroups_short
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_BROADCAST(char, char, i8)
DEFN_INTEL_SUB_GROUP_BROADCAST(uchar, char, i8)
#endif // cl_intel_subgroups_char
INLINE int OVERLOADABLE sub_group_all( int predicate )
{
return SPIRV_BUILTIN(GroupAll, _i32_i1, )(Subgroup, predicate);
}
INLINE int OVERLOADABLE sub_group_any( int predicate )
{
return SPIRV_BUILTIN(GroupAny, _i32_i1, )(Subgroup, predicate);
}
#if defined(cl_khr_subgroup_non_uniform_vote)
INLINE int OVERLOADABLE sub_group_elect()
{
return SPIRV_BUILTIN(GroupNonUniformElect, _i32, )(Subgroup);
}
INLINE int OVERLOADABLE sub_group_non_uniform_all(int predicate)
{
return SPIRV_BUILTIN(GroupNonUniformAll, _i32_i1, )(Subgroup, predicate);
}
INLINE int OVERLOADABLE sub_group_non_uniform_any(int predicate)
{
return SPIRV_BUILTIN(GroupNonUniformAny, _i32_i1, )(Subgroup, predicate);
}
#define DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE int OVERLOADABLE sub_group_non_uniform_all_equal(TYPE value) \
{ \
return SPIRV_BUILTIN(GroupNonUniformAllEqual, _i32_##TYPE_ABBR, )(Subgroup, as_##SPV_TYPE(value)); \
}
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(char, char, i8)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(uchar, char, i8)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(short, short, i16)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(ushort, short, i16)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(int, int, i32)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(uint, int, i32)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(long, long, i64)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(ulong, long, i64)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(float, float, f32)
#if defined(cl_khr_fp64)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(double, double, f64)
#endif // defined(cl_khr_fp64)
#if defined(cl_khr_fp16)
DEFN_SUB_GROUP_NON_UNIFORM_ALL_EQUAL(half, half, f16)
#endif // defined(cl_khr_fp16)
#endif // defined(cl_khr_subgroup_non_uniform_vote)
#if defined(cl_khr_subgroup_ballot)
#define DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE sub_group_non_uniform_broadcast(TYPE value, uint index) \
{ \
return as_##TYPE(SPIRV_BUILTIN(GroupNonUniformBroadcast, _i32_##TYPE_ABBR##_i32, )(Subgroup, as_##SPV_TYPE(value), index)); \
} \
INLINE TYPE OVERLOADABLE sub_group_broadcast_first(TYPE value) \
{ \
return as_##TYPE(SPIRV_BUILTIN(GroupNonUniformBroadcastFirst, _i32_##TYPE_ABBR, )(Subgroup, as_##SPV_TYPE(value))); \
}
#define DEFN_NON_UNIFORM_BROADCAST(TYPE, SPV_TYPE, TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE, SPV_TYPE, TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE##2, SPV_TYPE##2, v2##TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE##3, SPV_TYPE##3, v3##TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE##4, SPV_TYPE##4, v4##TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE##8, SPV_TYPE##8, v8##TYPE_ABBR) \
DEFN_NON_UNIFORM_BROADCAST_BASE(TYPE##16, SPV_TYPE##16, v16##TYPE_ABBR)
DEFN_NON_UNIFORM_BROADCAST(char, char, i8)
DEFN_NON_UNIFORM_BROADCAST(uchar, char, i8)
DEFN_NON_UNIFORM_BROADCAST(short, short, i16)
DEFN_NON_UNIFORM_BROADCAST(ushort, short, i16)
DEFN_NON_UNIFORM_BROADCAST(int, int, i32)
DEFN_NON_UNIFORM_BROADCAST(uint, int, i32)
DEFN_NON_UNIFORM_BROADCAST(long, long, i64)
DEFN_NON_UNIFORM_BROADCAST(ulong, long, i64)
DEFN_NON_UNIFORM_BROADCAST(float, float, f32)
#if defined(cl_khr_fp64)
DEFN_NON_UNIFORM_BROADCAST(double, double, f64)
#endif // defined(cl_khr_fp64)
#if defined(cl_khr_fp16)
DEFN_NON_UNIFORM_BROADCAST(half, half, f16)
#endif // defined(cl_khr_fp16)
INLINE uint4 OVERLOADABLE sub_group_ballot(int predicate)
{
return SPIRV_BUILTIN(GroupNonUniformBallot, _i32_i1, )(Subgroup, predicate);
}
INLINE int OVERLOADABLE sub_group_inverse_ballot(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformInverseBallot, _i32_v4i32, )(Subgroup, value);
}
INLINE int OVERLOADABLE sub_group_ballot_bit_extract(uint4 value, uint index)
{
return SPIRV_BUILTIN(GroupNonUniformBallotBitExtract, _i32_v4i32_i32, )(Subgroup, value, index);
}
INLINE uint OVERLOADABLE sub_group_ballot_bit_count(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformBallotBitCount, _i32_i32_v4i32, )(Subgroup, GroupOperationReduce, value);
}
INLINE uint OVERLOADABLE sub_group_ballot_inclusive_scan(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformBallotBitCount, _i32_i32_v4i32, )(Subgroup, GroupOperationInclusiveScan, value);
}
INLINE uint OVERLOADABLE sub_group_ballot_exclusive_scan(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformBallotBitCount, _i32_i32_v4i32, )(Subgroup, GroupOperationExclusiveScan, value);
}
INLINE uint OVERLOADABLE sub_group_ballot_find_lsb(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformBallotFindLSB, _i32_v4i32, )(Subgroup, value);
}
INLINE uint OVERLOADABLE sub_group_ballot_find_msb(uint4 value)
{
return SPIRV_BUILTIN(GroupNonUniformBallotFindMSB, _i32_v4i32, )(Subgroup, value);
}
INLINE uint4 OVERLOADABLE get_sub_group_eq_mask()
{
return as_uint4(SPIRV_BUILTIN_NO_OP(BuiltInSubgroupEqMask, , )());
}
INLINE uint4 OVERLOADABLE get_sub_group_ge_mask()
{
return as_uint4(SPIRV_BUILTIN_NO_OP(BuiltInSubgroupGeMask, , )());
}
INLINE uint4 OVERLOADABLE get_sub_group_gt_mask()
{
return as_uint4(SPIRV_BUILTIN_NO_OP(BuiltInSubgroupGtMask, , )());
}
INLINE uint4 OVERLOADABLE get_sub_group_le_mask()
{
return as_uint4(SPIRV_BUILTIN_NO_OP(BuiltInSubgroupLeMask, , )());
}
INLINE uint4 OVERLOADABLE get_sub_group_lt_mask()
{
return as_uint4(SPIRV_BUILTIN_NO_OP(BuiltInSubgroupLtMask, , )());
}
#endif // defined(cl_khr_subgroup_ballot)
/////////////////////////////////////////////////////////////////////////////////////
// Media block read/write extension
#define DEFN_MEDIA_BLOCK_READ_RO(TYPE, TYPE_POSTFIX, TYPE_ABBR, LEN) \
OVERLOADABLE INLINE \
TYPE##LEN intel_sub_group_media_block_read_##TYPE_POSTFIX##LEN( \
int2 src_offset, \
int width, \
int height, \
read_only image2d_t image) \
{ \
int id = (int)__builtin_astype(image, global void*); \
return __builtin_IB_media_block_read_##TYPE##LEN( \
id, src_offset, width, height); \
}
#if SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_MEDIA_BLOCK_READ_RW(TYPE, TYPE_POSTFIX, TYPE_ABBR, LEN) \
OVERLOADABLE INLINE \
TYPE##LEN intel_sub_group_media_block_read_##TYPE_POSTFIX##LEN( \
int2 src_offset, \
int width, \
int height, \
read_write image2d_t image) \
{ \
int id = (int)__builtin_astype(image, global void*); \
return __builtin_IB_media_block_read_##TYPE##LEN( \
id, src_offset, width, height); \
}
#endif
#define DEFN_MEDIA_BLOCK_WRITE_WO(TYPE, TYPE_POSTFIX, TYPE_ABBR, LEN) \
OVERLOADABLE INLINE \
void intel_sub_group_media_block_write_##TYPE_POSTFIX##LEN( \
int2 src_offset, \
int width, \
int height, \
TYPE##LEN pixels, \
write_only image2d_t image) \
{ \
int id = (int)__builtin_astype(image, global void*); \
__builtin_IB_media_block_write_##TYPE##LEN( \
id, src_offset, width, height, pixels); \
}
#if SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_MEDIA_BLOCK_WRITE_RW(TYPE, TYPE_POSTFIX, TYPE_ABBR, LEN) \
OVERLOADABLE INLINE \
void intel_sub_group_media_block_write_##TYPE_POSTFIX##LEN( \
int2 src_offset, \
int width, \
int height, \
TYPE##LEN pixels, \
read_write image2d_t image) \
{ \
int id = (int)__builtin_astype(image, global void*); \
__builtin_IB_media_block_write_##TYPE##LEN( \
id, src_offset, width, height, pixels); \
}
#endif
#define DEFN_MEDIA_BLOCK(OPERATION, TYPE, TYPE_POSTFIX, TYPE_ABBR) \
DEFN_MEDIA_BLOCK_##OPERATION(TYPE, TYPE_POSTFIX, TYPE_ABBR, ) \
DEFN_MEDIA_BLOCK_##OPERATION(TYPE, TYPE_POSTFIX, v2##TYPE_ABBR, 2) \
DEFN_MEDIA_BLOCK_##OPERATION(TYPE, TYPE_POSTFIX, v4##TYPE_ABBR, 4) \
DEFN_MEDIA_BLOCK_##OPERATION(TYPE, TYPE_POSTFIX, v8##TYPE_ABBR, 8) \
MEDIA_IO_HAS16( DEFN_MEDIA_BLOCK_##OPERATION(TYPE, TYPE_POSTFIX, v16##TYPE_ABBR, 16) )
#define MEDIA_IO_HAS16(x) x
DEFN_MEDIA_BLOCK(READ_RO, uchar, uc, i8)
DEFN_MEDIA_BLOCK(READ_RO, ushort, us, i16)
DEFN_MEDIA_BLOCK(WRITE_WO, uchar, uc, i8)
DEFN_MEDIA_BLOCK(WRITE_WO, ushort, us, i16)
#if SUPPORT_ACCESS_QUAL_OVERLOAD
DEFN_MEDIA_BLOCK(READ_RW, uchar, uc, i8)
DEFN_MEDIA_BLOCK(READ_RW, ushort, us, i16)
DEFN_MEDIA_BLOCK(WRITE_RW, uchar, uc, i8)
DEFN_MEDIA_BLOCK(WRITE_RW, ushort, us, i16)
#endif
// Integer block read/writes don't have 16 element version.
#undef MEDIA_IO_HAS16
#define MEDIA_IO_HAS16(x)
DEFN_MEDIA_BLOCK(READ_RO, uint, ui, i32)
DEFN_MEDIA_BLOCK(WRITE_WO, uint, ui, i32)
#if SUPPORT_ACCESS_QUAL_OVERLOAD
DEFN_MEDIA_BLOCK(READ_RW, uint, ui, i32)
DEFN_MEDIA_BLOCK(WRITE_RW, uint, ui, i32)
#endif
#undef MEDIA_IO_HAS16
/////////////////////////////////////////////////////////////////////////////////////
// Block Read/Write Functions
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE(FUNC_NAME, TYPE, INTERNAL_FUNC) \
INLINE TYPE OVERLOADABLE FUNC_NAME( image2d_t image, int2 coord ) \
{ \
int id = (int)__builtin_astype(image, __global void*); \
return INTERNAL_FUNC(id, coord); \
}
#if SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(FUNC_NAME, TYPE, INTERNAL_FUNC) \
INLINE TYPE OVERLOADABLE FUNC_NAME( read_write image2d_t image, int2 coord ) \
{ \
int id = (int)__builtin_astype(image, __global void*); \
return INTERNAL_FUNC(id, coord); \
}
#else
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(FUNC_NAME, TYPE, INTERNAL_FUNC)
#endif // SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(FUNC_NAME, TYPE, ELEM_TYPE, TYPE_ABBR, INTERNAL_FUNC) \
INLINE TYPE OVERLOADABLE FUNC_NAME( const __global ELEM_TYPE* p ) \
{ \
return INTERNAL_FUNC((__global void *)p); \
} \
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(FUNC_NAME, TYPE, ELEM_TYPE, TYPE_ABBR, INTERNAL_FUNC) \
INLINE TYPE OVERLOADABLE FUNC_NAME( const __local ELEM_TYPE* p ) \
{ \
return INTERNAL_FUNC(p); \
} \
#if SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(FUNC_NAME, TYPE, TYPE_ABBR, INTERNAL_FUNC) \
INLINE void OVERLOADABLE FUNC_NAME( write_only image2d_t image, int2 coord, TYPE data ) \
{ \
int id = (int)__builtin_astype(image, __global void*); \
INTERNAL_FUNC(id, coord, data); \
}
#else
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(FUNC_NAME, TYPE, TYPE_ABBR, INTERNAL_FUNC)
#endif // SUPPORT_ACCESS_QUAL_OVERLOAD
#if SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(FUNC_NAME, TYPE, INTERNAL_FUNC) \
INLINE void OVERLOADABLE FUNC_NAME( read_write image2d_t image, int2 coord, TYPE data )\
{ \
int id = (int)__builtin_astype(image, __global void*); \
INTERNAL_FUNC(id, coord, data); \
}
#else
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(FUNC_NAME, TYPE, INTERNAL_FUNC)
#endif // SUPPORT_ACCESS_QUAL_OVERLOAD
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(FUNC_NAME, TYPE, ELEM_TYPE, PTR_TYPE, TYPE_ABBR, INTERNAL_FUNC) \
INLINE void OVERLOADABLE FUNC_NAME( __global ELEM_TYPE* p, TYPE data ) \
{ \
INTERNAL_FUNC(p, data); \
} \
#define DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(FUNC_NAME, TYPE, ELEM_TYPE, PTR_TYPE, TYPE_ABBR, INTERNAL_FUNC) \
INLINE void OVERLOADABLE FUNC_NAME( __local ELEM_TYPE* p, TYPE data ) \
{ \
INTERNAL_FUNC(p, data); \
} \
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read, uint, __builtin_IB_simd_media_block_read_1)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read2, uint2, __builtin_IB_simd_media_block_read_2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read4, uint4, __builtin_IB_simd_media_block_read_4)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read8, uint8, __builtin_IB_simd_media_block_read_8)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read, uint, uint, i32, __builtin_IB_simd_block_read_1_global)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read2, uint2, uint, i32, __builtin_IB_simd_block_read_2_global)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read4, uint4, uint, i32, __builtin_IB_simd_block_read_4_global)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read8, uint8, uint, i32, __builtin_IB_simd_block_read_8_global)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write, uint, i32, __builtin_IB_simd_media_block_write_1)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write2, uint2, v2i32, __builtin_IB_simd_media_block_write_2)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write4, uint4, v4i32, __builtin_IB_simd_media_block_write_4)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write8, uint8, v8i32, __builtin_IB_simd_media_block_write_8)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write, uint, __builtin_IB_simd_media_block_write_1)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write2, uint2, __builtin_IB_simd_media_block_write_2)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write4, uint4, __builtin_IB_simd_media_block_write_4)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write8, uint8, __builtin_IB_simd_media_block_write_8)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write, uint, uint, i32, i32, __builtin_IB_simd_block_write_1_global)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write2, uint2, uint, i32, v2i32, __builtin_IB_simd_block_write_2_global)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write4, uint4, uint, i32, v4i32, __builtin_IB_simd_block_write_4_global)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write8, uint8, uint, i32, v8i32, __builtin_IB_simd_block_write_8_global)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_us, ushort, __builtin_IB_simd_media_block_read_1_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_us2, ushort2, __builtin_IB_simd_media_block_read_2_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_us4, ushort4, __builtin_IB_simd_media_block_read_4_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_us8, ushort8, __builtin_IB_simd_media_block_read_8_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_us, ushort, ushort, i16, __builtin_IB_simd_block_read_1_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_us2, ushort2, ushort, i16, __builtin_IB_simd_block_read_2_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_us4, ushort4, ushort, i16, __builtin_IB_simd_block_read_4_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_us8, ushort8, ushort, i16, __builtin_IB_simd_block_read_8_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_us, ushort, i16, __builtin_IB_simd_media_block_write_1_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_us2, ushort2, v2i16, __builtin_IB_simd_media_block_write_2_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_us4, ushort4, v4i16, __builtin_IB_simd_media_block_write_4_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_us8, ushort8, v8i16, __builtin_IB_simd_media_block_write_8_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_us, ushort, __builtin_IB_simd_media_block_write_1_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_us2, ushort2, __builtin_IB_simd_media_block_write_2_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_us4, ushort4, __builtin_IB_simd_media_block_write_4_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_us8, ushort8, __builtin_IB_simd_media_block_write_8_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_us, ushort, ushort, i16, i16, __builtin_IB_simd_block_write_1_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_us2, ushort2, ushort, i16, v2i16, __builtin_IB_simd_block_write_2_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_us4, ushort4, ushort, i16, v4i16, __builtin_IB_simd_block_write_4_global_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_us8, ushort8, ushort, i16, v8i16, __builtin_IB_simd_block_write_8_global_h)
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_uc, uchar, __builtin_IB_simd_media_block_read_1_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_uc2, uchar2, __builtin_IB_simd_media_block_read_2_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_uc4, uchar4, __builtin_IB_simd_media_block_read_4_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_uc8, uchar8, __builtin_IB_simd_media_block_read_8_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_uc16, uchar16, __builtin_IB_simd_media_block_read_16_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_uc, uchar, uchar, i8, __builtin_IB_simd_block_read_1_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_uc2, uchar2, uchar, i8, __builtin_IB_simd_block_read_2_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_uc4, uchar4, uchar, i8, __builtin_IB_simd_block_read_4_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_uc8, uchar8, uchar, i8, __builtin_IB_simd_block_read_8_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_uc16, uchar16, uchar, i8, __builtin_IB_simd_block_read_16_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_uc, uchar, i8, __builtin_IB_simd_media_block_write_1_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_uc2, uchar2, v2i8, __builtin_IB_simd_media_block_write_2_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_uc4, uchar4, v4i8, __builtin_IB_simd_media_block_write_4_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_uc8, uchar8, v8i8, __builtin_IB_simd_media_block_write_8_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_uc16, uchar16, v16i8, __builtin_IB_simd_media_block_write_16_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_uc, uchar, __builtin_IB_simd_media_block_write_1_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_uc2, uchar2, __builtin_IB_simd_media_block_write_2_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_uc4, uchar4, __builtin_IB_simd_media_block_write_4_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_uc8, uchar8, __builtin_IB_simd_media_block_write_8_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_uc16, uchar16, __builtin_IB_simd_media_block_write_16_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_uc, uchar, uchar, i8, i8, __builtin_IB_simd_block_write_1_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_uc2, uchar2, uchar, i8, v2i8, __builtin_IB_simd_block_write_2_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_uc4, uchar4, uchar, i8, v4i8, __builtin_IB_simd_block_write_4_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_uc8, uchar8, uchar, i8, v8i8, __builtin_IB_simd_block_write_8_global_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_uc16, uchar16, uchar, i8, v16i8, __builtin_IB_simd_block_write_16_global_b)
#endif // cl_intel_subgroups_char
#ifdef cl_intel_subgroups_long
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_ul, ulong, __builtin_IB_simd_media_block_read_1_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_ul2, ulong2, __builtin_IB_simd_media_block_read_2_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_ul4, ulong4, __builtin_IB_simd_media_block_read_4_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_IMAGE_RW(intel_sub_group_block_read_ul8, ulong8, __builtin_IB_simd_media_block_read_8_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_ul, ulong, ulong, i64, __builtin_IB_simd_block_read_1_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_ul2, ulong2, ulong, i64, __builtin_IB_simd_block_read_2_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_ul4, ulong4, ulong, i64, __builtin_IB_simd_block_read_4_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_GLOBAL(intel_sub_group_block_read_ul8, ulong8, ulong, i64, __builtin_IB_simd_block_read_8_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_ul, ulong, i64, __builtin_IB_simd_media_block_write_1_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_ul2, ulong2, v2i64, __builtin_IB_simd_media_block_write_2_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_ul4, ulong4, v4i64, __builtin_IB_simd_media_block_write_4_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_WO(intel_sub_group_block_write_ul8, ulong8, v8i64, __builtin_IB_simd_media_block_write_8_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_ul, ulong, __builtin_IB_simd_media_block_write_1_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_ul2, ulong2, __builtin_IB_simd_media_block_write_2_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_ul4, ulong4, __builtin_IB_simd_media_block_write_4_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_IMAGE_RW(intel_sub_group_block_write_ul8, ulong8, __builtin_IB_simd_media_block_write_8_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_ul, ulong, ulong, i64, i64, __builtin_IB_simd_block_write_1_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_ul2, ulong2, ulong, i64, v2i64, __builtin_IB_simd_block_write_2_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_ul4, ulong4, ulong, i64, v4i64, __builtin_IB_simd_block_write_4_global_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_GLOBAL(intel_sub_group_block_write_ul8, ulong8, ulong, i64, v8i64, __builtin_IB_simd_block_write_8_global_l)
#endif // cl_intel_subgroups_long
#ifdef cl_intel_subgroup_local_block_io
//
// SLM simd block read/write
//
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read, uint, uint, i32, __builtin_IB_simd_block_read_1_local)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read2, uint2, uint, i32, __builtin_IB_simd_block_read_2_local)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read4, uint4, uint, i32, __builtin_IB_simd_block_read_4_local)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read8, uint8, uint, i32, __builtin_IB_simd_block_read_8_local)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write, uint, uint, i32, i32, __builtin_IB_simd_block_write_1_local)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write2, uint2, uint, i32, v2i32, __builtin_IB_simd_block_write_2_local)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write4, uint4, uint, i32, v4i32, __builtin_IB_simd_block_write_4_local)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write8, uint8, uint, i32, v8i32, __builtin_IB_simd_block_write_8_local)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_us, ushort, ushort, i16, __builtin_IB_simd_block_read_1_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_us2, ushort2, ushort, i16, __builtin_IB_simd_block_read_2_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_us4, ushort4, ushort, i16, __builtin_IB_simd_block_read_4_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_us8, ushort8, ushort, i16, __builtin_IB_simd_block_read_8_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_us, ushort, ushort, i16, i16, __builtin_IB_simd_block_write_1_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_us2, ushort2, ushort, i16, v2i16, __builtin_IB_simd_block_write_2_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_us4, ushort4, ushort, i16, v4i16, __builtin_IB_simd_block_write_4_local_h)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_us8, ushort8, ushort, i16, v8i16, __builtin_IB_simd_block_write_8_local_h)
#ifdef cl_intel_subgroups_char
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_uc, uchar, uchar, i8, __builtin_IB_simd_block_read_1_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_uc2, uchar2, uchar, i8, __builtin_IB_simd_block_read_2_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_uc4, uchar4, uchar, i8, __builtin_IB_simd_block_read_4_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_uc8, uchar8, uchar, i8, __builtin_IB_simd_block_read_8_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_uc16, uchar16, uchar, i8, __builtin_IB_simd_block_read_16_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_uc, uchar, uchar, i8, i8, __builtin_IB_simd_block_write_1_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_uc2, uchar2, uchar, i8, v2i8, __builtin_IB_simd_block_write_2_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_uc4, uchar4, uchar, i8, v4i8, __builtin_IB_simd_block_write_4_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_uc8, uchar8, uchar, i8, v8i8, __builtin_IB_simd_block_write_8_local_b)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_uc16, uchar16, uchar, i8, v16i8, __builtin_IB_simd_block_write_16_local_b)
#endif // cl_intel_subgroups_char
#ifdef cl_intel_subgroups_long
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_ul, ulong, ulong, i64, __builtin_IB_simd_block_read_1_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_ul2, ulong2, ulong, i64, __builtin_IB_simd_block_read_2_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_ul4, ulong4, ulong, i64, __builtin_IB_simd_block_read_4_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LOCAL(intel_sub_group_block_read_ul8, ulong8, ulong, i64, __builtin_IB_simd_block_read_8_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_ul, ulong, ulong, i64, i64, __builtin_IB_simd_block_write_1_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_ul2, ulong2, ulong, i64, v2i64, __builtin_IB_simd_block_write_2_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_ul4, ulong4, ulong, i64, v4i64, __builtin_IB_simd_block_write_4_local_l)
DEFN_INTEL_SUB_GROUP_BLOCK_WRITE_LOCAL(intel_sub_group_block_write_ul8, ulong8, ulong, i64, v8i64, __builtin_IB_simd_block_write_8_local_l)
#endif // cl_intel_subgroups_long
#endif // cl_intel_subgroup_local_block_io
#if defined(cl_intel_subgroup_extended_block_read)
#define DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(FUNC_NAME, TYPE, INTERNAL_FUNC) \
INLINE TYPE FUNC_NAME( __global void* base_address, int width, int height, int pitch, int2 coord ) \
{ \
long baseoffset = as_long(base_address); \
int width_minus_one = width - 1; \
int height_minus_one = height - 1; \
int pitch_minus_one = pitch - 1; \
return INTERNAL_FUNC(baseoffset, width_minus_one, height_minus_one, pitch_minus_one, coord); \
}
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u8_m1k32v2, ushort2, __builtin_IB_subgroup_block_read_flat_u8_m1k32v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u8_m2k32v2, ushort4, __builtin_IB_subgroup_block_read_flat_u8_m2k32v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u8_m4k32v2, ushort8, __builtin_IB_subgroup_block_read_flat_u8_m4k32v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u8_m8k32v2, ushort16, __builtin_IB_subgroup_block_read_flat_u8_m8k32v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u16_m1k16v2, ushort2, __builtin_IB_subgroup_block_read_flat_u16_m1k16v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u16_m2k16v2, ushort4, __builtin_IB_subgroup_block_read_flat_u16_m2k16v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u16_m4k16v2, ushort8, __builtin_IB_subgroup_block_read_flat_u16_m4k16v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_u16_m8k16v2, ushort16, __builtin_IB_subgroup_block_read_flat_u16_m8k16v2)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_transform_u8_k32, uint8, __builtin_IB_subgroup_block_read_flat_transform_u8_k32)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_transform_u16_k16, uint8, __builtin_IB_subgroup_block_read_flat_transform_u16_k16)
DEFN_INTEL_SUB_GROUP_BLOCK_READ_LSC_FLAT(intel_subgroup_block_read_transpose_u32_k8, uint8, __builtin_IB_subgroup_block_read_flat_transpose_u32_k8)
#endif // defined(cl_intel_subgroup_extended_block_read)
#if defined(cl_khr_subgroup_shuffle)
#define DEFN_SUB_GROUP_SHUFFLE(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE sub_group_shuffle(TYPE value, uint index) \
{ \
return SPIRV_BUILTIN(GroupNonUniformShuffle, _i32_##TYPE_ABBR##_i32, )(Subgroup, as_##SPV_TYPE(value), index); \
} \
INLINE TYPE OVERLOADABLE sub_group_shuffle_xor(TYPE value, uint mask) \
{ \
return SPIRV_BUILTIN(GroupNonUniformShuffleXor, _i32_##TYPE_ABBR##_i32, )(Subgroup, as_##SPV_TYPE(value), mask); \
}
DEFN_SUB_GROUP_SHUFFLE(char, char, i8)
DEFN_SUB_GROUP_SHUFFLE(uchar, char, i8)
DEFN_SUB_GROUP_SHUFFLE(short, short, i16)
DEFN_SUB_GROUP_SHUFFLE(ushort, short, i16)
DEFN_SUB_GROUP_SHUFFLE(int, int, i32)
DEFN_SUB_GROUP_SHUFFLE(uint, int, i32)
DEFN_SUB_GROUP_SHUFFLE(long, long, i64)
DEFN_SUB_GROUP_SHUFFLE(ulong, long, i64)
DEFN_SUB_GROUP_SHUFFLE(float, float, f32)
#if defined(cl_khr_fp64)
DEFN_SUB_GROUP_SHUFFLE(double, double, f64)
#endif // defined(cl_khr_fp64)
#if defined(cl_khr_fp16)
DEFN_SUB_GROUP_SHUFFLE(half, half, f16)
#endif // defined(cl_khr_fp16)
#endif // defined(cl_khr_subgroup_shuffle)
#if defined(cl_khr_subgroup_shuffle_relative)
#define DEFN_SUB_GROUP_SHUFFLE_RELATIVE(TYPE, SPV_TYPE, TYPE_ABBR) \
INLINE TYPE OVERLOADABLE sub_group_shuffle_up(TYPE value, uint delta) \
{ \
return SPIRV_BUILTIN(GroupNonUniformShuffleUp, _i32_##TYPE_ABBR##_i32, )(Subgroup, as_##SPV_TYPE(value), delta); \
} \
INLINE TYPE OVERLOADABLE sub_group_shuffle_down(TYPE value, uint delta) \
{ \
return SPIRV_BUILTIN(GroupNonUniformShuffleDown, _i32_##TYPE_ABBR##_i32, )(Subgroup, as_##SPV_TYPE(value), delta); \
}
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(char, char, i8)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(uchar, char, i8)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(short, short, i16)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(ushort, short, i16)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(int, int, i32)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(uint, int, i32)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(long, long, i64)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(ulong, long, i64)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(float, float, f32)
#if defined(cl_khr_fp64)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(double, double, f64)
#endif // defined(cl_khr_fp64)
#if defined(cl_khr_fp16)
DEFN_SUB_GROUP_SHUFFLE_RELATIVE(half, half, f16)
#endif // defined(cl_khr_fp16)
#endif // defined(cl_khr_subgroup_shuffle_relative)
|
