File: llvm.amdgcn.mfma.gfx940.ll

package info (click to toggle)
llvm-toolchain-15 1%3A15.0.6-4
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,554,644 kB
  • sloc: cpp: 5,922,452; ansic: 1,012,136; asm: 674,362; python: 191,568; objc: 73,855; f90: 42,327; lisp: 31,913; pascal: 11,973; javascript: 10,144; sh: 9,421; perl: 7,447; ml: 5,527; awk: 3,523; makefile: 2,520; xml: 885; cs: 573; fortran: 567
file content (494 lines) | stat: -rw-r--r-- 33,944 bytes parent folder | download
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
; RUN: llc -march=amdgcn -mcpu=gfx940 -verify-machineinstrs < %s | FileCheck -enable-var-scope --check-prefixes=GCN,GFX940,VGPRCD %s
; RUN: llc -march=amdgcn -mcpu=gfx940 -global-isel -verify-machineinstrs < %s | FileCheck -enable-var-scope --check-prefixes=GCN,GISEL,VGPRCD %s
; RUN: llc -march=amdgcn -mcpu=gfx940 -stress-regalloc=10 -verify-machineinstrs < %s | FileCheck -enable-var-scope --check-prefixes=GCN,GFX940,AGPRCD %s
; RUN: llc -march=amdgcn -mcpu=gfx940 -stress-regalloc=10 -global-isel -verify-machineinstrs < %s | FileCheck -enable-var-scope --check-prefixes=GCN,GISEL,AGPRCD %s

declare <4 x i32> @llvm.amdgcn.mfma.i32.16x16x32.i8(i64, i64, <4 x i32>, i32, i32, i32)
declare <16 x i32> @llvm.amdgcn.mfma.i32.32x32x16.i8(i64, i64, <16 x i32>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.mfma.f32.16x16x8.xf32(<2 x float>, <2 x float>, <4 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.mfma.f32.32x32x4.xf32(<2 x float>, <2 x float>, <16 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.bf8.bf8(i64, i64, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.bf8.fp8(i64, i64, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.fp8.bf8(i64, i64, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.fp8.fp8(i64, i64, <4 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.bf8.bf8(i64, i64, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.bf8.fp8(i64, i64, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.fp8.bf8(i64, i64, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.fp8.fp8(i64, i64, <16 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x32.f16(<4 x half>, <8 x half>, <4 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x16.f16(<4 x half>, <8 x half>, <16 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x32.bf16(<4 x i16>, <8 x i16>, <4 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x16.bf16(<4 x i16>, <8 x i16>, <16 x float>, i32, i32, i32)
declare <4 x i32> @llvm.amdgcn.smfmac.i32.16x16x64.i8(<2 x i32>, <4 x i32>, <4 x i32>, i32, i32, i32)
declare <16 x i32> @llvm.amdgcn.smfmac.i32.32x32x32.i8(<2 x i32>, <4 x i32>, <16 x i32>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.bf8.bf8(<2 x i32>, <4 x i32>, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.bf8.fp8(<2 x i32>, <4 x i32>, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.fp8.bf8(<2 x i32>, <4 x i32>, <4 x float>, i32, i32, i32)
declare <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.fp8.fp8(<2 x i32>, <4 x i32>, <4 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.bf8.bf8(<2 x i32>, <4 x i32>, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.bf8.fp8(<2 x i32>, <4 x i32>, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.fp8.bf8(<2 x i32>, <4 x i32>, <16 x float>, i32, i32, i32)
declare <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.fp8.fp8(<2 x i32>, <4 x i32>, <16 x float>, i32, i32, i32)

; GCN-LABEL: {{^}}test_mfma_i32_16x16x32i8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_i32_16x16x32_i8 a[{{[0-9]+:[0-9]+}}], v[[[TWO]]:[[ONE]]], v[[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_i32_16x16x32_i8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_i32_16x16x32i8(<4 x i32> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x i32>, <4 x i32> addrspace(1)* %arg
  %mai.1 = tail call <4 x i32> @llvm.amdgcn.mfma.i32.16x16x32.i8(i64 4294967298, i64 12884901892, <4 x i32> %in.1, i32 1, i32 2, i32 3)
  store <4 x i32> %mai.1, <4 x i32> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_i32_32x32x16i8:
; GFX940-DAG:   v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:   v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:   v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:   v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-16: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:       v_mfma_i32_32x32x16_i8 a[{{[0-9]+:[0-9]+}}], v[[[TWO]]:[[ONE]]], v[[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:        v_mfma_i32_32x32x16_i8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:      v_accvgpr_read_b32
; GCN-COUNT-4:  global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_i32_32x32x16i8(<16 x i32> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x i32>, <16 x i32> addrspace(1)* %arg
  %mai.1 = tail call <16 x i32> @llvm.amdgcn.mfma.i32.32x32x16.i8(i64 4294967298, i64 12884901892, <16 x i32> %in.1, i32 1, i32 2, i32 3)
  store <16 x i32> %mai.1, <16 x i32> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_16x16x8xf32:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1.0
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2.0
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 0x40400000
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4.0
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_16x16x8_xf32 a[{{[0-9]+:[0-9]+}}], v[[[ONE]]:[[TWO]]], v[[[THREE]]:[[FOUR]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_16x16x8_xf32 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_16x16x8xf32(<4 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.16x16x8.xf32(<2 x float> <float 1.0, float 2.0>, <2 x float> <float 3.0, float 4.0>, <4 x float> %in.1, i32 1, i32 2, i32 3)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_32x32x4xf32:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1.0
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2.0
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 0x40400000
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4.0
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_32x32x4_xf32 a[{{[0-9]+:[0-9]+}}], v[[[ONE]]:[[TWO]]], v[[[THREE]]:[[FOUR]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_32x32x4_xf32 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_32x32x4xf32(<16 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.mfma.f32.32x32x4.xf32(<2 x float> <float 1.0, float 2.0>, <2 x float> <float 3.0, float 4.0>, <16 x float> %in.1, i32 1, i32 2, i32 3)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_16x16x32_bf8_bf8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_16x16x32_bf8_bf8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_16x16x32_bf8_bf8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_16x16x32_bf8_bf8(<4 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.bf8.bf8(i64 4294967298, i64 12884901892, <4 x float> %in.1, i32 1, i32 2, i32 3)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_16x16x32_bf8_fp8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_16x16x32_bf8_fp8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_16x16x32_bf8_fp8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_16x16x32_bf8_fp8(<4 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.bf8.fp8(i64 4294967298, i64 12884901892, <4 x float> %in.1, i32 1, i32 2, i32 3)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_16x16x32_fp8_bf8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_16x16x32_fp8_bf8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_16x16x32_fp8_bf8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_16x16x32_fp8_bf8(<4 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.fp8.bf8(i64 4294967298, i64 12884901892, <4 x float> %in.1, i32 1, i32 2, i32 3)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_16x16x32_fp8_fp8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_16x16x32_fp8_fp8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_16x16x32_fp8_fp8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_16x16x32_fp8_fp8(<4 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.mfma.f32.16x16x32.fp8.fp8(i64 4294967298, i64 12884901892, <4 x float> %in.1, i32 1, i32 2, i32 3)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_32x32x16_bf8_bf8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_32x32x16_bf8_bf8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_32x32x16_bf8_bf8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_32x32x16_bf8_bf8(<16 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.bf8.bf8(i64 4294967298, i64 12884901892, <16 x float> %in.1, i32 1, i32 2, i32 3)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_32x32x16_bf8_fp8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_32x32x16_bf8_fp8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_32x32x16_bf8_fp8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_32x32x16_bf8_fp8(<16 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.bf8.fp8(i64 4294967298, i64 12884901892, <16 x float> %in.1, i32 1, i32 2, i32 3)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_32x32x16_fp8_bf8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_32x32x16_fp8_bf8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_32x32x16_fp8_bf8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_32x32x16_fp8_bf8(<16 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.fp8.bf8(i64 4294967298, i64 12884901892, <16 x float> %in.1, i32 1, i32 2, i32 3)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_mfma_f32_32x32x16_fp8_fp8:
; GFX940-DAG:  v_mov_b32_e32 v[[ONE:[0-9]+]], 1
; GFX940-DAG:  v_mov_b32_e32 v[[TWO:[0-9]+]], 2
; GFX940-DAG:  v_mov_b32_e32 v[[THREE:[0-9]+]], 3
; GFX940-DAG:  v_mov_b32_e32 v[[FOUR:[0-9]+]], 4
; GCN-COUNT-4: v_accvgpr_write_b32 a{{[0-9]+}}, s{{[0-9]+}}
; GFX940:      v_mfma_f32_32x32x16_fp8_fp8 a[{{[0-9]+:[0-9]+}}], v{{\[}}[[TWO]]:[[ONE]]], v{{\[}}[[FOUR]]:[[THREE]]], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GISEL:       v_mfma_f32_32x32x16_fp8_fp8 a[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], v[{{[0-9]+:[0-9]+}}], a[{{[0-9]+:[0-9]+}}] cbsz:1 abid:2 blgp:3
; GCN-NOT:     v_accvgpr_read_b32
; GCN:         global_store_dwordx4 v{{[0-9]+}}, a[{{[0-9:]+}}]
define amdgpu_kernel void @test_mfma_f32_32x32x16_fp8_fp8(<16 x float> addrspace(1)* %arg) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.mfma.f32.32x32x16.fp8.fp8(i64 4294967298, i64 12884901892, <16 x float> %in.1, i32 1, i32 2, i32 3)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_f32_16x16x32_f16:
; GCN:        s_load_dwordx4 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x32_f16 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_f32_16x16x32_f16(<4 x float> addrspace(1)* %arg, <4 x half> %a, <8 x half> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x32.f16(<4 x half> %a, <8 x half> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_f32_32x32x16_f16:
; GCN:        s_load_dwordx16 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x16_f16 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_f32_32x32x16_f16(<16 x float> addrspace(1)* %arg, <4 x half> %a, <8 x half> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x16.f16(<4 x half> %a, <8 x half> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_f32_16x16x32_bf16:
; GCN:        s_load_dwordx4 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x32_bf16 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_f32_16x16x32_bf16(<4 x float> addrspace(1)* %arg, <4 x i16> %a, <8 x i16> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x32.bf16(<4 x i16> %a, <8 x i16> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_f32_32x32x16_bf16:
; GCN:        s_load_dwordx16 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x16_bf16 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_f32_32x32x16_bf16(<16 x float> addrspace(1)* %arg, <4 x i16> %a, <8 x i16> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x16.bf16(<4 x i16> %a, <8 x i16> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_16x16x64_i8:
; GCN:        s_load_dwordx4 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_i32_16x16x64_i8 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_i32_16x16x64_i8(<4 x i32> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x i32>, <4 x i32> addrspace(1)* %arg
  %mai.1 = tail call <4 x i32> @llvm.amdgcn.smfmac.i32.16x16x64.i8(<2 x i32> %a, <4 x i32> %b, <4 x i32> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x i32> %mai.1, <4 x i32> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_32x32x32_i8:
; GCN:        s_load_dwordx16 s[[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]][[[RLO:[0-9]+]]:{{[0-9]+}}], s[[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_i32_32x32x32_i8 [[CD]][[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_i32_32x32x32_i8(<16 x i32> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x i32>, <16 x i32> addrspace(1)* %arg
  %mai.1 = tail call <16 x i32> @llvm.amdgcn.smfmac.i32.32x32x32.i8(<2 x i32> %a, <4 x i32> %b, <16 x i32> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x i32> %mai.1, <16 x i32> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_16x16x64_bf8_bf8:
; GCN:        s_load_dwordx4 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x64_bf8_bf8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_i32_16x16x64_bf8_bf8(<4 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.bf8.bf8(<2 x i32> %a, <4 x i32> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_16x16x64_bf8_fp8:
; GCN:        s_load_dwordx4 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x64_bf8_fp8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_i32_16x16x64_bf8_fp8(<4 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.bf8.fp8(<2 x i32> %a, <4 x i32> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_16x16x64_fp8_bf8:
; GCN:        s_load_dwordx4 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x64_fp8_bf8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_i32_16x16x64_fp8_bf8(<4 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.fp8.bf8(<2 x i32> %a, <4 x i32> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_16x16x64_fp8_fp8:
; GCN:        s_load_dwordx4 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_16x16x64_fp8_fp8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN:        global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:[[RHI]]]
define amdgpu_kernel void @test_smfmac_i32_16x16x64_fp8_fp8(<4 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <4 x float>, <4 x float> addrspace(1)* %arg
  %mai.1 = tail call <4 x float> @llvm.amdgcn.smfmac.f32.16x16x64.fp8.fp8(<2 x i32> %a, <4 x i32> %b, <4 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <4 x float> %mai.1, <4 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_32x32x32_bf8_bf8:
; GCN:        s_load_dwordx16 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x32_bf8_bf8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_i32_32x32x32_bf8_bf8(<16 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.bf8.bf8(<2 x i32> %a, <4 x i32> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_32x32x32_bf8_fp8:
; GCN:        s_load_dwordx16 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x32_bf8_fp8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_i32_32x32x32_bf8_fp8(<16 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.bf8.fp8(<2 x i32> %a, <4 x i32> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_32x32x32_fp8_bf8:
; GCN:        s_load_dwordx16 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x32_fp8_bf8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_i32_32x32x32_fp8_bf8(<16 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.fp8.bf8(<2 x i32> %a, <4 x i32> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

; GCN-LABEL: {{^}}test_smfmac_i32_32x32x32_fp8_fp8:
; GCN:        s_load_dwordx16 s{{\[}}[[SLO:[0-9]+]]:[[SHI:[0-9]+]]], s[{{[0-9:]+}}], 0x0{{$}}
; VGPRCD-DAG: v_mov_b64_e32 [[CD:v]]{{\[}}[[RLO:[0-9]+]]:{{[0-9]+}}], s{{\[}}[[SLO]]:{{[0-9]+}}]{{$}}
; VGPRCD-DAG: v_mov_b64_e32 v[{{[0-9]+}}:[[RHI:[0-9]+]]], s[{{[0-9]+}}:[[SHI]]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 [[CD:a]][[RLO:[0-9]+]], s[[SLO]]{{$}}
; AGPRCD-DAG: v_accvgpr_write_b32 a[[RHI:[0-9]+]], s[[SHI]]{{$}}
; GCN:        v_smfmac_f32_32x32x32_fp8_fp8 [[CD]]{{\[}}[[RLO]]:[[RHI]]], {{[av]}}[{{[0-9:]+}}], {{[av]}}[{{[0-9:]+}}], v{{[0-9]+}} cbsz:1 abid:2
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]]{{\[}}[[RLO]]:{{[0-9]+}}], s[{{[0-9:]+}}]{{$}}
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:16
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9:]+}}], s[{{[0-9:]+}}] offset:32
; GCN-DAG:    global_store_dwordx4 v{{[0-9]+}}, [[CD]][{{[0-9]+}}:[[RHI]]], s[{{[0-9:]+}}] offset:48
define amdgpu_kernel void @test_smfmac_i32_32x32x32_fp8_fp8(<16 x float> addrspace(1)* %arg, <2 x i32> %a, <4 x i32> %b, i32 %idx) #0 {
bb:
  %in.1 = load <16 x float>, <16 x float> addrspace(1)* %arg
  %mai.1 = tail call <16 x float> @llvm.amdgcn.smfmac.f32.32x32x32.fp8.fp8(<2 x i32> %a, <4 x i32> %b, <16 x float> %in.1, i32 %idx, i32 1, i32 2)
  store <16 x float> %mai.1, <16 x float> addrspace(1)* %arg
  ret void
}

attributes #0 = { "amdgpu-flat-work-group-size"="1,256" }