llvm.org GIT mirror llvm / 69f971e test / CodeGen / AMDGPU / smrd.ll
69f971e

Tree @69f971e (Download .tar.gz)

smrd.ll @69f971eraw · history · blame

  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
; RUN: llc -march=amdgcn -mcpu=tahiti  -verify-machineinstrs -show-mc-encoding < %s | FileCheck -check-prefix=SI   -check-prefix=GCN -check-prefix=SICIVI -check-prefix=SICI -check-prefix=SIVIGFX9 %s
; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs -show-mc-encoding < %s | FileCheck -check-prefix=CI   -check-prefix=GCN -check-prefix=SICIVI -check-prefix=SICI %s
; RUN: llc -march=amdgcn -mcpu=tonga   -verify-machineinstrs -show-mc-encoding < %s | FileCheck -check-prefix=VI   -check-prefix=GCN -check-prefix=SICIVI -check-prefix=VIGFX9 -check-prefix=SIVIGFX9 %s
; RUN: llc -march=amdgcn -mcpu=gfx900  -verify-machineinstrs -show-mc-encoding < %s | FileCheck -check-prefix=GFX9 -check-prefix=GCN -check-prefix=VIGFX9 -check-prefix=SIVIGFX9  %s

; SMRD load with an immediate offset.
; GCN-LABEL: {{^}}smrd0:
; SICI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x1 ; encoding: [0x01
; VIGFX9: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x4
define amdgpu_kernel void @smrd0(i32 addrspace(1)* %out, i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 1
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; SMRD load with the largest possible immediate offset.
; GCN-LABEL: {{^}}smrd1:
; SICI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xff ; encoding: [0xff,0x{{[0-9]+[137]}}
; VIGFX9: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3fc
define amdgpu_kernel void @smrd1(i32 addrspace(1)* %out, i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 255
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; SMRD load with an offset greater than the largest possible immediate.
; GCN-LABEL: {{^}}smrd2:
; SI: s_movk_i32 s[[OFFSET:[0-9]]], 0x400
; SI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], s[[OFFSET]] ; encoding: [0x0[[OFFSET]]
; CI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x100
; VIGFX9: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x400
; GCN: s_endpgm
define amdgpu_kernel void @smrd2(i32 addrspace(1)* %out, i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 256
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; SMRD load with a 64-bit offset
; GCN-LABEL: {{^}}smrd3:
; FIXME: There are too many copies here because we don't fold immediates
;        through REG_SEQUENCE
; SI: s_load_dwordx2 s[{{[0-9]:[0-9]}}], s[{{[0-9]:[0-9]}}], 0x13 ; encoding: [0x13
; TODO: Add VI checks
; GCN: s_endpgm
define amdgpu_kernel void @smrd3(i32 addrspace(1)* %out, [8 x i32], i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 4294967296
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; SMRD load with the largest possible immediate offset on VI
; GCN-LABEL: {{^}}smrd4:
; SI: s_mov_b32 [[OFFSET:s[0-9]+]], 0xffffc
; SI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; CI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3ffff
; VIGFX9: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xffffc
define amdgpu_kernel void @smrd4(i32 addrspace(1)* %out, i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 262143
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; SMRD load with an offset greater than the largest possible immediate on VI
; GCN-LABEL: {{^}}smrd5:
; SIVIGFX9: s_mov_b32 [[OFFSET:s[0-9]+]], 0x100000
; SIVIGFX9: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; CI: s_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x40000
; GCN: s_endpgm
define amdgpu_kernel void @smrd5(i32 addrspace(1)* %out, i32 addrspace(4)* %ptr) #0 {
entry:
  %tmp = getelementptr i32, i32 addrspace(4)* %ptr, i64 262144
  %tmp1 = load i32, i32 addrspace(4)* %tmp
  store i32 %tmp1, i32 addrspace(1)* %out
  ret void
}

; GCN-LABEL: {{^}}smrd_hazard:
; GCN-DAG: s_mov_b32 s3, 3
; GCN-DAG: s_mov_b32 s2, 2
; GCN-DAG: s_mov_b32 s1, 1
; GCN-DAG: s_mov_b32 s0, 0
; SI-NEXT: nop 3
; GCN-NEXT: s_buffer_load_dword s0, s[0:3], 0x0
define amdgpu_ps float @smrd_hazard(<4 x i32> inreg %desc) #0 {
main_body:
  %d0 = insertelement <4 x i32> undef, i32 0, i32 0
  %d1 = insertelement <4 x i32> %d0, i32 1, i32 1
  %d2 = insertelement <4 x i32> %d1, i32 2, i32 2
  %d3 = insertelement <4 x i32> %d2, i32 3, i32 3
  %r = call float @llvm.SI.load.const.v4i32(<4 x i32> %d3, i32 0)
  ret float %r
}

; SMRD load using the load.const.v4i32 intrinsic with an immediate offset
; GCN-LABEL: {{^}}smrd_load_const0:
; SICI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x4 ; encoding: [0x04
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x10
define amdgpu_ps void @smrd_load_const0(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 16)
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %tmp21, i1 true, i1 true) #0
  ret void
}

; SMRD load using the load.const.v4i32 intrinsic with the largest possible immediate
; offset.
; GCN-LABEL: {{^}}smrd_load_const1:
; SICI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xff ; encoding: [0xff
; SICI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xff glc ; encoding: [0xff
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3fc
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3fc glc
define amdgpu_ps void @smrd_load_const1(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 1020)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 1020, i32 1)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load using the load.const.v4i32 intrinsic with an offset greater than the
; largets possible immediate.
; immediate offset.
; GCN-LABEL: {{^}}smrd_load_const2:
; SI: s_movk_i32 s[[OFFSET:[0-9]]], 0x400
; SI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], s[[OFFSET]] ; encoding: [0x0[[OFFSET]]
; SI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], s[[OFFSET]] ; encoding: [0x0[[OFFSET]]
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x100
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x100
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x400
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x400
define amdgpu_ps void @smrd_load_const2(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 1024)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 1024, i32 0)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load with the largest possible immediate offset on VI
; GCN-LABEL: {{^}}smrd_load_const3:
; SI: s_mov_b32 [[OFFSET:s[0-9]+]], 0xffffc
; SI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; SI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3ffff
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x3ffff
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xffffc
; VIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0xffffc
define amdgpu_ps void @smrd_load_const3(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 1048572)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 1048572, i32 0)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load with an offset greater than the largest possible immediate on VI
; GCN-LABEL: {{^}}smrd_load_const4:
; SIVIGFX9: s_mov_b32 [[OFFSET:s[0-9]+]], 0x100000
; SIVIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; SIVIGFX9: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], [[OFFSET]]
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x40000
; CI: s_buffer_load_dword s{{[0-9]}}, s[{{[0-9]:[0-9]}}], 0x40000
; GCN: s_endpgm
define amdgpu_ps void @smrd_load_const4(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 1048576)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 1048576, i32 0)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; dwordx2 s.buffer.load
; GCN-LABEL: {{^}}s_buffer_load_dwordx2:
; VIGFX9: s_buffer_load_dwordx2 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x80
; SICI: s_buffer_load_dwordx2 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x20
define amdgpu_ps void @s_buffer_load_dwordx2(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <2 x i32> @llvm.amdgcn.s.buffer.load.v2i32(<4 x i32> %tmp22, i32 128, i32 0)
  %s.buffer.0 = extractelement <2 x i32> %s.buffer, i32 0
  %s.buffer.0.float = bitcast i32 %s.buffer.0 to float
  %s.buffer.1 = extractelement <2 x i32> %s.buffer, i32 1
  %s.buffer.1.float = bitcast i32 %s.buffer.1 to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %s.buffer.0.float, float %s.buffer.1.float, float %s.buffer.0.float, float %s.buffer.1.float, i1 true, i1 true) #0
  ret void
}

; dwordx4 s.buffer.load
; GCN-LABEL: {{^}}s_buffer_load_dwordx4:
; VIGFX9: s_buffer_load_dwordx4 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x80
; SICI: s_buffer_load_dwordx4 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x20
define amdgpu_ps void @s_buffer_load_dwordx4(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <4 x i32> @llvm.amdgcn.s.buffer.load.v4i32(<4 x i32> %tmp22, i32 128, i32 0)
  %s.buffer.0 = extractelement <4 x i32> %s.buffer, i32 0
  %s.buffer.0.float = bitcast i32 %s.buffer.0 to float
  %s.buffer.1 = extractelement <4 x i32> %s.buffer, i32 1
  %s.buffer.1.float = bitcast i32 %s.buffer.1 to float
  %s.buffer.2 = extractelement <4 x i32> %s.buffer, i32 2
  %s.buffer.2.float = bitcast i32 %s.buffer.2 to float
  %s.buffer.3 = extractelement <4 x i32> %s.buffer, i32 3
  %s.buffer.3.float = bitcast i32 %s.buffer.3 to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %s.buffer.0.float, float %s.buffer.1.float, float %s.buffer.2.float, float %s.buffer.3.float, i1 true, i1 true) #0
  ret void
}

; dwordx8 s.buffer.load
; GCN-LABEL: {{^}}s_buffer_load_dwordx8:
; VIGFX9: s_buffer_load_dwordx8 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x80
; SICI: s_buffer_load_dwordx8 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x20
define amdgpu_ps void @s_buffer_load_dwordx8(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <8 x i32> @llvm.amdgcn.s.buffer.load.v8i32(<4 x i32> %tmp22, i32 128, i32 0)
  %s.buffer.0 = extractelement <8 x i32> %s.buffer, i32 0
  %s.buffer.0.float = bitcast i32 %s.buffer.0 to float
  %s.buffer.1 = extractelement <8 x i32> %s.buffer, i32 2
  %s.buffer.1.float = bitcast i32 %s.buffer.1 to float
  %s.buffer.2 = extractelement <8 x i32> %s.buffer, i32 5
  %s.buffer.2.float = bitcast i32 %s.buffer.2 to float
  %s.buffer.3 = extractelement <8 x i32> %s.buffer, i32 7
  %s.buffer.3.float = bitcast i32 %s.buffer.3 to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %s.buffer.0.float, float %s.buffer.1.float, float %s.buffer.2.float, float %s.buffer.3.float, i1 true, i1 true) #0
  ret void
}

; dwordx16 s.buffer.load
; GCN-LABEL: {{^}}s_buffer_load_dwordx16:
; VIGFX9: s_buffer_load_dwordx16 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x80
; SICI: s_buffer_load_dwordx16 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]:[0-9]}}], 0x20
define amdgpu_ps void @s_buffer_load_dwordx16(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in) #0 {
main_body:
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <16 x i32> @llvm.amdgcn.s.buffer.load.v16i32(<4 x i32> %tmp22, i32 128, i32 0)
  %s.buffer.0 = extractelement <16 x i32> %s.buffer, i32 0
  %s.buffer.0.float = bitcast i32 %s.buffer.0 to float
  %s.buffer.1 = extractelement <16 x i32> %s.buffer, i32 3
  %s.buffer.1.float = bitcast i32 %s.buffer.1 to float
  %s.buffer.2 = extractelement <16 x i32> %s.buffer, i32 12
  %s.buffer.2.float = bitcast i32 %s.buffer.2 to float
  %s.buffer.3 = extractelement <16 x i32> %s.buffer, i32 15
  %s.buffer.3.float = bitcast i32 %s.buffer.3 to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %s.buffer.0.float, float %s.buffer.1.float, float %s.buffer.2.float, float %s.buffer.3.float, i1 true, i1 true) #0
  ret void
}

; GCN-LABEL: {{^}}smrd_sgpr_offset:
; GCN: s_buffer_load_dword s{{[0-9]}}, s[0:3], s4
define amdgpu_ps float @smrd_sgpr_offset(<4 x i32> inreg %desc, i32 inreg %offset) #0 {
main_body:
  %r = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %offset)
  ret float %r
}

; GCN-LABEL: {{^}}smrd_vgpr_offset:
; GCN: buffer_load_dword v{{[0-9]}}, v0, s[0:3], 0 offen ;
define amdgpu_ps float @smrd_vgpr_offset(<4 x i32> inreg %desc, i32 %offset) #0 {
main_body:
  %r = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %offset)
  ret float %r
}

; GCN-LABEL: {{^}}smrd_vgpr_offset_imm:
; GCN-NEXT: %bb.
; GCN-NEXT: buffer_load_dword v{{[0-9]}}, v0, s[0:3], 0 offen offset:4095 ;
define amdgpu_ps float @smrd_vgpr_offset_imm(<4 x i32> inreg %desc, i32 %offset) #0 {
main_body:
  %off = add i32 %offset, 4095
  %r = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %off)
  ret float %r
}

; GCN-LABEL: {{^}}smrd_vgpr_offset_imm_too_large:
; GCN-NEXT: %bb.
; GCN-NEXT: v_add_{{i|u}}32_e32 v0, {{(vcc, )?}}0x1000, v0
; GCN-NEXT: buffer_load_dword v{{[0-9]}}, v0, s[0:3], 0 offen ;
define amdgpu_ps float @smrd_vgpr_offset_imm_too_large(<4 x i32> inreg %desc, i32 %offset) #0 {
main_body:
  %off = add i32 %offset, 4096
  %r = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %off)
  ret float %r
}

; GCN-LABEL: {{^}}smrd_imm_merged:
; GCN-NEXT: %bb.
; SICI-NEXT: s_buffer_load_dwordx4 s[{{[0-9]}}:{{[0-9]}}], s[0:3], 0x1
; SICI-NEXT: s_buffer_load_dwordx2 s[{{[0-9]}}:{{[0-9]}}], s[0:3], 0x7
; VIGFX9-NEXT: s_buffer_load_dwordx4 s[{{[0-9]}}:{{[0-9]}}], s[0:3], 0x4
; VIGFX9-NEXT: s_buffer_load_dwordx2 s[{{[0-9]}}:{{[0-9]}}], s[0:3], 0x1c
define amdgpu_ps void @smrd_imm_merged(<4 x i32> inreg %desc) #0 {
main_body:
  %r1 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 4)
  %r2 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 8)
  %r3 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 12)
  %r4 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 16)
  %r5 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 28)
  %r6 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 32)
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %r1, float %r2, float %r3, float %r4, i1 true, i1 true) #0
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %r5, float %r6, float undef, float undef, i1 true, i1 true) #0
  ret void
}

; GCN-LABEL: {{^}}smrd_imm_merge_m0:
;
; SICIVI: s_buffer_load_dwordx2
; SICIVI: s_mov_b32 m0
; SICIVI_DAG: v_interp_p1_f32
; SICIVI_DAG: v_interp_p1_f32
; SICIVI_DAG: v_interp_p1_f32
; SICIVI_DAG: v_interp_p2_f32
; SICIVI_DAG: v_interp_p2_f32
; SICIVI_DAG: v_interp_p2_f32
; SICIVI: s_mov_b32 m0
; SICIVI: v_movrels_b32_e32
;
; Merging is still thwarted on GFX9 due to s_set_gpr_idx
;
; GFX9: s_buffer_load_dword
; GFX9: s_buffer_load_dword
define amdgpu_ps float @smrd_imm_merge_m0(<4 x i32> inreg %desc, i32 inreg %prim, float %u, float %v) #0 {
main_body:
  %idx1.f = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 0)
  %idx1 = bitcast float %idx1.f to i32

  %v0.x1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 0, i32 0, i32 %prim)
  %v0.x = call nsz float @llvm.amdgcn.interp.p2(float %v0.x1, float %v, i32 0, i32 0, i32 %prim)
  %v0.y1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 0, i32 1, i32 %prim)
  %v0.y = call nsz float @llvm.amdgcn.interp.p2(float %v0.y1, float %v, i32 0, i32 1, i32 %prim)
  %v0.z1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 0, i32 2, i32 %prim)
  %v0.z = call nsz float @llvm.amdgcn.interp.p2(float %v0.z1, float %v, i32 0, i32 2, i32 %prim)
  %v0.tmp0 = insertelement <3 x float> undef, float %v0.x, i32 0
  %v0.tmp1 = insertelement <3 x float> %v0.tmp0, float %v0.y, i32 1
  %v0 = insertelement <3 x float> %v0.tmp1, float %v0.z, i32 2
  %a = extractelement <3 x float> %v0, i32 %idx1

  %v1.x1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 1, i32 0, i32 %prim)
  %v1.x = call nsz float @llvm.amdgcn.interp.p2(float %v1.x1, float %v, i32 1, i32 0, i32 %prim)
  %v1.y1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 1, i32 1, i32 %prim)
  %v1.y = call nsz float @llvm.amdgcn.interp.p2(float %v1.y1, float %v, i32 1, i32 1, i32 %prim)
  %v1.z1 = call nsz float @llvm.amdgcn.interp.p1(float %u, i32 1, i32 2, i32 %prim)
  %v1.z = call nsz float @llvm.amdgcn.interp.p2(float %v1.z1, float %v, i32 1, i32 2, i32 %prim)
  %v1.tmp0 = insertelement <3 x float> undef, float %v0.x, i32 0
  %v1.tmp1 = insertelement <3 x float> %v0.tmp0, float %v0.y, i32 1
  %v1 = insertelement <3 x float> %v0.tmp1, float %v0.z, i32 2

  %b = extractelement <3 x float> %v1, i32 %idx1
  %c = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 4)

  %res.tmp = fadd float %a, %b
  %res = fadd float %res.tmp, %c
  ret float %res
}

; GCN-LABEL: {{^}}smrd_vgpr_merged:
; GCN-NEXT: %bb.
; GCN-NEXT: buffer_load_dwordx4 v[{{[0-9]}}:{{[0-9]}}], v0, s[0:3], 0 offen offset:4
; GCN-NEXT: buffer_load_dwordx2 v[{{[0-9]}}:{{[0-9]}}], v0, s[0:3], 0 offen offset:28
define amdgpu_ps void @smrd_vgpr_merged(<4 x i32> inreg %desc, i32 %a) #0 {
main_body:
  %a1 = add i32 %a, 4
  %a2 = add i32 %a, 8
  %a3 = add i32 %a, 12
  %a4 = add i32 %a, 16
  %a5 = add i32 %a, 28
  %a6 = add i32 %a, 32
  %r1 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a1)
  %r2 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a2)
  %r3 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a3)
  %r4 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a4)
  %r5 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a5)
  %r6 = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %a6)
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %r1, float %r2, float %r3, float %r4, i1 true, i1 true) #0
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %r5, float %r6, float undef, float undef, i1 true, i1 true) #0
  ret void
}

; GCN-LABEL: {{^}}smrd_sgpr_descriptor_promoted
; GCN: v_readfirstlane
define amdgpu_cs void @smrd_sgpr_descriptor_promoted([0 x i8] addrspace(4)* inreg noalias dereferenceable(18446744073709551615), i32) #0 {
main_body:
  %descptr = bitcast [0 x i8] addrspace(4)* %0 to <4 x i32> addrspace(4)*, !amdgpu.uniform !0
  br label %.outer_loop_header

ret_block:                                       ; preds = %.outer, %.label22, %main_body
  ret void

.outer_loop_header:
  br label %.inner_loop_header

.inner_loop_header:                                     ; preds = %.inner_loop_body, %.outer_loop_header
  %loopctr.1 = phi i32 [ 0, %.outer_loop_header ], [ %loopctr.2, %.inner_loop_body ]
  %loopctr.2 = add i32 %loopctr.1, 1
  %inner_br1 = icmp slt i32 %loopctr.2, 10
  br i1 %inner_br1, label %.inner_loop_body, label %ret_block

.inner_loop_body:
  %descriptor = load <4 x i32>, <4 x i32> addrspace(4)* %descptr, align 16, !invariant.load !0
  %load1result = call float @llvm.SI.load.const.v4i32(<4 x i32> %descriptor, i32 0)
  store float %load1result, float addrspace(1)* undef
  %inner_br2 = icmp uge i32 %1, 10
  br i1 %inner_br2, label %.inner_loop_header, label %.outer_loop_body

.outer_loop_body:
  %offset = shl i32 %loopctr.2, 6
  %load2result = call float @llvm.SI.load.const.v4i32(<4 x i32> %descriptor, i32 %offset)
  %outer_br = fcmp ueq float %load2result, 0x0
  br i1 %outer_br, label %.outer_loop_header, label %ret_block
}

; SMRD load with a non-const offset
; GCN-LABEL: {{^}}smrd_load_nonconst0:
; SIVIGFX9: s_buffer_load_dword s{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; SIVIGFX9: s_buffer_load_dword s{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; CI: s_buffer_load_dword s{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; CI: s_buffer_load_dword s{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; GCN: s_endpgm
define amdgpu_ps void @smrd_load_nonconst0(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in, i32 inreg %ncoff) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 %ncoff)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 %ncoff, i32 0)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load with a non-const non-uniform offset
; GCN-LABEL: {{^}}smrd_load_nonconst1:
; SIVIGFX9: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; SIVIGFX9: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; CI: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; CI: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; GCN: s_endpgm
define amdgpu_ps void @smrd_load_nonconst1(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in, i32 %ncoff) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 %ncoff)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32> %tmp22, i32 %ncoff, i32 0)
  %s.buffer.float = bitcast i32 %s.buffer to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load with a non-const non-uniform offset of > 4 dwords (requires splitting)
; GCN-LABEL: {{^}}smrd_load_nonconst2:
; SIVIGFX9: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; SIVIGFX9: buffer_load_dwordx4 v[{{[0-9]+:[0-9]+}}], v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; CI: buffer_load_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; CI: buffer_load_dwordx4 v[{{[0-9]+:[0-9]+}}], v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 offen
; GCN: s_endpgm
define amdgpu_ps void @smrd_load_nonconst2(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in, i32 %ncoff) #0 {
main_body:
  %tmp = getelementptr <4 x i32>, <4 x i32> addrspace(4)* %arg, i32 0
  %tmp20 = load <4 x i32>, <4 x i32> addrspace(4)* %tmp
  %tmp21 = call float @llvm.SI.load.const.v4i32(<4 x i32> %tmp20, i32 %ncoff)
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <8 x i32> @llvm.amdgcn.s.buffer.load.v8i32(<4 x i32> %tmp22, i32 %ncoff, i32 0)
  %s.buffer.elt = extractelement <8 x i32> %s.buffer, i32 1
  %s.buffer.float = bitcast i32 %s.buffer.elt to float
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %tmp21, float %tmp21, float %tmp21, float %s.buffer.float, i1 true, i1 true) #0
  ret void
}

; SMRD load with a non-const non-uniform offset of > 4 dwords (requires splitting)
; GCN-LABEL: {{^}}smrd_load_nonconst3:
; GCN-DAG: buffer_load_dwordx4 v[0:3], v{{[0-9]+}}, s[0:3], 0 offen ;
; GCN-DAG: buffer_load_dwordx4 v[4:7], v{{[0-9]+}}, s[0:3], 0 offen offset:16 ;
; GCN-DAG: buffer_load_dwordx4 v[8:11], v{{[0-9]+}}, s[0:3], 0 offen offset:32 ;
; GCN-DAG: buffer_load_dwordx4 v[12:15], v{{[0-9]+}}, s[0:3], 0 offen offset:48 ;
; GCN: ; return to shader part epilog
define amdgpu_ps <16 x float> @smrd_load_nonconst3(<4 x i32> inreg %rsrc, i32 %off) #0 {
main_body:
  %ld = call <16 x i32> @llvm.amdgcn.s.buffer.load.v16i32(<4 x i32> %rsrc, i32 %off, i32 0)
  %bc = bitcast <16 x i32> %ld to <16 x float>
  ret <16 x float> %bc
}

; GCN-LABEL: {{^}}smrd_load_nonconst4:
; SICIVI: v_add_{{i32|u32}}_e32 v{{[0-9]+}}, vcc, 0xff8, v0 ;
; GFX9: v_add_u32_e32 v{{[0-9]+}}, 0xff8, v0 ;
; GCN-DAG: buffer_load_dwordx4 v[0:3], v{{[0-9]+}}, s[0:3], 0 offen ;
; GCN-DAG: buffer_load_dwordx4 v[4:7], v{{[0-9]+}}, s[0:3], 0 offen offset:16 ;
; GCN-DAG: buffer_load_dwordx4 v[8:11], v{{[0-9]+}}, s[0:3], 0 offen offset:32 ;
; GCN-DAG: buffer_load_dwordx4 v[12:15], v{{[0-9]+}}, s[0:3], 0 offen offset:48 ;
; GCN: ; return to shader part epilog
define amdgpu_ps <16 x float> @smrd_load_nonconst4(<4 x i32> inreg %rsrc, i32 %off) #0 {
main_body:
  %off.2 = add i32 %off, 4088
  %ld = call <16 x i32> @llvm.amdgcn.s.buffer.load.v16i32(<4 x i32> %rsrc, i32 %off.2, i32 0)
  %bc = bitcast <16 x i32> %ld to <16 x float>
  ret <16 x float> %bc
}

; GCN-LABEL: {{^}}smrd_load_nonconst5:
; SICIVI: v_add_{{i32|u32}}_e32 v{{[0-9]+}}, vcc, 0x1004, v0
; GFX9: v_add_u32_e32 v{{[0-9]+}}, 0x1004, v0
; GCN-DAG: buffer_load_dwordx4 v[0:3], v{{[0-9]+}}, s[0:3], 0 offen ;
; GCN-DAG: buffer_load_dwordx4 v[4:7], v{{[0-9]+}}, s[0:3], 0 offen offset:16 ;
; GCN-DAG: buffer_load_dwordx4 v[8:11], v{{[0-9]+}}, s[0:3], 0 offen offset:32 ;
; GCN-DAG: buffer_load_dwordx4 v[12:15], v{{[0-9]+}}, s[0:3], 0 offen offset:48 ;
; GCN: ; return to shader part epilog
define amdgpu_ps <16 x float> @smrd_load_nonconst5(<4 x i32> inreg %rsrc, i32 %off) #0 {
main_body:
  %off.2 = add i32 %off, 4100
  %ld = call <16 x i32> @llvm.amdgcn.s.buffer.load.v16i32(<4 x i32> %rsrc, i32 %off.2, i32 0)
  %bc = bitcast <16 x i32> %ld to <16 x float>
  ret <16 x float> %bc
}

; SMRD load dwordx2
; GCN-LABEL: {{^}}smrd_load_dwordx2:
; SIVIGFX9: s_buffer_load_dwordx2 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; CI: s_buffer_load_dwordx2 s[{{[0-9]+:[0-9]+}}], s[{{[0-9]+:[0-9]+}}], s{{[0-9]+}}
; GCN: s_endpgm
define amdgpu_ps void @smrd_load_dwordx2(<4 x i32> addrspace(4)* inreg %arg, <4 x i32> addrspace(4)* inreg %arg1, <32 x i8> addrspace(4)* inreg %arg2, i32 inreg %arg3, <2 x i32> %arg4, <2 x i32> %arg5, <2 x i32> %arg6, <3 x i32> %arg7, <2 x i32> %arg8, <2 x i32> %arg9, <2 x i32> %arg10, float %arg11, float %arg12, float %arg13, float %arg14, float %arg15, float %arg16, float %arg17, float %arg18, float %arg19, <4 x i32> addrspace(4)* inreg %in, i32 inreg %ncoff) #0 {
main_body:
  %tmp22 = load <4 x i32>, <4 x i32> addrspace(4)* %in
  %s.buffer = call <2 x i32> @llvm.amdgcn.s.buffer.load.v2i32(<4 x i32> %tmp22, i32 %ncoff, i32 0)
  %s.buffer.float = bitcast <2 x i32> %s.buffer to <2 x float>
  %r.1 = extractelement <2 x float> %s.buffer.float, i32 0
  %r.2 = extractelement <2 x float> %s.buffer.float, i32 1
  call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %r.1, float %r.1, float %r.1, float %r.2, i1 true, i1 true) #0
  ret void
}

; GCN-LABEL: {{^}}smrd_uniform_loop:
;
; TODO: this should use an s_buffer_load
;
; GCN: buffer_load_dword
define amdgpu_ps float @smrd_uniform_loop(<4 x i32> inreg %desc, i32 %bound) #0 {
main_body:
  br label %loop

loop:
  %counter = phi i32 [ 0, %main_body ], [ %counter.next, %loop ]
  %sum = phi float [ 0.0, %main_body ], [ %sum.next, %loop ]
  %offset = shl i32 %counter, 2
  %v = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %offset)
  %sum.next = fadd float %sum, %v
  %counter.next = add i32 %counter, 1
  %cc = icmp uge i32 %counter.next, %bound
  br i1 %cc, label %exit, label %loop

exit:
  ret float %sum.next
}


; GCN-LABEL: {{^}}smrd_uniform_loop2:
; (this test differs from smrd_uniform_loop by the more complex structure of phis,
; which used to confuse the DivergenceAnalysis after structurization)
;
; TODO: we should keep the loop counter in an SGPR and use an S_BUFFER_LOAD
;
; GCN: buffer_load_dword
define amdgpu_ps float @smrd_uniform_loop2(<4 x i32> inreg %desc, i32 %bound, i32 %bound.a) #0 {
main_body:
  br label %loop

loop:
  %counter = phi i32 [ 0, %main_body ], [ %counter.next, %loop.a ], [ %counter.next, %loop.b ]
  %sum = phi float [ 0.0, %main_body ], [ %sum.next, %loop.a ], [ %sum.next.b, %loop.b ]
  %offset = shl i32 %counter, 2
  %v = call float @llvm.SI.load.const.v4i32(<4 x i32> %desc, i32 %offset)
  %sum.next = fadd float %sum, %v
  %counter.next = add i32 %counter, 1
  %cc = icmp uge i32 %counter.next, %bound
  br i1 %cc, label %exit, label %loop.a

loop.a:
  %cc.a = icmp uge i32 %counter.next, %bound.a
  br i1 %cc, label %loop, label %loop.b

loop.b:
  %sum.next.b = fadd float %sum.next, 1.0
  br label %loop

exit:
  ret float %sum.next
}


declare void @llvm.amdgcn.exp.f32(i32, i32, float, float, float, float, i1, i1) #0
declare float @llvm.SI.load.const.v4i32(<4 x i32>, i32) #1
declare float @llvm.amdgcn.interp.p1(float, i32, i32, i32) #2
declare float @llvm.amdgcn.interp.p2(float, float, i32, i32, i32) #2
declare i32 @llvm.amdgcn.s.buffer.load.i32(<4 x i32>, i32, i32)
declare <2 x i32> @llvm.amdgcn.s.buffer.load.v2i32(<4 x i32>, i32, i32)
declare <4 x i32> @llvm.amdgcn.s.buffer.load.v4i32(<4 x i32>, i32, i32)
declare <8 x i32> @llvm.amdgcn.s.buffer.load.v8i32(<4 x i32>, i32, i32)
declare <16 x i32> @llvm.amdgcn.s.buffer.load.v16i32(<4 x i32>, i32, i32)

attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }
attributes #2 = { nounwind readnone speculatable }

!0 = !{}