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
|
// RUN: mlir-opt %s -test-linalg-greedy-fusion -split-input-file | FileCheck %s
func.func @f1(%A: memref<?x?xf32, strided<[?, 1], offset: 0>>,
%B: memref<?x?xf32, strided<[?, 1], offset: 0>>,
%C: memref<?x?xf32, strided<[?, 1], offset: 0>>,
%D: memref<?x?xf32, strided<[?, 1], offset: 0>>,
%E: memref<?x?xf32, strided<[?, 1], offset: 0>>
) -> memref<?x?xf32, strided<[?, 1], offset: 0>> {
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%c1 = arith.constant 1 : index
%0 = memref.dim %A, %c0 : memref<?x?xf32, strided<[?, 1], offset: 0>>
%1 = memref.dim %A, %c1 : memref<?x?xf32, strided<[?, 1], offset: 0>>
%2 = memref.dim %B, %c1 : memref<?x?xf32, strided<[?, 1], offset: 0>>
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, 1], offset: 0>>,
memref<?x?xf32, strided<[?, 1], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, 1], offset: 0>>)
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%5 = memref.subview %A[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, 1], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %B[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, 1], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %C[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, 1], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8: memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, 1], offset: 0>>
}
// CHECK-LABEL: func @f1
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK: scf.for
// CHECK: scf.for
// CHECK: scf.for
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// -----
func.func @f2(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C: memref<?x?xf32, strided<[?, ?], offset: 0>>)
%0 = memref.dim %C, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %C, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%5 = memref.subview %C[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %D[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f2
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK-DAG: %[[C_0:.*]] = memref.dim %[[C]], %c0{{[_0-9]*}} : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[C_1:.*]] = memref.dim %[[C]], %c1{{[_0-9]*}} : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[D_1:.*]] = memref.dim %[[D]], %c1{{[_0-9]*}} : memref<?x?xf32, strided<[?, ?]>>
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_0]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[D_1]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_1]] step %{{.*}} {
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// -----
func.func @f3(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
%0 = memref.dim %D, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %C, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%5 = memref.subview %D[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %C[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f3
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
// CHECK: %[[D_0:.*]] = memref.dim %[[D]], %[[C0]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[D_1:.*]] = memref.dim %[[D]], %[[C1]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[C_1:.*]] = memref.dim %[[C]], %[[C1]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[D_0]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_1]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[D_1]] step %{{.*}} {
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// -----
func.func @f4(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%D : memref<?x?xf32, strided<[?, ?], offset: 0>>)
%0 = memref.dim %C, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %C, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%5 = memref.subview %C[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %D[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f4
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
// CHECK: %[[C_0:.*]] = memref.dim %[[C]], %[[C0:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[C_1:.*]] = memref.dim %[[C]], %[[C1:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[D_1:.*]] = memref.dim %[[D]], %[[C1:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_0]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[D_1]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_1]] step %{{.*}} {
// Fuse D then fuse C, no false dependence prevent it.
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// -----
func.func @f5(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%0 = memref.dim %B, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %D, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
linalg.matmul ins(%C, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%D : memref<?x?xf32, strided<[?, ?], offset: 0>>)
scf.for %arg5 = %c0 to %1 step %c2 {
scf.for %arg6 = %c0 to %0 step %c3 {
scf.for %arg7 = %c0 to %2 step %c4 {
%5 = memref.subview %D[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %B[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-DAG: #[[BOUND_2_MAP:.+]] = affine_map<(d0)[s0] -> (-d0 + s0, 2)>
// CHECK-DAG: #[[BOUND_2_MAP_2:.+]] = affine_map<(d0)[s0, s1] -> (-d0 + s1, -d0 + s0, 2)>
// CHECK-DAG: #[[BOUND_4_MAP:.+]] = affine_map<(d0)[s0] -> (-d0 + s0, 4)>
// CHECK: func @f5
// CHECK-SAME: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
// CHECK-DAG: %[[A_0:.*]] = memref.dim %[[A]], %[[C0]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[B_1:.*]] = memref.dim %[[B]], %[[C1]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[C_0:.*]] = memref.dim %[[C]], %[[C0]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[D_0:.*]] = memref.dim %[[D]], %[[C0]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[D_1:.*]] = memref.dim %[[D]], %[[C1]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK-DAG: %[[B_00:.*]] = memref.subview %[[B]][0, 0]{{.*}}
// CHECK: scf.for %[[I:.*]] = %{{.*}} to %[[D_0]] step %{{.*}} {
// CHECK: %[[BOUND_2_C0:.+]] = affine.min #[[BOUND_2_MAP]](%[[I]])[%[[C_0]]]
// CHECK: %[[C_I0:.*]] = memref.subview %[[C]][%[[I]], 0] [%[[BOUND_2_C0]]
// CHECK: %[[BOUND_ID_C0:.+]] = affine.min #[[BOUND_2_MAP_2]](%[[I]])[%[[A_0]], %[[C_0]]]
// CHECK: %[[A_I0:.*]] = memref.subview %[[A]][%[[I]], 0]
// CHECK: %[[C_I0_OUT:.*]] = memref.subview %[[C]][%[[I]], 0] [%[[BOUND_ID_C0]]
// CHECK: scf.for %[[J:.*]] = %{{.*}} to %[[B_1]] step %{{.*}} {
// CHECK: %[[E_IJ:.*]] = memref.subview %[[E]][%[[I]], %[[J]]]
// CHECK: scf.for %[[K:.*]] = %{{.*}} to %[[D_1]] step %{{.*}} {
// CHECK: %[[D_IK:.*]] = memref.subview %[[D]][%[[I]], %[[K]]] [2, 4]
// CHECK: %[[B_KJ:.*]] = memref.subview %[[B]][%[[K]], %[[J]]]
// CHECK: %[[BOUND_4_B1:.*]] = affine.min #[[BOUND_4_MAP]](%[[K]])[%[[B_1]]]
// CHECK: %[[B_0K:.*]] = memref.subview %[[B]][0, %[[K]]]
// CHECK: %[[D_IK_OUT:.+]] = memref.subview %[[D]][%[[I]], %[[K]]] [%[[BOUND_2_C0]], %[[BOUND_4_B1]]]
// CHECK: linalg.matmul ins(%[[A_I0]], %[[B_00]]{{.*}} outs(%[[C_I0_OUT]]
// CHECK: linalg.matmul ins(%[[C_I0]], %[[B_0K]]{{.*}} outs(%[[D_IK_OUT]]
// CHECK: linalg.matmul ins(%[[D_IK]], %[[B_KJ]]{{.*}} outs(%[[E_IJ]]
// -----
#map0 = affine_map<(d0) -> (d0 + 2)>
#map1 = affine_map<(d0) -> (d0 + 4)>
#map2 = affine_map<(d0) -> (d0 + 3)>
func.func @f6(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%0 = memref.dim %C, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
linalg.matmul ins(%A, %C : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%E : memref<?x?xf32, strided<[?, ?], offset: 0>>)
%1 = memref.dim %C, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg5 = %c0 to %1 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %0 step %c4 {
%3 = affine.apply #map0(%arg5)
%4 = affine.apply #map1(%arg7)
%5 = memref.subview %C[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%6 = affine.apply #map2(%arg6)
%7 = memref.subview %D[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f6
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// Fuse the producer of E (WAW) then the producer of C (WAR).
// CHECK: scf.for
// CHECK: scf.for
// CHECK: scf.for
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// -----
func.func @f7(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%0 = memref.dim %A, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %A, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%2 = memref.dim %C, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%3 = memref.dim %C, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%4 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
linalg.matmul ins(%A, %C : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%E : memref<?x?xf32, strided<[?, ?], offset: 0>>)
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%7 = memref.subview %A[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%9 = memref.subview %C[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%10 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%7, %9 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%10 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
scf.for %arg5 = %c0 to %3 step %c2 {
scf.for %arg6 = %c0 to %4 step %c3 {
scf.for %arg7 = %c0 to %2 step %c4 {
%7 = memref.subview %C[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%9 = memref.subview %D[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%10 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%7, %9 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%10 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f7
// CHECK: (%[[A:.*]]:{{.*}}, %[[B:.*]]:{{.*}}, %[[C:.*]]:{{.*}}, %[[D:.*]]:{{.*}}, %[[E:.*]]:{{.*}})
// CHECK-DAG: %[[C0:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
// CHECK: %[[A_0:.*]] = memref.dim %[[A]], %[[C0:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[A_1:.*]] = memref.dim %[[A]], %[[C1:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[C_1:.*]] = memref.dim %[[C]], %[[C1:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[C_0:.*]] = memref.dim %[[C]], %[[C0:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: %[[D_1:.*]] = memref.dim %[[D]], %[[C1:.*]] : memref<?x?xf32, strided<[?, ?]>>
// CHECK: linalg.matmul ins(%[[A]], %[[C]]{{.*}} outs(%[[E]]
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[A_0]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_1]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[A_1]] step %{{.*}} {
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_0]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[D_1]] step %{{.*}} {
// CHECK: scf.for %{{.*}} = %{{.*}} to %[[C_1]] step %{{.*}} {
// CHECK: linalg.matmul
// CHECK-NOT: linalg.matmul
// -----
#map0 = affine_map<(d0) -> (d0 + 2)>
#map1 = affine_map<(d0) -> (d0 + 4)>
#map2 = affine_map<(d0) -> (d0 + 3)>
func.func @f8(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%E: memref<?x?xf32, strided<[?, ?], offset: 0>>
) -> memref<?x?xf32, strided<[?, ?], offset: 0>> {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c4 = arith.constant 4 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%0 = memref.dim %A, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %A, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
linalg.matmul ins(%A, %C : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%D : memref<?x?xf32, strided<[?, ?], offset: 0>>)
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
%2 = memref.dim %D, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg5 = %c0 to %0 step %c2 {
scf.for %arg6 = %c0 to %2 step %c3 {
scf.for %arg7 = %c0 to %1 step %c4 {
%3 = affine.apply #map0(%arg5)
%4 = affine.apply #map1(%arg7)
%5 = memref.subview %A[%arg5, %arg7][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%6 = affine.apply #map2(%arg6)
%7 = memref.subview %D[%arg7, %arg6][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%8 = memref.subview %E[%arg5, %arg6][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%5, %7 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%8 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return %E : memref<?x?xf32, strided<[?, ?], offset: 0>>
}
// CHECK-LABEL: func @f8
// CHECK: (%[[A:.*]]: memref{{.*}}, %[[B:.*]]: memref{{.*}}, %[[C:.*]]: memref{{.*}}, %[[D:.*]]: memref{{.*}}, %[[E:.*]]: memref{{.*}})
// CHECK: linalg.matmul
// CHECK: linalg.matmul
// CHECK: scf.for
// CHECK: scf.for
// CHECK: scf.for
// CHECK: linalg.matmul
// CHECK-NOT: linalg.matmul
// -----
#id_2d = affine_map<(i, j) -> (i, j)>
#pointwise_2d_trait = {
indexing_maps = [#id_2d, #id_2d, #id_2d],
iterator_types = ["parallel", "parallel"]
}
func.func @pointwise(%A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%B: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%C: memref<?x?xf32, strided<[?, ?], offset: 0>>,
%D: memref<?x?xf32, strided<[?, ?], offset: 0>>) {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
linalg.generic #pointwise_2d_trait
ins(%A, %A: memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%B : memref<?x?xf32, strided<[?, ?], offset: 0>>) {
^bb0(%E: f32, %arg5: f32, %arg6: f32):
%2 = arith.addf %E, %arg5 : f32
linalg.yield %2 : f32
}
%0 = memref.dim %B, %c0 : memref<?x?xf32, strided<[?, ?], offset: 0>>
%1 = memref.dim %B, %c1 : memref<?x?xf32, strided<[?, ?], offset: 0>>
scf.for %arg4 = %c0 to %0 step %c2 {
scf.for %arg5 = %c0 to %1 step %c3 {
%4 = memref.subview %B[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%5 = memref.subview %C[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%6 = memref.subview %D[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.generic #pointwise_2d_trait
ins(%4, %5: memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%6 : memref<?x?xf32, strided<[?, ?], offset: ?>>) {
^bb0(%arg6: f32, %arg7: f32, %arg8: f32):
%7 = arith.mulf %arg6, %arg7 : f32
linalg.yield %7 : f32
}
}
}
return
}
// CHECK-LABEL: func @pointwise
// CHECK: scf.for
// CHECK: scf.for
// CHECK-NOT: scf.for
// CHECK: linalg.generic
// CHECK: arith.addf
// CHECK: linalg.generic
// CHECK: arith.mulf
// -----
#id_2d = affine_map<(i, j) -> (i, j)>
#pointwise_2d_trait = {
indexing_maps = [#id_2d, #id_2d, #id_2d],
iterator_types = ["parallel", "parallel"]
}
func.func @pointwise_no_view(%M: index, %N: index) {
%c1 = arith.constant 1 : index
%c0 = arith.constant 0 : index
%c3 = arith.constant 3 : index
%c2 = arith.constant 2 : index
%A = memref.alloc (%M, %N): memref<?x?xf32>
%B = memref.alloc (%M, %N): memref<?x?xf32>
%C = memref.alloc (%M, %N): memref<?x?xf32>
%D = memref.alloc (%M, %N): memref<?x?xf32>
%E = memref.alloc (%M, %N): memref<?x?xf32>
linalg.generic #pointwise_2d_trait
ins(%A, %A : memref<?x?xf32>, memref<?x?xf32>)
outs(%B : memref<?x?xf32>) {
^bb0(%e: f32, %arg5: f32, %arg6: f32):
%2 = arith.addf %e, %arg5 : f32
linalg.yield %2 : f32
}
%0 = memref.dim %B, %c0 : memref<?x?xf32>
%1 = memref.dim %B, %c1 : memref<?x?xf32>
scf.for %arg4 = %c0 to %0 step %c2 {
scf.for %arg5 = %c0 to %1 step %c3 {
%4 = memref.subview %B[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%5 = memref.subview %C[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%6 = memref.subview %D[%arg4, %arg5][%c2, %c3][%c1, %c1] :
memref<?x?xf32> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.generic #pointwise_2d_trait
ins(%4, %5: memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%6 : memref<?x?xf32, strided<[?, ?], offset: ?>>) {
^bb0(%arg6: f32, %arg7: f32, %arg8: f32):
%7 = arith.mulf %arg6, %arg7 : f32
linalg.yield %7 : f32
}
}
}
return
}
// CHECK-LABEL: func @pointwise_no_view
// CHECK: scf.for
// CHECK: scf.for
// CHECK-NOT: scf.for
// CHECK: linalg.generic
// CHECK: arith.addf
// CHECK: linalg.generic
// CHECK: arith.mulf
// -----
#map0 = affine_map<(d0, d1) -> (d0)>
#map1 = affine_map<(d0, d1) -> (d0, d1)>
func.func @fusion_of_three(%arg0: memref<100x10xf32>,
%arg1: memref<100xf32>,
%arg2: memref<100x10xf32>) {
%c0 = arith.constant 0 : index
%c1 = arith.constant 1 : index
%0 = memref.alloc() {temp = true} : memref<100x10xf32>
linalg.generic {
indexing_maps = [#map0, #map1],
iterator_types = ["parallel", "parallel"]}
ins(%arg1 : memref<100xf32>)
outs(%0 : memref<100x10xf32>) {
^bb0(%arg3: f32, %arg4: f32):
linalg.yield %arg3 : f32
}
%1 = memref.alloc() {temp = true} : memref<100x10xf32>
linalg.generic {
indexing_maps = [#map1, #map1, #map1],
iterator_types = ["parallel", "parallel"]}
ins(%arg0, %0: memref<100x10xf32>, memref<100x10xf32>)
outs(%1 : memref<100x10xf32>) {
^bb0(%arg3: f32, %arg4: f32, %arg5: f32):
%2 = arith.subf %arg3, %arg4 : f32
linalg.yield %2 : f32
}
memref.dealloc %0 : memref<100x10xf32>
%2 = memref.dim %1, %c0 : memref<100x10xf32>
%3 = memref.dim %1, %c1 : memref<100x10xf32>
%4 = memref.dim %arg2, %c0 : memref<100x10xf32>
%5 = memref.dim %arg2, %c1 : memref<100x10xf32>
scf.for %i = %c0 to %2 step %c1 {
scf.for %j = %c0 to %3 step %c1 {
%6 = memref.subview %1[%i, %j][%c1, %c1][%c1, %c1] :
memref<100x10xf32> to memref<?x?xf32, strided<[?, ?], offset: ?>>
%7 = memref.subview %arg2[%i, %j][%c1, %c1][%c1, %c1] :
memref<100x10xf32> to memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.generic {
indexing_maps = [#map1, #map1],
iterator_types = ["parallel", "parallel"]}
ins(%6 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%7 : memref<?x?xf32, strided<[?, ?], offset: ?>>) {
^bb0(%arg3: f32, %arg4: f32):
%8 = math.exp %arg3 : f32
linalg.yield %8 : f32
}
}
}
memref.dealloc %1 : memref<100x10xf32>
return
}
// CHECK-LABEL: func @fusion
// CHECK-NOT: linalg.generic
// CHECK: scf.for
// CHECK: scf.for
// CHECK-NOT: scf.for
// CHECK: linalg.generic
// CHECK: linalg.yield
// CHECK: linalg.generic
// CHECK: arith.subf
// CHECK: linalg.yield
// CHECK: linalg.generic
// CHECK: exp
// CHECK: linalg.yield
// -----
#map0 = affine_map<(d0)[s0] -> (2, -d0 + s0)>
#map1 = affine_map<(d0)[s0] -> (3, -d0 + s0)>
#map3 = affine_map<(d0)[s0, s1] -> (s0 + 1, -d0 + s0 + s1)>
#map4 = affine_map<(d0)[s0, s1] -> (s0 + 2, -d0 + s0 + s1)>
func.func @fill_and_conv(%arg0: memref<?x?xf32>, %arg1: memref<?x?xf32>, %arg2: memref<?x?xf32>) {
%cst = arith.constant 0.000000e+00 : f32
%c2 = arith.constant 2 : index
%c3 = arith.constant 3 : index
%c0 = arith.constant 0 : index
%c1 = arith.constant 1 : index
linalg.fill ins(%cst : f32) outs(%arg0 : memref<?x?xf32>)
%2 = memref.dim %arg1, %c0 : memref<?x?xf32>
%3 = memref.dim %arg1, %c1 : memref<?x?xf32>
%4 = memref.dim %arg2, %c0 : memref<?x?xf32>
%5 = memref.dim %arg2, %c1 : memref<?x?xf32>
scf.for %arg3 = %c0 to %4 step %c2 {
scf.for %arg4 = %c0 to %5 step %c3 {
%6 = affine.min #map3(%arg3)[%2, %4]
%7 = affine.min #map4(%arg4)[%3, %5]
%8 = memref.subview %arg0[%arg3, %arg4] [%6, %7] [1, 1] : memref<?x?xf32> to memref<?x?xf32, strided<[?, 1], offset: ?>>
%9 = affine.min #map0(%arg3)[%4]
%10 = affine.min #map1(%arg4)[%5]
%11 = memref.subview %arg2[%arg3, %arg4] [%9, %10] [1, 1] : memref<?x?xf32> to memref<?x?xf32, strided<[?, 1], offset: ?>>
linalg.conv_2d ins(%8, %arg1 : memref<?x?xf32, strided<[?, 1], offset: ?>>, memref<?x?xf32>) outs(%11 : memref<?x?xf32, strided<[?, 1], offset: ?>>)
}
}
return
}
// CHECK-LABEL: func @fill_and_conv
// CHECK: scf.for
// CHECK: scf.for
// CHECK: linalg.fill
// CHECK: linalg.conv_2d
// -----
// Test that different allocation-like ops are recognized and properly handled.
func.func @accept_different_alloc_ops(%dim: index, %s0 : index, %s1: index) {
%c0 = arith.constant 0 : index
%c1 = arith.constant 1 : index
%c2 = arith.constant 2 : index
%c3 = arith.constant 3 : index
%c4 = arith.constant 4 : index
%A = memref.alloca(%dim, %dim)[%s0, %s1] : memref<?x?xf32, strided<[?, ?], offset: 0>>
%B = memref.alloca(%dim, %dim)[%s0, %s1] : memref<?x?xf32, strided<[?, ?], offset: 0>>
%C = memref.alloc(%dim, %dim)[%s0, %s1] : memref<?x?xf32, strided<[?, ?], offset: 0>>
linalg.matmul ins(%A, %B : memref<?x?xf32, strided<[?, ?], offset: 0>>,
memref<?x?xf32, strided<[?, ?], offset: 0>>)
outs(%C : memref<?x?xf32, strided<[?, ?], offset: 0>>)
scf.for %i = %c0 to %dim step %c2 {
scf.for %j = %c0 to %dim step %c3 {
scf.for %k = %c0 to %dim step %c4 {
%0 = memref.subview %A[%i, %k][%c2, %c4][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%1 = memref.subview %B[%k, %j][%c4, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
%2 = memref.subview %C[%i, %j][%c2, %c3][%c1, %c1] :
memref<?x?xf32, strided<[?, ?], offset: 0>> to
memref<?x?xf32, strided<[?, ?], offset: ?>>
linalg.matmul ins(%0, %1 : memref<?x?xf32, strided<[?, ?], offset: ?>>,
memref<?x?xf32, strided<[?, ?], offset: ?>>)
outs(%2 : memref<?x?xf32, strided<[?, ?], offset: ?>>)
}
}
}
return
}
// CHECK-LABEL: func @accept_different_alloc_ops
// CHECK-COUNT-3: scf.for
// CHECK-COUNT-2: linalg.matmul
|
