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
|
// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s
// Note: unlike the 'private' recipe checks, this is just for spot-checking,
// so this test isn't as comprehensive. The same code paths are used for
// 'private', so we just count on those to catch the errors.
struct NoOps {
int i;
~NoOps();
};
void do_things(unsigned A, unsigned B) {
NoOps ThreeArr[5][5][5];
#pragma acc parallel reduction(+:ThreeArr[B][B][B])
// CHECK:acc.reduction.recipe @reduction_add__Bcnt3__ZTSA5_A5_A5_5NoOps : !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> reduction_operator <add> init {
// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: %[[TL_ALLOCA:.*]] = cir.alloca !cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>, !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>>, ["openacc.reduction.init"] {alignment = 4 : i64}
//
// Init Section:
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB3_CAST]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR3_LOAD]], %[[UB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_DECAY:.*]] = cir.cast array_to_ptrdecay %[[TL_ALLOCA]] : !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_DECAY]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB2_CAST]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR2_LOAD]], %[[UB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> -> !cir.ptr<!cir.array<!rec_NoOps x 5>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_DECAY]] : !cir.ptr<!cir.array<!rec_NoOps x 5>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.array<!rec_NoOps x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB1_CAST]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR1_LOAD]], %[[UB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND2_STRIDE]] : !cir.ptr<!cir.array<!rec_NoOps x 5>> -> !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[BOUND1_STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_DECAY]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[BOUND1_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i>
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i
// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: acc.yield
// CHECK-NEXT: } combiner {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: acc.yield
// CHECK-NEXT:} destroy {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[LAST_SUB_ONE:.*]] = cir.binop(sub, %[[UB3_CAST]], %[[ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[LAST_SUB_ONE]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR3_LOAD]], %[[LB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_DECAY:.*]] = cir.cast array_to_ptrdecay %[[PRIVATE]] : !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_DECAY]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[LAST_SUB_ONE:.*]] = cir.binop(sub, %[[UB2_CAST]], %[[ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[LAST_SUB_ONE]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR2_LOAD]], %[[LB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> -> !cir.ptr<!cir.array<!rec_NoOps x 5>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_DECAY]] : !cir.ptr<!cir.array<!rec_NoOps x 5>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.array<!rec_NoOps x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[LAST_SUB_ONE:.*]] = cir.binop(sub, %[[UB1_CAST]], %[[ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[LAST_SUB_ONE]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR1_LOAD]], %[[LB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND2_STRIDE]] : !cir.ptr<!cir.array<!rec_NoOps x 5>> -> !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[BOUND1_STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_DECAY]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: cir.call @_ZN5NoOpsD1Ev(%[[BOUND1_STRIDE]]) nothrow : (!cir.ptr<!rec_NoOps>) -> ()
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT:acc.yield
// CHECK-NEXT:}
;
NoOps ***ThreePtr;
#pragma acc parallel reduction(*:ThreePtr[B][B][A:B])
// CHECK: acc.reduction.recipe @reduction_mul__Bcnt3__ZTSPPP5NoOps : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> reduction_operator <mul> init {
// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: %[[TOP_LEVEL_ALLOCA:.*]] = cir.alloca !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, ["openacc.reduction.init"] {alignment = 8 : i64}
//
// CHECK-NEXT: %[[INT_PTR_PTR_PTR_UPPER_BOUND:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UPPER_BOUND_CAST:.*]] = builtin.unrealized_conversion_cast %[[INT_PTR_PTR_PTR_UPPER_BOUND]] : index to !u64i
// CHECK-NEXT: %[[SIZEOF_PTR:.*]] = cir.const #cir.int<8> : !u64i
// CHECK-NEXT: %[[CALC_ALLOCA_SIZE:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST]], %[[SIZEOF_PTR]]) : !u64i
// CHECK-NEXT: %[[INT_PTR_PTR_VLA_ALLOCA:.*]] = cir.alloca !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, %[[CALC_ALLOCA_SIZE]] : !u64i, ["openacc.init.bounds"] {alignment = 8 : i64}
//
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["itr"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[UPPER_LIMIT:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UPPER_LIMIT]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
//
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[SRC_IDX:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST]], %[[ITR_LOAD]]) : !u64i
// CHECK-NEXT: %[[SRC_STRIDE:.*]] = cir.ptr_stride(%[[INT_PTR_PTR_VLA_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, %[[SRC_IDX]] : !u64i), !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: %[[DEST_STRIDE:.*]] = cir.ptr_stride(%[[TOP_LEVEL_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, %[[ITR_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>
// CHECK-NEXT: cir.store %[[SRC_STRIDE]], %[[DEST_STRIDE]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>
// CHECK-NEXT: cir.yield
//
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
//
// CHECK-NEXT: %[[INT_PTR_PTR_UPPER_BOUND:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UPPER_BOUND_CAST_2:.*]] = builtin.unrealized_conversion_cast %[[INT_PTR_PTR_UPPER_BOUND]] : index to !u64i
// CHECK-NEXT: %[[NUM_ELTS:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST_2]], %[[UPPER_BOUND_CAST]]) : !u64i
// CHECK-NEXT: %[[SIZEOF_PTR_PTR:.*]] = cir.const #cir.int<8> : !u64i
// CHECK-NEXT: %[[CALC_ALLOCA_SIZE:.*]] = cir.binop(mul, %[[NUM_ELTS]], %[[SIZEOF_PTR_PTR]]) : !u64i
// CHECK-NEXT: %[[INT_PTR_PTR_ALLOCA:.*]] = cir.alloca !cir.ptr<!rec_NoOps>, !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[CALC_ALLOCA_SIZE]] : !u64i, ["openacc.init.bounds"] {alignment = 8 : i64}
//
//
// Copy array pointer to the original alloca.
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["itr"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UPPER_BOUND_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
//
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[SRC_IDX:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST_2]], %[[ITR_LOAD]]) : !u64i
// CHECK-NEXT: %[[SRC_STRIDE:.*]] = cir.ptr_stride(%[[INT_PTR_PTR_ALLOCA]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[SRC_IDX]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[DEST_STRIDE:.*]] = cir.ptr_stride(%[[INT_PTR_PTR_VLA_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, %[[ITR_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: cir.store %[[SRC_STRIDE]], %[[DEST_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: cir.yield
//
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
//
// CHECK-NEXT: %[[INT_PTR_UPPER_BOUND:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UPPER_BOUND_CAST_3:.*]] = builtin.unrealized_conversion_cast %[[INT_PTR_UPPER_BOUND]] : index to !u64i
// CHECK-NEXT: %[[NUM_ELTS_2:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST_3]], %[[NUM_ELTS]]) : !u64i
// CHECK-NEXT: %[[SIZEOF_INT:.*]] = cir.const #cir.int<4> : !u64i
// CHECK-NEXT: %[[CALC_ALLOCA_SIZE:.*]] = cir.binop(mul, %[[NUM_ELTS_2]], %[[SIZEOF_INT]]) : !u64i
// CHECK-NEXT: %[[INT_PTR_ALLOCA:.*]] = cir.alloca !rec_NoOps, !cir.ptr<!rec_NoOps>, %[[CALC_ALLOCA_SIZE]] : !u64i, ["openacc.init.bounds"] {alignment = 4 : i64}
//
// Copy array pointer to the original alloca.
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["itr"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[NUM_ELTS]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
//
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[SRC_IDX:.*]] = cir.binop(mul, %[[UPPER_BOUND_CAST_3]], %[[ITR_LOAD]]) : !u64i
// CHECK-NEXT: %[[SRC_STRIDE:.*]] = cir.ptr_stride(%[[INT_PTR_ALLOCA]] : !cir.ptr<!rec_NoOps>, %[[SRC_IDX]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[DEST_STRIDE:.*]] = cir.ptr_stride(%[[INT_PTR_PTR_ALLOCA]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.store %[[SRC_STRIDE]], %[[DEST_STRIDE]] : !cir.ptr<!rec_NoOps>, !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.yield
//
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// Initialization Section
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB3_CAST]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR3_LOAD]], %[[UB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_LOAD:.*]] = cir.load %[[TOP_LEVEL_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_LOAD]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB2_CAST]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR2_LOAD]], %[[UB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_LOAD:.*]] = cir.load %[[BOUND3_STRIDE]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_LOAD]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB1_CAST]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR1_LOAD]], %[[UB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_LOAD:.*]] = cir.load %[[BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_LOAD]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i>
// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s32i
// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[GET_I]] : !s32i, !cir.ptr<!s32i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
//
// CHECK-NEXT: acc.yield
// CHECK-NEXT: } combiner {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: acc.yield
// CHECK-NEXT: } destroy {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB3:.*]] = cir.binop(sub, %[[UB3_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB3]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR3_LOAD]], %[[LB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_LOAD:.*]] = cir.load %[[PRIVATE]] : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_LOAD]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB2:.*]] = cir.binop(sub, %[[UB2_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB2]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR2_LOAD]], %[[LB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_LOAD:.*]] = cir.load %[[BOUND3_STRIDE]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_LOAD]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB1:.*]] = cir.binop(sub, %[[UB1_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB1]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR1_LOAD]], %[[LB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_LOAD:.*]] = cir.load %[[BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_LOAD]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: cir.call @_ZN5NoOpsD1Ev(%[[STRIDE]]) nothrow : (!cir.ptr<!rec_NoOps>) -> ()
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
//
// CHECK-NEXT: acc.yield
// CHECK-NEXT: }
;
using PtrTArrayTy = NoOps*[5];
PtrTArrayTy *PtrArrayPtr;
#pragma acc parallel reduction(||:PtrArrayPtr[B][B][B])
// CHECK-NEXT: acc.reduction.recipe @reduction_lor__Bcnt3__ZTSPA5_P5NoOps : !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> reduction_operator <lor> init {
// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: %[[TL_ALLOCA:.*]] = cir.alloca !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>, ["openacc.reduction.init"] {alignment = 8 : i64}
//
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ARR_SIZE:.*]] = cir.const #cir.int<40> : !u64i
// CHECK-NEXT: %[[ALLOCA_SIZE:.*]] = cir.binop(mul, %[[UB3_CAST]], %[[ARR_SIZE]]) : !u64i
// CHECK-NEXT: %[[ARR_ALLOCA:.*]] = cir.alloca !cir.array<!cir.ptr<!rec_NoOps> x 5>, !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, %[[ALLOCA_SIZE]] : !u64i, ["openacc.init.bounds"] {alignment = 8 : i64}
//
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["itr"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[UPP_BOUND:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UPP_BOUND]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[SRC_IDX:.*]] = cir.binop(mul, %[[UB3_CAST]], %[[ITR_LOAD]]) : !u64i
// CHECK-NEXT: %[[SRC:.*]] = cir.ptr_stride(%[[ARR_ALLOCA]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, %[[SRC_IDX]] : !u64i), !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>
// CHECK-NEXT: %[[DEST:.*]] = cir.ptr_stride(%[[TL_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>, %[[ITR_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>
// CHECK-NEXT: cir.store %[[SRC]], %[[DEST]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
//
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[NUM_ELTS:.*]] = cir.binop(mul, %[[UB2_CAST]], %[[UB3_CAST]]) : !u64i
//
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ARR_ALLOCA]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>> -> !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride(%[[DECAY]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ZERO]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
//
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[NUM_ELTS2:.*]] = cir.binop(mul, %[[UB1_CAST]], %[[NUM_ELTS]]) : !u64i
// CHECK-NEXT: %[[ELT_SIZE:.*]] = cir.const #cir.int<4> : !u64i
// CHECK-NEXT: %[[ALLOCA_SIZE:.*]] = cir.binop(mul, %[[NUM_ELTS2]], %[[ELT_SIZE]]) : !u64i
// CHECK-NEXT: %[[ARR_ALLOCA2:.*]] = cir.alloca !rec_NoOps, !cir.ptr<!rec_NoOps>, %[[ALLOCA_SIZE]] : !u64i, ["openacc.init.bounds"] {alignment = 4 : i64}
//
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["itr"] {alignment = 8 : i64}
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u64i
// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[NUM_ELTS]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[SRC_IDX:.*]] = cir.binop(mul, %[[UB1_CAST]], %[[ITR_LOAD]]) : !u64i
// CHECK-NEXT: %[[SRC:.*]] = cir.ptr_stride(%[[ARR_ALLOCA2]] : !cir.ptr<!rec_NoOps>, %[[SRC_IDX]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[DEST:.*]] = cir.ptr_stride(%[[STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.store %[[SRC]], %[[DEST]] : !cir.ptr<!rec_NoOps>, !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// Init Section:
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB3_CAST]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR3_LOAD]], %[[UB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[CMP]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_LOAD:.*]] = cir.load %[[TL_ALLOCA]] : !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>, !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_LOAD]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB2_CAST]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR2_LOAD]], %[[UB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>> -> !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_DECAY]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: cir.store %[[LB1_CAST]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR1_LOAD]], %[[UB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_LOAD:.*]] = cir.load %[[BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_LOAD]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i>
// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i
// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[INC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: acc.yield
// CHECK-NEXT: } combiner {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: acc.yield
// CHECK-NEXT: } destroy {
// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}):
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i
// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i
// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB3:.*]] = cir.binop(sub, %[[UB3_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB3]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
//
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR3_LOAD]], %[[LB3_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[TLA_LOAD:.*]] = cir.load %[[PRIVATE]] : !cir.ptr<!cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>>, !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>
// CHECK-NEXT: %[[BOUND3_STRIDE:.*]] = cir.ptr_stride(%[[TLA_LOAD]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>, %[[ITR3_LOAD]] : !u64i), !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i
// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i
// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB2:.*]] = cir.binop(sub, %[[UB2_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB2]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR2_LOAD]], %[[LB2_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.ptr<!rec_NoOps> x 5>> -> !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: %[[BOUND2_STRIDE:.*]] = cir.ptr_stride(%[[BOUND3_STRIDE_DECAY]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, %[[ITR2_LOAD]] : !u64i), !cir.ptr<!cir.ptr<!rec_NoOps>>
// CHECK-NEXT: cir.scope {
// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i
// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index
// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i
// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64}
// CHECK-NEXT: %[[CONST_ONE:.*]] = cir.const #cir.int<1> : !u64i
// CHECK-NEXT: %[[ONE_BELOW_UB1:.*]] = cir.binop(sub, %[[UB1_CAST]], %[[CONST_ONE]]) : !u64i
// CHECK-NEXT: cir.store %[[ONE_BELOW_UB1]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.for : cond {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[COND:.*]] = cir.cmp(ge, %[[ITR1_LOAD]], %[[LB1_CAST]]) : !u64i, !cir.bool
// CHECK-NEXT: cir.condition(%[[COND]])
// CHECK-NEXT: } body {
// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[BOUND2_STRIDE_LOAD:.*]] = cir.load %[[BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps>
// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride(%[[BOUND2_STRIDE_LOAD]] : !cir.ptr<!rec_NoOps>, %[[ITR1_LOAD]] : !u64i), !cir.ptr<!rec_NoOps>
// CHECK-NEXT: cir.call @_ZN5NoOpsD1Ev(%[[STRIDE]]) nothrow : (!cir.ptr<!rec_NoOps>) -> ()
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR1_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR2_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: cir.yield
// CHECK-NEXT: } step {
// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i
// CHECK-NEXT: %[[DEC:.*]] = cir.unary(dec, %[[ITR3_LOAD]]) : !u64i, !u64i
// CHECK-NEXT: cir.store %[[DEC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i>
// CHECK-NEXT: cir.yield
// CHECK-NEXT: }
// CHECK-NEXT: }
// CHECK-NEXT: acc.yield
// CHECK-NEXT: }
;
}
|
