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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -o - %s | FileCheck %s
target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128"
target triple = "arm64-apple-ios"
; CHECK-LABEL: lCPI0_0:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 16 ; 0x10
; CHECK-NEXT: .byte 20 ; 0x14
; CHECK-NEXT: .byte 24 ; 0x18
; CHECK-NEXT: .byte 28 ; 0x1c
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; It's profitable to convert the fptoui float -> i8 to first convert from
; float -> i32 and then use tbl for the truncate in a loop, so the mask can be
; materialized outside the loop.
define void @fptoui_v8f32_to_v8i8_in_loop(ptr %A, ptr %dst) {
; CHECK-LABEL: fptoui_v8f32_to_v8i8_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh0:
; CHECK-NEXT: adrp x9, lCPI0_0@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh1:
; CHECK-NEXT: ldr q0, [x9, lCPI0_0@PAGEOFF]
; CHECK-NEXT: LBB0_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: add x9, x0, x8, lsl #5
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: ldp q2, q1, [x9]
; CHECK-NEXT: fcvtzu.4s v4, v1
; CHECK-NEXT: fcvtzu.4s v3, v2
; CHECK-NEXT: tbl.16b v1, { v3, v4 }, v0
; CHECK-NEXT: str d1, [x1], #16
; CHECK-NEXT: b.eq LBB0_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh0, Lloh1
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <8 x float>, ptr %A, i64 %iv
%l.A = load <8 x float>, ptr %gep.A
%c = fptoui <8 x float> %l.A to <8 x i8>
%gep.dst = getelementptr inbounds <16 x i8>, ptr %dst, i64 %iv
store <8 x i8> %c, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; Not profitable to use tbl, as materializing the masks requires more
; instructions.
define void @fptoui_v8f32_to_v8i8_no_loop(ptr %A, ptr %dst) {
; CHECK-LABEL: fptoui_v8f32_to_v8i8_no_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: ldp q0, q1, [x0]
; CHECK-NEXT: fcvtzs.4s v0, v0
; CHECK-NEXT: fcvtzs.4s v1, v1
; CHECK-NEXT: xtn.4h v0, v0
; CHECK-NEXT: xtn.4h v1, v1
; CHECK-NEXT: uzp1.8b v0, v0, v1
; CHECK-NEXT: str d0, [x1]
; CHECK-NEXT: ret
entry:
%l.A = load <8 x float>, ptr %A
%c = fptoui <8 x float> %l.A to <8 x i8>
store <8 x i8> %c, ptr %dst
ret void
}
; CHECK-LABEL: lCPI2_0:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 16 ; 0x10
; CHECK-NEXT: .byte 20 ; 0x14
; CHECK-NEXT: .byte 24 ; 0x18
; CHECK-NEXT: .byte 28 ; 0x1c
; CHECK-NEXT: .byte 32 ; 0x20
; CHECK-NEXT: .byte 36 ; 0x24
; CHECK-NEXT: .byte 40 ; 0x28
; CHECK-NEXT: .byte 44 ; 0x2c
; CHECK-NEXT: .byte 48 ; 0x30
; CHECK-NEXT: .byte 52 ; 0x34
; CHECK-NEXT: .byte 56 ; 0x38
; CHECK-NEXT: .byte 60 ; 0x3c
; Tbl can also be used when combining multiple fptoui using a shuffle. The loop
; vectorizer may create such patterns.
define void @fptoui_2x_v8f32_to_v8i8_in_loop(ptr %A, ptr %B, ptr %dst) {
; CHECK-LABEL: fptoui_2x_v8f32_to_v8i8_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh2:
; CHECK-NEXT: adrp x9, lCPI2_0@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh3:
; CHECK-NEXT: ldr q0, [x9, lCPI2_0@PAGEOFF]
; CHECK-NEXT: LBB2_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: lsl x9, x8, #5
; CHECK-NEXT: add x10, x0, x9
; CHECK-NEXT: add x9, x1, x9
; CHECK-NEXT: ldp q2, q1, [x10]
; CHECK-NEXT: ldp q4, q3, [x9]
; CHECK-NEXT: fcvtzu.4s v17, v1
; CHECK-NEXT: fcvtzu.4s v16, v2
; CHECK-NEXT: fcvtzu.4s v19, v3
; CHECK-NEXT: fcvtzu.4s v18, v4
; CHECK-NEXT: tbl.16b v1, { v16, v17, v18, v19 }, v0
; CHECK-NEXT: str q1, [x2, x8, lsl #4]
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: b.eq LBB2_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh2, Lloh3
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <8 x float>, ptr %A, i64 %iv
%gep.B = getelementptr inbounds <8x float>, ptr %B, i64 %iv
%l.A = load <8 x float>, ptr %gep.A
%l.B = load <8 x float>, ptr %gep.B
%c1 = fptoui <8 x float> %l.A to <8 x i8>
%c2 = fptoui <8 x float> %l.B to <8 x i8>
%s = shufflevector <8 x i8> %c1, <8 x i8> %c2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%gep.dst = getelementptr inbounds <16 x i8>, ptr %dst, i64 %iv
store <16 x i8> %s, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; CHECK-LABEL: lCPI3_0:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 36 ; 0x24
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 16 ; 0x10
; CHECK-NEXT: .byte 20 ; 0x14
; CHECK-NEXT: .byte 24 ; 0x18
; CHECK-NEXT: .byte 44 ; 0x2c
; CHECK-NEXT: .byte 32 ; 0x20
; CHECK-NEXT: .byte 36 ; 0x24
; CHECK-NEXT: .byte 40 ; 0x28
; CHECK-NEXT: .byte 44 ; 0x2c
; CHECK-NEXT: .byte 48 ; 0x30
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 56 ; 0x38
; CHECK-NEXT: .byte 60 ; 0x3c
define void @fptoui_2x_v8f32_to_v8i8_in_loop_no_concat_shuffle(ptr %A, ptr %B, ptr %dst) {
; CHECK-LABEL: fptoui_2x_v8f32_to_v8i8_in_loop_no_concat_shuffle:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh4:
; CHECK-NEXT: adrp x9, lCPI3_0@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh5:
; CHECK-NEXT: ldr q0, [x9, lCPI3_0@PAGEOFF]
; CHECK-NEXT: LBB3_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: lsl x9, x8, #5
; CHECK-NEXT: add x10, x0, x9
; CHECK-NEXT: add x9, x1, x9
; CHECK-NEXT: ldp q2, q1, [x10]
; CHECK-NEXT: ldp q4, q3, [x9]
; CHECK-NEXT: fcvtzu.4s v17, v1
; CHECK-NEXT: fcvtzu.4s v16, v2
; CHECK-NEXT: fcvtzu.4s v19, v3
; CHECK-NEXT: fcvtzu.4s v18, v4
; CHECK-NEXT: tbl.16b v1, { v16, v17, v18, v19 }, v0
; CHECK-NEXT: str q1, [x2, x8, lsl #4]
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: b.eq LBB3_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh4, Lloh5
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <8 x float>, ptr %A, i64 %iv
%gep.B = getelementptr inbounds <8x float>, ptr %B, i64 %iv
%l.A = load <8 x float>, ptr %gep.A
%l.B = load <8 x float>, ptr %gep.B
%c1 = fptoui <8 x float> %l.A to <8 x i8>
%c2 = fptoui <8 x float> %l.B to <8 x i8>
%s = shufflevector <8 x i8> %c1, <8 x i8> %c2, <16 x i32> <i32 0, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 11, i32 8, i32 9, i32 10, i32 11, i32 12, i32 3, i32 14, i32 15>
%gep.dst = getelementptr inbounds <16 x i8>, ptr %dst, i64 %iv
store <16 x i8> %s, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; CHECK-LABEL: lCPI4_0:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 16 ; 0x10
; CHECK-NEXT: .byte 20 ; 0x14
; CHECK-NEXT: .byte 24 ; 0x18
; CHECK-NEXT: .byte 28 ; 0x1c
; CHECK-NEXT: .byte 32 ; 0x20
; CHECK-NEXT: .byte 36 ; 0x24
; CHECK-NEXT: .byte 40 ; 0x28
; CHECK-NEXT: .byte 44 ; 0x2c
; CHECK-NEXT: .byte 48 ; 0x30
; CHECK-NEXT: .byte 52 ; 0x34
; CHECK-NEXT: .byte 56 ; 0x38
; CHECK-NEXT: .byte 60 ; 0x3c
define void @fptoui_v16f32_to_v16i8_in_loop(ptr %A, ptr %dst) {
; CHECK-LABEL: fptoui_v16f32_to_v16i8_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh6:
; CHECK-NEXT: adrp x9, lCPI4_0@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh7:
; CHECK-NEXT: ldr q0, [x9, lCPI4_0@PAGEOFF]
; CHECK-NEXT: LBB4_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: add x9, x0, x8, lsl #6
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: ldp q2, q1, [x9, #32]
; CHECK-NEXT: ldp q4, q3, [x9]
; CHECK-NEXT: fcvtzu.4s v19, v1
; CHECK-NEXT: fcvtzu.4s v18, v2
; CHECK-NEXT: fcvtzu.4s v17, v3
; CHECK-NEXT: fcvtzu.4s v16, v4
; CHECK-NEXT: tbl.16b v1, { v16, v17, v18, v19 }, v0
; CHECK-NEXT: str q1, [x1], #32
; CHECK-NEXT: b.eq LBB4_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh6, Lloh7
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <16 x float>, ptr %A, i64 %iv
%l.A = load <16 x float>, ptr %gep.A
%c = fptoui <16 x float> %l.A to <16 x i8>
%gep.dst = getelementptr inbounds <32 x i8>, ptr %dst, i64 %iv
store <16 x i8> %c, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; CHECK-LABEL: lCPI5_0:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 16 ; 0x10
; CHECK-NEXT: .byte 20 ; 0x14
; CHECK-NEXT: .byte 24 ; 0x18
; CHECK-NEXT: .byte 28 ; 0x1c
; CHECK-NEXT: .byte 32 ; 0x20
; CHECK-NEXT: .byte 36 ; 0x24
; CHECK-NEXT: .byte 40 ; 0x28
; CHECK-NEXT: .byte 44 ; 0x2c
; CHECK-NEXT: .byte 48 ; 0x30
; CHECK-NEXT: .byte 52 ; 0x34
; CHECK-NEXT: .byte 56 ; 0x38
; CHECK-NEXT: .byte 60 ; 0x3c
define void @fptoui_2x_v16f32_to_v16i8_in_loop(ptr %A, ptr %B, ptr %dst) {
; CHECK-LABEL: fptoui_2x_v16f32_to_v16i8_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh8:
; CHECK-NEXT: adrp x9, lCPI5_0@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh9:
; CHECK-NEXT: ldr q0, [x9, lCPI5_0@PAGEOFF]
; CHECK-NEXT: LBB5_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: lsl x9, x8, #6
; CHECK-NEXT: add x10, x0, x9
; CHECK-NEXT: add x9, x1, x9
; CHECK-NEXT: ldp q1, q2, [x10, #32]
; CHECK-NEXT: ldp q3, q4, [x9, #32]
; CHECK-NEXT: fcvtzu.4s v24, v2
; CHECK-NEXT: fcvtzu.4s v23, v1
; CHECK-NEXT: ldp q5, q6, [x9]
; CHECK-NEXT: fcvtzu.4s v20, v4
; CHECK-NEXT: add x9, x2, x8, lsl #5
; CHECK-NEXT: fcvtzu.4s v19, v3
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: ldp q7, q16, [x10]
; CHECK-NEXT: fcvtzu.4s v18, v6
; CHECK-NEXT: fcvtzu.4s v17, v5
; CHECK-NEXT: fcvtzu.4s v22, v16
; CHECK-NEXT: fcvtzu.4s v21, v7
; CHECK-NEXT: tbl.16b v1, { v17, v18, v19, v20 }, v0
; CHECK-NEXT: tbl.16b v2, { v21, v22, v23, v24 }, v0
; CHECK-NEXT: stp q2, q1, [x9]
; CHECK-NEXT: b.eq LBB5_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh8, Lloh9
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <16 x float>, ptr %A, i64 %iv
%gep.B = getelementptr inbounds <16 x float>, ptr %B, i64 %iv
%l.A = load <16 x float>, ptr %gep.A
%l.B = load <16 x float>, ptr %gep.B
%c1 = fptoui <16 x float> %l.A to <16 x i8>
%c2 = fptoui <16 x float> %l.B to <16 x i8>
%s = shufflevector <16 x i8> %c1, <16 x i8> %c2, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
%gep.dst = getelementptr inbounds <32 x i8>, ptr %dst, i64 %iv
store <32 x i8> %s, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
define void @fptoui_v8f32_to_v8i16_in_loop(ptr %A, ptr %dst) {
; CHECK-LABEL: fptoui_v8f32_to_v8i16_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: LBB6_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: add x9, x0, x8, lsl #5
; CHECK-NEXT: ldp q0, q1, [x9]
; CHECK-NEXT: fcvtzu.4s v0, v0
; CHECK-NEXT: fcvtzu.4s v1, v1
; CHECK-NEXT: uzp1.8h v0, v0, v1
; CHECK-NEXT: str q0, [x1, x8, lsl #4]
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: b.eq LBB6_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <8 x float>, ptr %A, i64 %iv
%l.A = load <8 x float>, ptr %gep.A
%c = fptoui <8 x float> %l.A to <8 x i16>
%gep.dst = getelementptr inbounds <8 x i16>, ptr %dst, i64 %iv
store <8 x i16> %c, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
define void @fptoui_2x_v8f32_to_v8i16_in_loop(ptr %A, ptr %B, ptr %dst) {
; CHECK-LABEL: fptoui_2x_v8f32_to_v8i16_in_loop:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: LBB7_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: lsl x9, x8, #5
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: add x10, x0, x9
; CHECK-NEXT: add x11, x1, x9
; CHECK-NEXT: add x9, x2, x9
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: ldp q0, q1, [x10]
; CHECK-NEXT: fcvtzu.4s v0, v0
; CHECK-NEXT: ldp q2, q3, [x11]
; CHECK-NEXT: fcvtzu.4s v1, v1
; CHECK-NEXT: fcvtzu.4s v2, v2
; CHECK-NEXT: uzp1.8h v0, v0, v1
; CHECK-NEXT: fcvtzu.4s v3, v3
; CHECK-NEXT: uzp1.8h v1, v2, v3
; CHECK-NEXT: stp q0, q1, [x9]
; CHECK-NEXT: b.eq LBB7_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.A = getelementptr inbounds <8 x float>, ptr %A, i64 %iv
%gep.B = getelementptr inbounds <8 x float>, ptr %B, i64 %iv
%l.A = load <8 x float>, ptr %gep.A
%l.B = load <8 x float>, ptr %gep.B
%c1 = fptoui <8 x float> %l.A to <8 x i16>
%c2 = fptoui <8 x float> %l.B to <8 x i16>
%s = shufflevector <8 x i16> %c1, <8 x i16> %c2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%gep.dst = getelementptr inbounds <16 x i16>, ptr %dst, i64 %iv
store <16 x i16> %s, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; CHECK-LABEL: lCPI8_0:
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 5 ; 0x5
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 6 ; 0x6
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 7 ; 0x7
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: lCPI8_1:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 1 ; 0x1
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 2 ; 0x2
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 3 ; 0x3
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
define void @uitofp_v8i8_to_v8f32(ptr %src, ptr %dst) {
; CHECK-LABEL: uitofp_v8i8_to_v8f32:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh10:
; CHECK-NEXT: adrp x9, lCPI8_0@PAGE
; CHECK-NEXT: Lloh11:
; CHECK-NEXT: adrp x10, lCPI8_1@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh12:
; CHECK-NEXT: ldr q0, [x9, lCPI8_0@PAGEOFF]
; CHECK-NEXT: Lloh13:
; CHECK-NEXT: ldr q1, [x10, lCPI8_1@PAGEOFF]
; CHECK-NEXT: LBB8_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: ldr d2, [x0, x8, lsl #3]
; CHECK-NEXT: add x9, x1, x8, lsl #5
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: tbl.16b v3, { v2 }, v0
; CHECK-NEXT: tbl.16b v2, { v2 }, v1
; CHECK-NEXT: ucvtf.4s v3, v3
; CHECK-NEXT: ucvtf.4s v2, v2
; CHECK-NEXT: stp q2, q3, [x9]
; CHECK-NEXT: b.eq LBB8_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh11, Lloh13
; CHECK-NEXT: .loh AdrpLdr Lloh10, Lloh12
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.src = getelementptr inbounds <8 x i8>, ptr %src, i64 %iv
%l = load <8 x i8>, ptr %gep.src
%conv = uitofp <8 x i8> %l to <8 x float>
%gep.dst = getelementptr inbounds <8 x float>, ptr %dst, i64 %iv
store <8 x float> %conv, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
; CHECK-LABEL: lCPI9_0:
; CHECK-NEXT: .byte 12 ; 0xc
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 13 ; 0xd
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 14 ; 0xe
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 15 ; 0xf
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: lCPI9_1:
; CHECK-NEXT: .byte 8 ; 0x8
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 9 ; 0x9
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 10 ; 0xa
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 11 ; 0xb
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: lCPI9_2:
; CHECK-NEXT: .byte 4 ; 0x4
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 5 ; 0x5
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 6 ; 0x6
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 7 ; 0x7
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: lCPI9_3:
; CHECK-NEXT: .byte 0 ; 0x0
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 1 ; 0x1
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 2 ; 0x2
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 3 ; 0x3
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
; CHECK-NEXT: .byte 255 ; 0xff
define void @uitofp_v16i8_to_v16f32(ptr %src, ptr %dst) {
; CHECK-LABEL: uitofp_v16i8_to_v16f32:
; CHECK: ; %bb.0: ; %entry
; CHECK-NEXT: Lloh14:
; CHECK-NEXT: adrp x9, lCPI9_0@PAGE
; CHECK-NEXT: Lloh15:
; CHECK-NEXT: adrp x10, lCPI9_1@PAGE
; CHECK-NEXT: Lloh16:
; CHECK-NEXT: adrp x11, lCPI9_2@PAGE
; CHECK-NEXT: Lloh17:
; CHECK-NEXT: adrp x12, lCPI9_3@PAGE
; CHECK-NEXT: mov x8, xzr
; CHECK-NEXT: Lloh18:
; CHECK-NEXT: ldr q0, [x9, lCPI9_0@PAGEOFF]
; CHECK-NEXT: Lloh19:
; CHECK-NEXT: ldr q1, [x10, lCPI9_1@PAGEOFF]
; CHECK-NEXT: Lloh20:
; CHECK-NEXT: ldr q2, [x11, lCPI9_2@PAGEOFF]
; CHECK-NEXT: Lloh21:
; CHECK-NEXT: ldr q3, [x12, lCPI9_3@PAGEOFF]
; CHECK-NEXT: LBB9_1: ; %loop
; CHECK-NEXT: ; =>This Inner Loop Header: Depth=1
; CHECK-NEXT: ldr q4, [x0, x8, lsl #4]
; CHECK-NEXT: add x9, x1, x8, lsl #6
; CHECK-NEXT: add x8, x8, #1
; CHECK-NEXT: cmp x8, #1000
; CHECK-NEXT: tbl.16b v5, { v4 }, v0
; CHECK-NEXT: tbl.16b v6, { v4 }, v1
; CHECK-NEXT: tbl.16b v7, { v4 }, v2
; CHECK-NEXT: tbl.16b v4, { v4 }, v3
; CHECK-NEXT: ucvtf.4s v5, v5
; CHECK-NEXT: ucvtf.4s v6, v6
; CHECK-NEXT: ucvtf.4s v7, v7
; CHECK-NEXT: ucvtf.4s v4, v4
; CHECK-NEXT: stp q6, q5, [x9, #32]
; CHECK-NEXT: stp q4, q7, [x9]
; CHECK-NEXT: b.eq LBB9_1
; CHECK-NEXT: ; %bb.2: ; %exit
; CHECK-NEXT: ret
; CHECK-NEXT: .loh AdrpLdr Lloh17, Lloh21
; CHECK-NEXT: .loh AdrpLdr Lloh16, Lloh20
; CHECK-NEXT: .loh AdrpLdr Lloh15, Lloh19
; CHECK-NEXT: .loh AdrpLdr Lloh14, Lloh18
entry:
br label %loop
loop:
%iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
%gep.src = getelementptr inbounds <16 x i8>, ptr %src, i64 %iv
%l = load <16 x i8>, ptr %gep.src
%conv = uitofp <16 x i8> %l to <16 x float>
%gep.dst = getelementptr inbounds <16 x float>, ptr %dst, i64 %iv
store <16 x float> %conv, ptr %gep.dst
%iv.next = add i64 %iv, 1
%ec = icmp eq i64 %iv.next, 1000
br i1 %ec, label %loop, label %exit
exit:
ret void
}
|
