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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64--linux-gnu -mattr=sve < %s | FileCheck %s
; NOTE: invalid, upper and lower bound immediate values of the regimm
; addressing mode are checked only for the byte version of each
; instruction (`ld<N>b`), as the code for detecting the immediate is
; common to all instructions, and varies only for the number of
; elements of the structure store, which is <N> = 2, 3, 4.
; ld2b
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #2
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 2
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8*
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #-16
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -16
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #14
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 14
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8_not_multiple_of_2(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8_not_multiple_of_2:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #3
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 3
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8_outside_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8_outside_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #-18
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -18
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8> } @ld2.nxv32i8_outside_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld2.nxv32i8_outside_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #16
; CHECK-NEXT: ld2b { z0.b, z1.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 16
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
; ld2h
define { <vscale x 8 x i16>, <vscale x 8 x i16> } @ld2.nxv16i16(<vscale x 8 x i1> %Pg, <vscale x 8 x i16>* %addr) {
; CHECK-LABEL: ld2.nxv16i16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #14
; CHECK-NEXT: ld2h { z0.h, z1.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %addr, i64 14
%base_ptr = bitcast <vscale x 8 x i16>* %base to i16 *
%res = call { <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld2.sret.nxv16i16.nxv8i1.p0i16(<vscale x 8 x i1> %Pg, i16 *%base_ptr)
ret { <vscale x 8 x i16>, <vscale x 8 x i16> } %res
}
define { <vscale x 8 x half>, <vscale x 8 x half> } @ld2.nxv16f16(<vscale x 8 x i1> %Pg, <vscale x 8 x half>* %addr) {
; CHECK-LABEL: ld2.nxv16f16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-16
; CHECK-NEXT: ld2h { z0.h, z1.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %addr, i64 -16
%base_ptr = bitcast <vscale x 8 x half>* %base to half *
%res = call { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld2.sret.nxv16f16.nxv8i1.p0f16(<vscale x 8 x i1> %Pg, half *%base_ptr)
ret { <vscale x 8 x half>, <vscale x 8 x half> } %res
}
define { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @ld2.nxv16bf16(<vscale x 8 x i1> %Pg, <vscale x 8 x bfloat>* %addr) #0 {
; CHECK-LABEL: ld2.nxv16bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #12
; CHECK-NEXT: ld2h { z0.h, z1.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %addr, i64 12
%base_ptr = bitcast <vscale x 8 x bfloat>* %base to bfloat *
%res = call { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld2.sret.nxv16bf16.nxv8i1.p0bf16(<vscale x 8 x i1> %Pg, bfloat *%base_ptr)
ret { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } %res
}
; ld2w
define { <vscale x 4 x i32>, <vscale x 4 x i32> } @ld2.nxv8i32(<vscale x 4 x i1> %Pg, <vscale x 4 x i32>* %addr) {
; CHECK-LABEL: ld2.nxv8i32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #14
; CHECK-NEXT: ld2w { z0.s, z1.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %addr, i64 14
%base_ptr = bitcast <vscale x 4 x i32>* %base to i32 *
%res = call { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld2.sret.nxv8i32.nxv4i1.p0i32(<vscale x 4 x i1> %Pg, i32 *%base_ptr)
ret { <vscale x 4 x i32>, <vscale x 4 x i32> } %res
}
define { <vscale x 4 x float>, <vscale x 4 x float> } @ld2.nxv8f32(<vscale x 4 x i1> %Pg, <vscale x 4 x float>* %addr) {
; CHECK-LABEL: ld2.nxv8f32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-16
; CHECK-NEXT: ld2w { z0.s, z1.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %addr, i64 -16
%base_ptr = bitcast <vscale x 4 x float>* %base to float *
%res = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld2.sret.nxv8f32.nxv4i1.p0f32(<vscale x 4 x i1> %Pg, float *%base_ptr)
ret { <vscale x 4 x float>, <vscale x 4 x float> } %res
}
; ld2d
define { <vscale x 2 x i64>, <vscale x 2 x i64> } @ld2.nxv4i64(<vscale x 2 x i1> %Pg, <vscale x 2 x i64>* %addr) {
; CHECK-LABEL: ld2.nxv4i64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #14
; CHECK-NEXT: ld2d { z0.d, z1.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %addr, i64 14
%base_ptr = bitcast <vscale x 2 x i64>* %base to i64 *
%res = call { <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld2.sret.nxv4i64.nxv2i1.p0i64(<vscale x 2 x i1> %Pg, i64 *%base_ptr)
ret { <vscale x 2 x i64>, <vscale x 2 x i64> } %res
}
define { <vscale x 2 x double>, <vscale x 2 x double> } @ld2.nxv4f64(<vscale x 2 x i1> %Pg, <vscale x 2 x double>* %addr) {
; CHECK-LABEL: ld2.nxv4f64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-16
; CHECK-NEXT: ld2d { z0.d, z1.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %addr, i64 -16
%base_ptr = bitcast <vscale x 2 x double>* %base to double *
%res = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld2.sret.nxv4f64.nxv2i1.p0f64(<vscale x 2 x i1> %Pg, double *%base_ptr)
ret { <vscale x 2 x double>, <vscale x 2 x double> } %res
}
; ld3b
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #3
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 3
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #-24
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -24
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #21
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 21
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_not_multiple_of_3_01(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_not_multiple_of_3_01:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #4
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 4
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_not_multiple_of_3_02(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_not_multiple_of_3_02:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #5
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 5
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_outside_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_outside_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #-27
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -27
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld3.nxv48i8_outside_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld3.nxv48i8_outside_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #24
; CHECK-NEXT: ld3b { z0.b, z1.b, z2.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 24
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
; ld3h
define { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @ld3.nxv24i16(<vscale x 8 x i1> %Pg, <vscale x 8 x i16> *%addr) {
; CHECK-LABEL: ld3.nxv24i16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #21
; CHECK-NEXT: ld3h { z0.h, z1.h, z2.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %addr, i64 21
%base_ptr = bitcast <vscale x 8 x i16>* %base to i16 *
%res = call { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld3.sret.nxv24i16.nxv8i1.p0i16(<vscale x 8 x i1> %Pg, i16 *%base_ptr)
ret { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } %res
}
define { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @ld3.nxv24f16(<vscale x 8 x i1> %Pg, <vscale x 8 x half> *%addr) {
; CHECK-LABEL: ld3.nxv24f16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #21
; CHECK-NEXT: ld3h { z0.h, z1.h, z2.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %addr, i64 21
%base_ptr = bitcast <vscale x 8 x half>* %base to half *
%res = call { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld3.sret.nxv24f16.nxv8i1.p0f16(<vscale x 8 x i1> %Pg, half *%base_ptr)
ret { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } %res
}
define { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @ld3.nxv24bf16(<vscale x 8 x i1> %Pg, <vscale x 8 x bfloat> *%addr) #0 {
; CHECK-LABEL: ld3.nxv24bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-24
; CHECK-NEXT: ld3h { z0.h, z1.h, z2.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %addr, i64 -24
%base_ptr = bitcast <vscale x 8 x bfloat>* %base to bfloat *
%res = call { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld3.sret.nxv24bf16.nxv8i1.p0bf16(<vscale x 8 x i1> %Pg, bfloat *%base_ptr)
ret { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } %res
}
; ld3w
define { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @ld3.nxv12i32(<vscale x 4 x i1> %Pg, <vscale x 4 x i32> *%addr) {
; CHECK-LABEL: ld3.nxv12i32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #21
; CHECK-NEXT: ld3w { z0.s, z1.s, z2.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %addr, i64 21
%base_ptr = bitcast <vscale x 4 x i32>* %base to i32 *
%res = call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld3.sret.nxv12i32.nxv4i1.p0i32(<vscale x 4 x i1> %Pg, i32 *%base_ptr)
ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } %res
}
define { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @ld3.nxv12f32(<vscale x 4 x i1> %Pg, <vscale x 4 x float> *%addr) {
; CHECK-LABEL: ld3.nxv12f32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-24
; CHECK-NEXT: ld3w { z0.s, z1.s, z2.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %addr, i64 -24
%base_ptr = bitcast <vscale x 4 x float>* %base to float *
%res = call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld3.sret.nxv12f32.nxv4i1.p0f32(<vscale x 4 x i1> %Pg, float *%base_ptr)
ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } %res
}
; ld3d
define { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @ld3.nxv6i64(<vscale x 2 x i1> %Pg, <vscale x 2 x i64> *%addr) {
; CHECK-LABEL: ld3.nxv6i64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #21
; CHECK-NEXT: ld3d { z0.d, z1.d, z2.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %addr, i64 21
%base_ptr = bitcast <vscale x 2 x i64>* %base to i64 *
%res = call { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld3.sret.nxv6i64.nxv2i1.p0i64(<vscale x 2 x i1> %Pg, i64 *%base_ptr)
ret { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } %res
}
define { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @ld3.nxv6f64(<vscale x 2 x i1> %Pg, <vscale x 2 x double> *%addr) {
; CHECK-LABEL: ld3.nxv6f64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-24
; CHECK-NEXT: ld3d { z0.d, z1.d, z2.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %addr, i64 -24
%base_ptr = bitcast <vscale x 2 x double>* %base to double *
%res = call { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld3.sret.nxv6f64.nxv2i1.p0f64(<vscale x 2 x i1> %Pg, double *%base_ptr)
ret { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } %res
}
; ; ld4b
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #4
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 4
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #-32
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -32
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #28
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 28
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_not_multiple_of_4_01(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_not_multiple_of_4_01:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #5
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 5
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_not_multiple_of_4_02(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_not_multiple_of_4_02:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #6
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 6
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_not_multiple_of_4_03(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_not_multiple_of_4_03:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #7
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 7
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_outside_lower_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_outside_lower_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #1
; CHECK-NEXT: mov x9, #-576
; CHECK-NEXT: lsr x8, x8, #4
; CHECK-NEXT: mul x8, x8, x9
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
; FIXME: optimize OFFSET computation so that xOFFSET = (mul (RDVL #4) #9)
; xM = -9 * 2^6
; xP = RDVL * 2^-4
; xOFFSET = RDVL * 2^-4 * -9 * 2^6 = RDVL * -36
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 -36
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
define { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @ld4.nxv64i8_outside_upper_bound(<vscale x 16 x i1> %Pg, <vscale x 16 x i8> *%addr) {
; CHECK-LABEL: ld4.nxv64i8_outside_upper_bound:
; CHECK: // %bb.0:
; CHECK-NEXT: rdvl x8, #1
; CHECK-NEXT: mov w9, #512
; CHECK-NEXT: lsr x8, x8, #4
; CHECK-NEXT: mul x8, x8, x9
; CHECK-NEXT: ld4b { z0.b, z1.b, z2.b, z3.b }, p0/z, [x0, x8]
; CHECK-NEXT: ret
; FIXME: optimize OFFSET computation so that xOFFSET = (mul (RDVL #16) #2)
; xM = 2^9
; xP = RDVL * 2^-4
; xOFFSET = RDVL * 2^-4 * 2^9 = RDVL * 32
%base = getelementptr <vscale x 16 x i8>, <vscale x 16 x i8>* %addr, i64 32
%base_ptr = bitcast <vscale x 16 x i8>* %base to i8 *
%res = call { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1> %Pg, i8 *%base_ptr)
ret { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } %res
}
; ld4h
define { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @ld4.nxv32i16(<vscale x 8 x i1> %Pg, <vscale x 8 x i16> *%addr) {
; CHECK-LABEL: ld4.nxv32i16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #8
; CHECK-NEXT: ld4h { z0.h, z1.h, z2.h, z3.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x i16>, <vscale x 8 x i16>* %addr, i64 8
%base_ptr = bitcast <vscale x 8 x i16>* %base to i16 *
%res = call { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld4.sret.nxv32i16.nxv8i1.p0i16(<vscale x 8 x i1> %Pg, i16 *%base_ptr)
ret { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } %res
}
define { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @ld4.nxv32f16(<vscale x 8 x i1> %Pg, <vscale x 8 x half> *%addr) {
; CHECK-LABEL: ld4.nxv32f16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #28
; CHECK-NEXT: ld4h { z0.h, z1.h, z2.h, z3.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x half>, <vscale x 8 x half>* %addr, i64 28
%base_ptr = bitcast <vscale x 8 x half>* %base to half *
%res = call { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld4.sret.nxv32f16.nxv8i1.p0f16(<vscale x 8 x i1> %Pg, half *%base_ptr)
ret { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } %res
}
define { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @ld4.nxv32bf16(<vscale x 8 x i1> %Pg, <vscale x 8 x bfloat> *%addr) #0 {
; CHECK-LABEL: ld4.nxv32bf16:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-32
; CHECK-NEXT: ld4h { z0.h, z1.h, z2.h, z3.h }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %addr, i64 -32
%base_ptr = bitcast <vscale x 8 x bfloat>* %base to bfloat *
%res = call { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld4.sret.nxv32bf16.nxv8i1.p0bf16(<vscale x 8 x i1> %Pg, bfloat *%base_ptr)
ret { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } %res
}
; ld4w
define { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @ld4.nxv16i32(<vscale x 4 x i1> %Pg, <vscale x 4 x i32> *%addr) {
; CHECK-LABEL: ld4.nxv16i32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #28
; CHECK-NEXT: ld4w { z0.s, z1.s, z2.s, z3.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x i32>, <vscale x 4 x i32>* %addr, i64 28
%base_ptr = bitcast <vscale x 4 x i32>* %base to i32 *
%res = call { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld4.sret.nxv16i32.nxv4i1.p0i32(<vscale x 4 x i1> %Pg, i32 *%base_ptr)
ret { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } %res
}
define { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @ld4.nxv16f32(<vscale x 4 x i1> %Pg, <vscale x 4 x float>* %addr) {
; CHECK-LABEL: ld4.nxv16f32:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-32
; CHECK-NEXT: ld4w { z0.s, z1.s, z2.s, z3.s }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 4 x float>, <vscale x 4 x float>* %addr, i64 -32
%base_ptr = bitcast <vscale x 4 x float>* %base to float *
%res = call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld4.sret.nxv16f32.nxv4i1.p0f32(<vscale x 4 x i1> %Pg, float *%base_ptr)
ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } %res
}
; ld4d
define { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @ld4.nxv8i64(<vscale x 2 x i1> %Pg, <vscale x 2 x i64> *%addr) {
; CHECK-LABEL: ld4.nxv8i64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #28
; CHECK-NEXT: ld4d { z0.d, z1.d, z2.d, z3.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x i64>, <vscale x 2 x i64>* %addr, i64 28
%base_ptr = bitcast <vscale x 2 x i64>* %base to i64 *
%res = call { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld4.sret.nxv8i64.nxv2i1.p0i64(<vscale x 2 x i1> %Pg, i64 *%base_ptr)
ret { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } %res
}
define { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @ld4.nxv8f64(<vscale x 2 x i1> %Pg, <vscale x 2 x double> *%addr) {
; CHECK-LABEL: ld4.nxv8f64:
; CHECK: // %bb.0:
; CHECK-NEXT: addvl x8, x0, #-32
; CHECK-NEXT: ld4d { z0.d, z1.d, z2.d, z3.d }, p0/z, [x8]
; CHECK-NEXT: ret
%base = getelementptr <vscale x 2 x double>, <vscale x 2 x double>* %addr, i64 -32
%base_ptr = bitcast <vscale x 2 x double>* %base to double *
%res = call { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld4.sret.nxv8f64.nxv2i1.p0f64(<vscale x 2 x i1> %Pg, double * %base_ptr)
ret { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } %res
}
declare { <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld2.sret.nxv32i8.nxv16i1.p0i8(<vscale x 16 x i1>, i8*)
declare { <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld2.sret.nxv16i16.nxv8i1.p0i16(<vscale x 8 x i1>, i16*)
declare { <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld2.sret.nxv8i32.nxv4i1.p0i32(<vscale x 4 x i1>, i32*)
declare { <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld2.sret.nxv4i64.nxv2i1.p0i64(<vscale x 2 x i1>, i64*)
declare { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld2.sret.nxv16f16.nxv8i1.p0f16(<vscale x 8 x i1>, half*)
declare { <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld2.sret.nxv16bf16.nxv8i1.p0bf16(<vscale x 8 x i1>, bfloat*)
declare { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld2.sret.nxv8f32.nxv4i1.p0f32(<vscale x 4 x i1>, float*)
declare { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld2.sret.nxv4f64.nxv2i1.p0f64(<vscale x 2 x i1>, double*)
declare { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld3.sret.nxv48i8.nxv16i1.p0i8(<vscale x 16 x i1>, i8*)
declare { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld3.sret.nxv24i16.nxv8i1.p0i16(<vscale x 8 x i1>, i16*)
declare { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld3.sret.nxv12i32.nxv4i1.p0i32(<vscale x 4 x i1>, i32*)
declare { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld3.sret.nxv6i64.nxv2i1.p0i64(<vscale x 2 x i1>, i64*)
declare { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld3.sret.nxv24f16.nxv8i1.p0f16(<vscale x 8 x i1>, half*)
declare { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld3.sret.nxv24bf16.nxv8i1.p0bf16(<vscale x 8 x i1>, bfloat*)
declare { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld3.sret.nxv12f32.nxv4i1.p0f32(<vscale x 4 x i1>, float*)
declare { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld3.sret.nxv6f64.nxv2i1.p0f64(<vscale x 2 x i1>, double*)
declare { <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8> } @llvm.aarch64.sve.ld4.sret.nxv64i8.nxv16i1.p0i8(<vscale x 16 x i1>, i8*)
declare { <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16> } @llvm.aarch64.sve.ld4.sret.nxv32i16.nxv8i1.p0i16(<vscale x 8 x i1>, i16*)
declare { <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32> } @llvm.aarch64.sve.ld4.sret.nxv16i32.nxv4i1.p0i32(<vscale x 4 x i1>, i32*)
declare { <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64> } @llvm.aarch64.sve.ld4.sret.nxv8i64.nxv2i1.p0i64(<vscale x 2 x i1>, i64*)
declare { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sve.ld4.sret.nxv32f16.nxv8i1.p0f16(<vscale x 8 x i1>, half*)
declare { <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat> } @llvm.aarch64.sve.ld4.sret.nxv32bf16.nxv8i1.p0bf16(<vscale x 8 x i1>, bfloat*)
declare { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sve.ld4.sret.nxv16f32.nxv4i1.p0f32(<vscale x 4 x i1>, float*)
declare { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sve.ld4.sret.nxv8f64.nxv2i1.p0f64(<vscale x 2 x i1>, double*)
; +bf16 is required for the bfloat version.
attributes #0 = { "target-features"="+sve,+bf16" }
|
