llvm.org GIT mirror llvm / release_60 test / CodeGen / X86 / vector-shuffle-combining-avx.ll
release_60

Tree @release_60 (Download .tar.gz)

vector-shuffle-combining-avx.ll @release_60raw · history · blame

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX --check-prefix=X32-AVX1
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX --check-prefix=X32-AVX2
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx512f | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX512
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX --check-prefix=X64-AVX1
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX --check-prefix=X64-AVX2
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512f | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX512
;
; Combine tests involving AVX target shuffles

declare <4 x float> @llvm.x86.avx.vpermil.ps(<4 x float>, i8)
declare <8 x float> @llvm.x86.avx.vpermil.ps.256(<8 x float>, i8)
declare <2 x double> @llvm.x86.avx.vpermil.pd(<2 x double>, i8)
declare <4 x double> @llvm.x86.avx.vpermil.pd.256(<4 x double>, i8)

declare <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>, <4 x i32>)
declare <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>, <8 x i32>)
declare <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double>, <2 x i64>)
declare <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double>, <4 x i64>)

declare <8 x i32> @llvm.x86.avx.vperm2f128.si.256(<8 x i32>, <8 x i32>, i8)
declare <8 x float> @llvm.x86.avx.vperm2f128.ps.256(<8 x float>, <8 x float>, i8)
declare <4 x double> @llvm.x86.avx.vperm2f128.pd.256(<4 x double>, <4 x double>, i8)

define <4 x float> @combine_vpermilvar_4f32_identity(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_identity:
; X32:       # %bb.0:
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_identity:
; X64:       # %bb.0:
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
  %2 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %1, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
  ret <4 x float> %2
}

define <4 x float> @combine_vpermilvar_4f32_movddup(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_movddup:
; X32:       # %bb.0:
; X32-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_movddup:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 0, i32 1, i32 0, i32 1>)
  ret <4 x float> %1
}
define <4 x float> @combine_vpermilvar_4f32_movddup_load(<4 x float> *%a0) {
; X32-LABEL: combine_vpermilvar_4f32_movddup_load:
; X32:       # %bb.0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_movddup_load:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
; X64-NEXT:    retq
  %1 = load <4 x float>, <4 x float> *%a0
  %2 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %1, <4 x i32> <i32 0, i32 1, i32 0, i32 1>)
  ret <4 x float> %2
}

define <4 x float> @combine_vpermilvar_4f32_movshdup(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_movshdup:
; X32:       # %bb.0:
; X32-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_movshdup:
; X64:       # %bb.0:
; X64-NEXT:    vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 undef, i32 1, i32 3, i32 3>)
  ret <4 x float> %1
}

define <4 x float> @combine_vpermilvar_4f32_movsldup(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_movsldup:
; X32:       # %bb.0:
; X32-NEXT:    vmovsldup {{.*#+}} xmm0 = xmm0[0,0,2,2]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_movsldup:
; X64:       # %bb.0:
; X64-NEXT:    vmovsldup {{.*#+}} xmm0 = xmm0[0,0,2,2]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 0, i32 0, i32 2, i32 undef>)
  ret <4 x float> %1
}

define <4 x float> @combine_vpermilvar_4f32_unpckh(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_unpckh:
; X32:       # %bb.0:
; X32-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_unpckh:
; X64:       # %bb.0:
; X64-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 2, i32 2, i32 3, i32 3>)
  ret <4 x float> %1
}

define <4 x float> @combine_vpermilvar_4f32_unpckl(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_unpckl:
; X32:       # %bb.0:
; X32-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_unpckl:
; X64:       # %bb.0:
; X64-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 0, i32 0, i32 1, i32 1>)
  ret <4 x float> %1
}

define <8 x float> @combine_vpermilvar_8f32_identity(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_identity:
; X32:       # %bb.0:
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_identity:
; X64:       # %bb.0:
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 2, i32 3, i32 0, i32 undef>)
  %2 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %1, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 2, i32 3, i32 0, i32 1>)
  ret <8 x float> %2
}

define <8 x float> @combine_vpermilvar_8f32_10326u4u(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_10326u4u:
; X32:       # %bb.0:
; X32-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[1,0,3,2,6,u,4,u]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_10326u4u:
; X64:       # %bb.0:
; X64-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[1,0,3,2,6,u,4,u]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 0, i32 1, i32 2, i32 undef>)
  %2 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %1, <8 x i32> <i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 undef>)
  ret <8 x float> %2
}

define <8 x float> @combine_vpermilvar_vperm2f128_8f32(<8 x float> %a0) {
; X32-AVX1-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X32-AVX1:       # %bb.0:
; X32-AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X32-AVX1-NEXT:    retl
;
; X32-AVX2-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X32-AVX2:       # %bb.0:
; X32-AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X32-AVX2-NEXT:    retl
;
; X32-AVX512-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X32-AVX512:       # %bb.0:
; X32-AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X32-AVX512-NEXT:    retl
;
; X64-AVX1-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X64-AVX1:       # %bb.0:
; X64-AVX1-NEXT:    vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X64-AVX1-NEXT:    retq
;
; X64-AVX2-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X64-AVX2:       # %bb.0:
; X64-AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X64-AVX2-NEXT:    retq
;
; X64-AVX512-LABEL: combine_vpermilvar_vperm2f128_8f32:
; X64-AVX512:       # %bb.0:
; X64-AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
; X64-AVX512-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  %2 = shufflevector <8 x float> %1, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
  %3 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %2, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  ret <8 x float> %3
}

define <8 x float> @combine_vpermilvar_vperm2f128_zero_8f32(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_vperm2f128_zero_8f32:
; X32:       # %bb.0:
; X32-NEXT:    vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_vperm2f128_zero_8f32:
; X64:       # %bb.0:
; X64-NEXT:    vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  %2 = shufflevector <8 x float> %1, <8 x float> zeroinitializer, <8 x i32> <i32 8, i32 8, i32 8, i32 8, i32 0, i32 1, i32 2, i32 3>
  %3 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %2, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  ret <8 x float> %3
}

define <4 x double> @combine_vperm2f128_vpermilvar_as_vpblendpd(<4 x double> %a0) {
; X32-LABEL: combine_vperm2f128_vpermilvar_as_vpblendpd:
; X32:       # %bb.0:
; X32-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
; X32-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vperm2f128_vpermilvar_as_vpblendpd:
; X64:       # %bb.0:
; X64-NEXT:    vxorpd %xmm1, %xmm1, %xmm1
; X64-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
; X64-NEXT:    retq
  %1 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> <i64 2, i64 0, i64 2, i64 0>)
  %2 = shufflevector <4 x double> %1, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
  %3 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %2, <4 x i64> <i64 2, i64 0, i64 2, i64 0>)
  ret <4 x double> %3
}

define <8 x float> @combine_vpermilvar_8f32_movddup(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_movddup:
; X32:       # %bb.0:
; X32-NEXT:    vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_movddup:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>)
  ret <8 x float> %1
}
define <8 x float> @combine_vpermilvar_8f32_movddup_load(<8 x float> *%a0) {
; X32-LABEL: combine_vpermilvar_8f32_movddup_load:
; X32:       # %bb.0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vmovddup {{.*#+}} ymm0 = mem[0,0,2,2]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_movddup_load:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} ymm0 = mem[0,0,2,2]
; X64-NEXT:    retq
  %1 = load <8 x float>, <8 x float> *%a0
  %2 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %1, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>)
  ret <8 x float> %2
}

define <8 x float> @combine_vpermilvar_8f32_movshdup(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_movshdup:
; X32:       # %bb.0:
; X32-NEXT:    vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_movshdup:
; X64:       # %bb.0:
; X64-NEXT:    vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 undef, i32 5, i32 7, i32 7>)
  ret <8 x float> %1
}

define <8 x float> @combine_vpermilvar_8f32_movsldup(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_movsldup:
; X32:       # %bb.0:
; X32-NEXT:    vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_movsldup:
; X64:       # %bb.0:
; X64-NEXT:    vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>)
  ret <8 x float> %1
}

define <2 x double> @combine_vpermilvar_2f64_identity(<2 x double> %a0) {
; X32-LABEL: combine_vpermilvar_2f64_identity:
; X32:       # %bb.0:
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_2f64_identity:
; X64:       # %bb.0:
; X64-NEXT:    retq
  %1 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %a0, <2 x i64> <i64 2, i64 0>)
  %2 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double>  %1, <2 x i64> <i64 2, i64 0>)
  ret <2 x double> %2
}

define <2 x double> @combine_vpermilvar_2f64_movddup(<2 x double> %a0) {
; X32-LABEL: combine_vpermilvar_2f64_movddup:
; X32:       # %bb.0:
; X32-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_2f64_movddup:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; X64-NEXT:    retq
  %1 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %a0, <2 x i64> <i64 0, i64 0>)
  ret <2 x double> %1
}

define <4 x double> @combine_vpermilvar_4f64_identity(<4 x double> %a0) {
; X32-LABEL: combine_vpermilvar_4f64_identity:
; X32:       # %bb.0:
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f64_identity:
; X64:       # %bb.0:
; X64-NEXT:    retq
  %1 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> <i64 2, i64 0, i64 2, i64 0>)
  %2 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double>  %1, <4 x i64> <i64 2, i64 0, i64 2, i64 0>)
  ret <4 x double> %2
}

define <4 x double> @combine_vpermilvar_4f64_movddup(<4 x double> %a0) {
; X32-LABEL: combine_vpermilvar_4f64_movddup:
; X32:       # %bb.0:
; X32-NEXT:    vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f64_movddup:
; X64:       # %bb.0:
; X64-NEXT:    vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
; X64-NEXT:    retq
  %1 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> <i64 0, i64 0, i64 4, i64 4>)
  ret <4 x double> %1
}

define <4 x float> @combine_vpermilvar_4f32_4stage(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_4stage:
; X32:       # %bb.0:
; X32-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,1]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_4stage:
; X64:       # %bb.0:
; X64-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,1]
; X64-NEXT:    retq
  %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
  %2 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %1, <4 x i32> <i32 2, i32 3, i32 0, i32 1>)
  %3 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %2, <4 x i32> <i32 0, i32 2, i32 1, i32 3>)
  %4 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %3, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
  ret <4 x float> %4
}

define <8 x float> @combine_vpermilvar_8f32_4stage(<8 x float> %a0) {
; X32-LABEL: combine_vpermilvar_8f32_4stage:
; X32:       # %bb.0:
; X32-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_8f32_4stage:
; X64:       # %bb.0:
; X64-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
; X64-NEXT:    retq
  %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  %2 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %1, <8 x i32> <i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1>)
  %3 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %2, <8 x i32> <i32 0, i32 2, i32 1, i32 3, i32 0, i32 2, i32 1, i32 3>)
  %4 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %3, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
  ret <8 x float> %4
}

define <4 x float> @combine_vpermilvar_4f32_as_insertps(<4 x float> %a0) {
; X32-LABEL: combine_vpermilvar_4f32_as_insertps:
; X32:       # %bb.0:
; X32-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[1],zero,xmm0[2],zero
; X32-NEXT:    retl
;
; X64-LABEL: combine_vpermilvar_4f32_as_insertps:
; X64:       # %bb.0:
; X64-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[1],zero,xmm0[2],zero
; X64-NEXT:    retq
  %1 = call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
  %2 = shufflevector <4 x float> %1, <4 x float> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 1, i32 4>
  ret <4 x float> %2
}

define <2 x double> @constant_fold_vpermilvar_pd() {
; X32-LABEL: constant_fold_vpermilvar_pd:
; X32:       # %bb.0:
; X32-NEXT:    vmovaps {{.*#+}} xmm0 = [2.000000e+00,1.000000e+00]
; X32-NEXT:    retl
;
; X64-LABEL: constant_fold_vpermilvar_pd:
; X64:       # %bb.0:
; X64-NEXT:    vmovaps {{.*#+}} xmm0 = [2.000000e+00,1.000000e+00]
; X64-NEXT:    retq
  %1 = call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> <double 1.0, double 2.0>, <2 x i64> <i64 2, i64 0>)
  ret <2 x double> %1
}

define <4 x double> @constant_fold_vpermilvar_pd_256() {
; X32-LABEL: constant_fold_vpermilvar_pd_256:
; X32:       # %bb.0:
; X32-NEXT:    vmovaps {{.*#+}} ymm0 = [2.000000e+00,1.000000e+00,3.000000e+00,4.000000e+00]
; X32-NEXT:    retl
;
; X64-LABEL: constant_fold_vpermilvar_pd_256:
; X64:       # %bb.0:
; X64-NEXT:    vmovaps {{.*#+}} ymm0 = [2.000000e+00,1.000000e+00,3.000000e+00,4.000000e+00]
; X64-NEXT:    retq
  %1 = call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> <double 1.0, double 2.0, double 3.0, double 4.0>, <4 x i64> <i64 2, i64 0, i64 0, i64 2>)
  ret <4 x double> %1
}

define <4 x float> @constant_fold_vpermilvar_ps() {
; X32-LABEL: constant_fold_vpermilvar_ps:
; X32:       # %bb.0:
; X32-NEXT:    vmovaps {{.*#+}} xmm0 = [4.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00]
; X32-NEXT:    retl
;
; X64-LABEL: constant_fold_vpermilvar_ps:
; X64:       # %bb.0:
; X64-NEXT:    vmovaps {{.*#+}} xmm0 = [4.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00]
; X64-NEXT:    retq
  %1 = call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> <float 1.0, float 2.0, float 3.0, float 4.0>, <4 x i32> <i32 3, i32 0, i32 2, i32 1>)
  ret <4 x float> %1
}

define <8 x float> @constant_fold_vpermilvar_ps_256() {
; X32-LABEL: constant_fold_vpermilvar_ps_256:
; X32:       # %bb.0:
; X32-NEXT:    vmovaps {{.*#+}} ymm0 = [1.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00,5.000000e+00,6.000000e+00,6.000000e+00,6.000000e+00]
; X32-NEXT:    retl
;
; X64-LABEL: constant_fold_vpermilvar_ps_256:
; X64:       # %bb.0:
; X64-NEXT:    vmovaps {{.*#+}} ymm0 = [1.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00,5.000000e+00,6.000000e+00,6.000000e+00,6.000000e+00]
; X64-NEXT:    retq
  %1 = call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> <float 1.0, float 2.0, float 3.0, float 4.0, float 5.0, float 6.0, float 7.0, float 8.0>, <8 x i32> <i32 4, i32 0, i32 2, i32 1, i32 0, i32 1, i32 1, i32 1>)
  ret <8 x float> %1
}