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
| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx,+xop | FileCheck %s --check-prefixes=CHECK,X86,AVX,X86-AVX
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx2,+xop | FileCheck %s --check-prefixes=CHECK,X86,AVX2,X86-AVX2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+xop | FileCheck %s --check-prefixes=CHECK,X64,AVX,X64-AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2,+xop | FileCheck %s --check-prefixes=CHECK,X64,AVX2,X64-AVX2
declare <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double>, <2 x double>, <2 x i64>, i8) nounwind readnone
declare <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double>, <4 x double>, <4 x i64>, i8) nounwind readnone
declare <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float>, <4 x float>, <4 x i32>, i8) nounwind readnone
declare <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float>, <8 x float>, <8 x i32>, i8) nounwind readnone
declare <16 x i8> @llvm.x86.xop.vpperm(<16 x i8>, <16 x i8>, <16 x i8>) nounwind readnone
define <2 x double> @combine_vpermil2pd_identity(<2 x double> %a0, <2 x double> %a1) {
; CHECK-LABEL: combine_vpermil2pd_identity:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double> %a1, <2 x double> %a0, <2 x i64> <i64 2, i64 0>, i8 0)
%res1 = call <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double> %res0, <2 x double> undef, <2 x i64> <i64 2, i64 0>, i8 0)
ret <2 x double> %res1
}
define <4 x double> @combine_vpermil2pd256_identity(<4 x double> %a0, <4 x double> %a1) {
; CHECK-LABEL: combine_vpermil2pd256_identity:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %ymm1, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> %a1, <4 x double> %a0, <4 x i64> <i64 2, i64 0, i64 2, i64 0>, i8 0)
%res1 = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> %res0, <4 x double> undef, <4 x i64> <i64 2, i64 0, i64 2, i64 0>, i8 0)
ret <4 x double> %res1
}
define <4 x double> @combine_vpermil2pd256_0z73(<4 x double> %a0, <4 x double> %a1) {
; CHECK-LABEL: combine_vpermil2pd256_0z73:
; CHECK: # %bb.0:
; CHECK-NEXT: vpermil2pd {{.*#+}} ymm0 = ymm0[0],zero,ymm1[3],ymm0[3]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = shufflevector <4 x double> %a0, <4 x double> %a1, <4 x i32> <i32 0, i32 undef, i32 7, i32 3>
%res1 = shufflevector <4 x double> %res0, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 7, i32 2, i32 3>
ret <4 x double> %res1
}
define <4 x float> @combine_vpermil2ps_identity(<4 x float> %a0, <4 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps_identity:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %a1, <4 x float> %a0, <4 x i32> <i32 3, i32 2, i32 1, i32 0>, i8 0)
%res1 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %res0, <4 x float> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>, i8 0)
ret <4 x float> %res1
}
define <4 x float> @combine_vpermil2ps_1z74(<4 x float> %a0, <4 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps_1z74:
; CHECK: # %bb.0:
; CHECK-NEXT: vshufps {{.*#+}} xmm0 = xmm0[1,1],xmm1[3,0]
; CHECK-NEXT: vxorps %xmm1, %xmm1, %xmm1
; CHECK-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 1, i32 1, i32 7, i32 4>, i8 0)
%res1 = shufflevector <4 x float> %res0, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 7, i32 2, i32 3>
ret <4 x float> %res1
}
define <4 x float> @combine_vpermil2ps_02zu(<4 x float> %a0, <4 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps_02zu:
; CHECK: # %bb.0:
; CHECK-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[0,2],zero,zero
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %a0, <4 x float> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 undef>, i8 0)
ret <4 x float> %res0
}
define <8 x float> @combine_vpermil2ps256_identity(<8 x float> %a0, <8 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps256_identity:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %ymm1, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float> %a1, <8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 1, i32 0, i32 3, i32 2>, i8 0)
%res1 = call <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float> %res0, <8 x float> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 1, i32 0, i32 3, i32 2>, i8 0)
ret <8 x float> %res1
}
define <8 x float> @combine_vpermil2ps256_08z945Az(<8 x float> %a0, <8 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps256_08z945Az:
; CHECK: # %bb.0:
; CHECK-NEXT: vpermil2ps {{.*#+}} ymm0 = ymm0[0],ymm1[0],zero,ymm1[1],ymm0[4,5],ymm1[6],zero
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float> %a0, <8 x float> %a1, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 0, i32 1, i32 6, i32 7>, i8 0)
%res1 = shufflevector <8 x float> %res0, <8 x float> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 8, i32 3, i32 4, i32 5, i32 6, i32 8>
ret <8 x float> %res1
}
define <8 x float> @combine_vpermil2ps256_zero(<8 x float> %a0, <8 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps256_zero:
; CHECK: # %bb.0:
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float> %a1, <8 x float> %a0, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 8, i32 9, i32 10, i32 11>, i8 2)
ret <8 x float> %res0
}
define <4 x float> @combine_vpermil2ps_blend_with_zero(<4 x float> %a0, <4 x float> %a1) {
; CHECK-LABEL: combine_vpermil2ps_blend_with_zero:
; CHECK: # %bb.0:
; CHECK-NEXT: vxorps %xmm1, %xmm1, %xmm1
; CHECK-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 8, i32 1, i32 2, i32 3>, i8 2)
ret <4 x float> %res0
}
define <2 x double> @combine_vpermil2pd_as_shufpd(<2 x double> %a0, <2 x double> %a1) {
; CHECK-LABEL: combine_vpermil2pd_as_shufpd:
; CHECK: # %bb.0:
; CHECK-NEXT: vshufpd {{.*#+}} xmm0 = xmm0[1],xmm1[0]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double> %a0, <2 x double> %a1, <2 x i64> <i64 2, i64 4>, i8 0)
ret <2 x double> %res0
}
define <4 x double> @combine_vpermil2pd256_as_shufpd(<4 x double> %a0, <4 x double> %a1) {
; CHECK-LABEL: combine_vpermil2pd256_as_shufpd:
; CHECK: # %bb.0:
; CHECK-NEXT: vshufpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[3],ymm1[3]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> %a0, <4 x double> %a1, <4 x i64> <i64 0, i64 4, i64 2, i64 7>, i8 0)
ret <4 x double> %res0
}
define <4 x double> @demandedelts_vpermil2pd256_as_shufpd(<4 x double> %a0, <4 x double> %a1, i64 %a2) {
; X86-AVX-LABEL: demandedelts_vpermil2pd256_as_shufpd:
; X86-AVX: # %bb.0:
; X86-AVX-NEXT: movl $4, %eax
; X86-AVX-NEXT: vmovd %eax, %xmm2
; X86-AVX-NEXT: vmovq {{.*#+}} xmm3 = mem[0],zero
; X86-AVX-NEXT: vpunpcklqdq {{.*#+}} xmm2 = xmm3[0],xmm2[0]
; X86-AVX-NEXT: vinsertf128 $1, {{\.LCPI.*}}, %ymm2, %ymm2
; X86-AVX-NEXT: vpermil2pd $0, %ymm2, %ymm1, %ymm0, %ymm0
; X86-AVX-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,2,3]
; X86-AVX-NEXT: retl
;
; X86-AVX2-LABEL: demandedelts_vpermil2pd256_as_shufpd:
; X86-AVX2: # %bb.0:
; X86-AVX2-NEXT: movl $4, %eax
; X86-AVX2-NEXT: vmovd %eax, %xmm2
; X86-AVX2-NEXT: vmovq {{.*#+}} xmm3 = mem[0],zero
; X86-AVX2-NEXT: vpunpcklqdq {{.*#+}} xmm2 = xmm3[0],xmm2[0]
; X86-AVX2-NEXT: vinserti128 $1, {{\.LCPI.*}}, %ymm2, %ymm2
; X86-AVX2-NEXT: vpermil2pd $0, %ymm2, %ymm1, %ymm0, %ymm0
; X86-AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,2,3]
; X86-AVX2-NEXT: retl
;
; X64-LABEL: demandedelts_vpermil2pd256_as_shufpd:
; X64: # %bb.0:
; X64-NEXT: vpermil2pd {{.*#+}} ymm0 = ymm1[0,0],ymm0[3],ymm1[3]
; X64-NEXT: retq
%res0 = insertelement <4 x i64> <i64 0, i64 4, i64 2, i64 7>, i64 %a2, i32 0
%res1 = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> %a0, <4 x double> %a1, <4 x i64> %res0, i8 0)
%res2 = shufflevector <4 x double> %res1, <4 x double> undef, <4 x i32> <i32 1, i32 1, i32 2, i32 3>
ret <4 x double> %res2
}
define <16 x i8> @combine_vpperm_identity(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_identity:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 31, i8 30, i8 29, i8 28, i8 27, i8 26, i8 25, i8 24, i8 23, i8 22, i8 21, i8 20, i8 19, i8 18, i8 17, i8 16>)
%res1 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res0, <16 x i8> undef, <16 x i8> <i8 15, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
ret <16 x i8> %res1
}
define <16 x i8> @combine_vpperm_zero(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_zero:
; CHECK: # %bb.0:
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 128, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>)
%res1 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res0, <16 x i8> undef, <16 x i8> <i8 0, i8 128, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>)
%res2 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res1, <16 x i8> undef, <16 x i8> <i8 0, i8 1, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128>)
ret <16 x i8> %res2
}
define <16 x i8> @combine_vpperm_identity_bitcast(<16 x i8> %a0, <16 x i8> %a1) {
; X86-LABEL: combine_vpperm_identity_bitcast:
; X86: # %bb.0:
; X86-NEXT: vpaddq {{\.LCPI.*}}, %xmm0, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: combine_vpperm_identity_bitcast:
; X64: # %bb.0:
; X64-NEXT: vpaddq {{.*}}(%rip), %xmm0, %xmm0
; X64-NEXT: retq
%mask = bitcast <2 x i64> <i64 1084818905618843912, i64 506097522914230528> to <16 x i8>
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> %mask)
%res1 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res0, <16 x i8> undef, <16 x i8> %mask)
%res2 = bitcast <16 x i8> %res1 to <2 x i64>
%res3 = add <2 x i64> %res2, <i64 1084818905618843912, i64 506097522914230528>
%res4 = bitcast <2 x i64> %res3 to <16 x i8>
ret <16 x i8> %res4
}
define <16 x i8> @combine_vpperm_as_blend_with_zero(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_as_blend_with_zero:
; CHECK: # %bb.0:
; CHECK-NEXT: vpxor %xmm1, %xmm1, %xmm1
; CHECK-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3],xmm1[4,5,6,7]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 0, i8 1, i8 128, i8 129, i8 4, i8 5, i8 6, i8 7, i8 130, i8 131, i8 132, i8 133, i8 134, i8 135, i8 136, i8 137>)
ret <16 x i8> %res0
}
define <16 x i8> @combine_vpperm_as_unary_unpckhbw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_as_unary_unpckhbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpunpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a0, <16 x i8> <i8 8, i8 undef, i8 9, i8 25, i8 10, i8 26, i8 11, i8 27, i8 12, i8 28, i8 13, i8 29, i8 14, i8 30, i8 15, i8 31>)
ret <16 x i8> %res0
}
define <16 x i8> @combine_vpperm_as_unpckhbw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_as_unpckhbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 8, i8 24, i8 9, i8 25, i8 10, i8 26, i8 11, i8 27, i8 12, i8 28, i8 13, i8 29, i8 14, i8 30, i8 15, i8 31>)
ret <16 x i8> %res0
}
define <16 x i8> @combine_vpperm_as_unpcklbw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: combine_vpperm_as_unpcklbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 16, i8 0, i8 17, i8 1, i8 18, i8 2, i8 19, i8 3, i8 20, i8 4, i8 21, i8 5, i8 22, i8 6, i8 23, i8 7>)
ret <16 x i8> %res0
}
define <4 x i32> @combine_vpperm_10zz32BA(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: combine_vpperm_10zz32BA:
; CHECK: # %bb.0:
; CHECK-NEXT: vpperm {{.*#+}} xmm0 = xmm0[2,3,0,1],zero,zero,zero,zero,xmm0[6,7,4,5],xmm1[6,7,4,5]
; CHECK-NEXT: ret{{[l|q]}}
%res0 = shufflevector <4 x i32> %a0, <4 x i32> %a1, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
%res1 = bitcast <4 x i32> %res0 to <16 x i8>
%res2 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %res1, <16 x i8> undef, <16 x i8> <i8 2, i8 3, i8 0, i8 1, i8 128, i8 128, i8 128, i8 128, i8 10, i8 11, i8 8, i8 9, i8 14, i8 15, i8 12, i8 13>)
%res3 = bitcast <16 x i8> %res2 to <4 x i32>
ret <4 x i32> %res3
}
define void @buildvector_v4f32_0404(float %a, float %b, <4 x float>* %ptr) {
; X86-LABEL: buildvector_v4f32_0404:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
; X86-NEXT: vmovaps %xmm0, (%eax)
; X86-NEXT: retl
;
; X64-AVX-LABEL: buildvector_v4f32_0404:
; X64-AVX: # %bb.0:
; X64-AVX-NEXT: vpermil2ps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[0],xmm1[0]
; X64-AVX-NEXT: vmovaps %xmm0, (%rdi)
; X64-AVX-NEXT: retq
;
; X64-AVX2-LABEL: buildvector_v4f32_0404:
; X64-AVX2: # %bb.0:
; X64-AVX2-NEXT: vunpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; X64-AVX2-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; X64-AVX2-NEXT: vmovaps %xmm0, (%rdi)
; X64-AVX2-NEXT: retq
%v0 = insertelement <4 x float> undef, float %a, i32 0
%v1 = insertelement <4 x float> %v0, float %b, i32 1
%v2 = insertelement <4 x float> %v1, float %a, i32 2
%v3 = insertelement <4 x float> %v2, float %b, i32 3
store <4 x float> %v3, <4 x float>* %ptr
ret void
}
define void @buildvector_v4f32_07z6(float %a, <4 x float> %b, <4 x float>* %ptr) {
; X86-LABEL: buildvector_v4f32_07z6:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
; X86-NEXT: vpermil2ps {{.*#+}} xmm0 = xmm1[0],xmm0[3],zero,xmm0[2]
; X86-NEXT: vmovaps %xmm0, (%eax)
; X86-NEXT: retl
;
; X64-LABEL: buildvector_v4f32_07z6:
; X64: # %bb.0:
; X64-NEXT: vpermil2ps {{.*#+}} xmm0 = xmm0[0],xmm1[3],zero,xmm1[2]
; X64-NEXT: vmovaps %xmm0, (%rdi)
; X64-NEXT: retq
%b2 = extractelement <4 x float> %b, i32 2
%b3 = extractelement <4 x float> %b, i32 3
%v0 = insertelement <4 x float> undef, float %a, i32 0
%v1 = insertelement <4 x float> %v0, float %b3, i32 1
%v2 = insertelement <4 x float> %v1, float 0.0, i32 2
%v3 = insertelement <4 x float> %v2, float %b2, i32 3
store <4 x float> %v3, <4 x float>* %ptr
ret void
}
define <2 x double> @constant_fold_vpermil2pd() {
; CHECK-LABEL: constant_fold_vpermil2pd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps {{.*#+}} xmm0 = [-2.0E+0,2.0E+0]
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double> <double 1.0, double 2.0>, <2 x double> <double -2.0, double -1.0>, <2 x i64> <i64 4, i64 2>, i8 2)
ret <2 x double> %1
}
define <4 x double> @constant_fold_vpermil2pd_256() {
; CHECK-LABEL: constant_fold_vpermil2pd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps {{.*#+}} ymm0 = [-4.0E+0,0.0E+0,4.0E+0,3.0E+0]
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double> <double 1.0, double 2.0, double 3.0, double 4.0>, <4 x double> <double -4.0, double -3.0, double -2.0, double -1.0>, <4 x i64> <i64 4, i64 8, i64 2, i64 0>, i8 2)
ret <4 x double> %1
}
define <4 x float> @constant_fold_vpermil2ps() {
; CHECK-LABEL: constant_fold_vpermil2ps:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps {{.*#+}} xmm0 = [-4.0E+0,1.0E+0,3.0E+0,0.0E+0]
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float> <float 1.0, float 2.0, float 3.0, float 4.0>, <4 x float> <float -4.0, float -3.0, float -2.0, float -1.0>, <4 x i32> <i32 4, i32 0, i32 2, i32 8>, i8 2)
ret <4 x float> %1
}
define <8 x float> @constant_fold_vpermil2ps_256() {
; CHECK-LABEL: constant_fold_vpermil2ps_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps {{.*#+}} ymm0 = [-8.0E+0,1.0E+0,3.0E+0,0.0E+0,5.0E+0,0.0E+0,5.0E+0,7.0E+0]
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <8 x float> @llvm.x86.xop.vpermil2ps.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 float> <float -8.0, float -7.0, float -6.0, float -5.0, float -4.0, float -3.0, float -2.0, float -1.0>, <8 x i32> <i32 4, i32 0, i32 2, i32 8, i32 0, i32 8, i32 0, i32 2>, i8 2)
ret <8 x float> %1
}
define <16 x i8> @constant_fold_vpperm() {
; CHECK-LABEL: constant_fold_vpperm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps {{.*#+}} xmm0 = [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15]
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> <i8 0, i8 -1, i8 -2, i8 -3, i8 -4, i8 -5, i8 -6, i8 -7, i8 -8, i8 -9, i8 -10, i8 -11, i8 -12, i8 -13, i8 -14, i8 -15>, <16 x i8> <i8 15, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>, <16 x i8> <i8 31, i8 30, i8 29, i8 28, i8 27, i8 26, i8 25, i8 24, i8 23, i8 22, i8 21, i8 20, i8 19, i8 18, i8 17, i8 16>)
ret <16 x i8> %1
}
define <4 x float> @PR31296(i8* %in) {
; X86-LABEL: PR31296:
; X86: # %bb.0: # %entry
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
; X86-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; X86-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[0],zero,zero,mem[0]
; X86-NEXT: retl
;
; X64-LABEL: PR31296:
; X64: # %bb.0: # %entry
; X64-NEXT: movl (%rdi), %eax
; X64-NEXT: vmovq %rax, %xmm0
; X64-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[0],zero,zero,mem[0]
; X64-NEXT: retq
entry:
%0 = getelementptr i8, i8* %in, i32 0
%1 = bitcast i8* %0 to i32*
%2 = load i32, i32* %1
%3 = zext i32 %2 to i128
%4 = bitcast i128 %3 to <4 x float>
%5 = shufflevector <4 x float> %4, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 4, i32 4, i32 5>
ret <4 x float> %5
}
|