1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX --check-prefix=X32-AVX1
3 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX --check-prefix=X32-AVX2
4 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx512f | FileCheck %s --check-prefix=X32 --check-prefix=X32-AVX512
5 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX --check-prefix=X64-AVX1
6 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX --check-prefix=X64-AVX2
7 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512f | FileCheck %s --check-prefix=X64 --check-prefix=X64-AVX512
9 ; Combine tests involving AVX target shuffles
11 declare <4 x float> @llvm.x86.avx.vpermil.ps(<4 x float>, i8)
12 declare <8 x float> @llvm.x86.avx.vpermil.ps.256(<8 x float>, i8)
13 declare <2 x double> @llvm.x86.avx.vpermil.pd(<2 x double>, i8)
14 declare <4 x double> @llvm.x86.avx.vpermil.pd.256(<4 x double>, i8)
16 declare <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>, <4 x i32>)
17 declare <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>, <8 x i32>)
18 declare <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double>, <2 x i64>)
19 declare <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double>, <4 x i64>)
21 declare <8 x i32> @llvm.x86.avx.vperm2f128.si.256(<8 x i32>, <8 x i32>, i8)
22 declare <8 x float> @llvm.x86.avx.vperm2f128.ps.256(<8 x float>, <8 x float>, i8)
23 declare <4 x double> @llvm.x86.avx.vperm2f128.pd.256(<4 x double>, <4 x double>, i8)
25 define <4 x float> @combine_vpermilvar_4f32_identity(<4 x float> %a0) {
26 ; X32-LABEL: combine_vpermilvar_4f32_identity:
30 ; X64-LABEL: combine_vpermilvar_4f32_identity:
33 %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>)
34 %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>)
38 define <4 x float> @combine_vpermilvar_4f32_movddup(<4 x float> %a0) {
39 ; X32-LABEL: combine_vpermilvar_4f32_movddup:
41 ; X32-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
44 ; X64-LABEL: combine_vpermilvar_4f32_movddup:
46 ; X64-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
48 %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>)
51 define <4 x float> @combine_vpermilvar_4f32_movddup_load(<4 x float> *%a0) {
52 ; X32-LABEL: combine_vpermilvar_4f32_movddup_load:
54 ; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
55 ; X32-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
58 ; X64-LABEL: combine_vpermilvar_4f32_movddup_load:
60 ; X64-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
62 %1 = load <4 x float>, <4 x float> *%a0
63 %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>)
67 define <4 x float> @combine_vpermilvar_4f32_movshdup(<4 x float> %a0) {
68 ; X32-LABEL: combine_vpermilvar_4f32_movshdup:
70 ; X32-NEXT: vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
73 ; X64-LABEL: combine_vpermilvar_4f32_movshdup:
75 ; X64-NEXT: vmovshdup {{.*#+}} xmm0 = xmm0[1,1,3,3]
77 %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>)
81 define <4 x float> @combine_vpermilvar_4f32_movsldup(<4 x float> %a0) {
82 ; X32-LABEL: combine_vpermilvar_4f32_movsldup:
84 ; X32-NEXT: vmovsldup {{.*#+}} xmm0 = xmm0[0,0,2,2]
87 ; X64-LABEL: combine_vpermilvar_4f32_movsldup:
89 ; X64-NEXT: vmovsldup {{.*#+}} xmm0 = xmm0[0,0,2,2]
91 %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>)
95 define <4 x float> @combine_vpermilvar_4f32_unpckh(<4 x float> %a0) {
96 ; X32-LABEL: combine_vpermilvar_4f32_unpckh:
98 ; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
101 ; X64-LABEL: combine_vpermilvar_4f32_unpckh:
103 ; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
105 %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>)
109 define <4 x float> @combine_vpermilvar_4f32_unpckl(<4 x float> %a0) {
110 ; X32-LABEL: combine_vpermilvar_4f32_unpckl:
112 ; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
115 ; X64-LABEL: combine_vpermilvar_4f32_unpckl:
117 ; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
119 %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>)
123 define <8 x float> @combine_vpermilvar_8f32_identity(<8 x float> %a0) {
124 ; X32-LABEL: combine_vpermilvar_8f32_identity:
128 ; X64-LABEL: combine_vpermilvar_8f32_identity:
131 %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>)
132 %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>)
136 define <8 x float> @combine_vpermilvar_8f32_10326u4u(<8 x float> %a0) {
137 ; X32-LABEL: combine_vpermilvar_8f32_10326u4u:
139 ; X32-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,0,3,2,6,u,4,u]
142 ; X64-LABEL: combine_vpermilvar_8f32_10326u4u:
144 ; X64-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,0,3,2,6,u,4,u]
146 %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>)
147 %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>)
151 define <8 x float> @combine_vpermilvar_vperm2f128_8f32(<8 x float> %a0) {
152 ; X32-AVX1-LABEL: combine_vpermilvar_vperm2f128_8f32:
154 ; X32-AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
155 ; X32-AVX1-NEXT: retl
157 ; X32-AVX2-LABEL: combine_vpermilvar_vperm2f128_8f32:
159 ; X32-AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
160 ; X32-AVX2-NEXT: retl
162 ; X32-AVX512-LABEL: combine_vpermilvar_vperm2f128_8f32:
163 ; X32-AVX512: # BB#0:
164 ; X32-AVX512-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
165 ; X32-AVX512-NEXT: retl
167 ; X64-AVX1-LABEL: combine_vpermilvar_vperm2f128_8f32:
169 ; X64-AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
170 ; X64-AVX1-NEXT: retq
172 ; X64-AVX2-LABEL: combine_vpermilvar_vperm2f128_8f32:
174 ; X64-AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
175 ; X64-AVX2-NEXT: retq
177 ; X64-AVX512-LABEL: combine_vpermilvar_vperm2f128_8f32:
178 ; X64-AVX512: # BB#0:
179 ; X64-AVX512-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,3,0,1]
180 ; X64-AVX512-NEXT: retq
181 %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>)
182 %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>
183 %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>)
187 define <8 x float> @combine_vpermilvar_vperm2f128_zero_8f32(<8 x float> %a0) {
188 ; X32-LABEL: combine_vpermilvar_vperm2f128_zero_8f32:
190 ; X32-NEXT: vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
193 ; X64-LABEL: combine_vpermilvar_vperm2f128_zero_8f32:
195 ; X64-NEXT: vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
197 %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>)
198 %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>
199 %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>)
203 define <4 x double> @combine_vperm2f128_vpermilvar_as_vpblendpd(<4 x double> %a0) {
204 ; X32-LABEL: combine_vperm2f128_vpermilvar_as_vpblendpd:
206 ; X32-NEXT: vxorpd %xmm1, %xmm1, %xmm1
207 ; X32-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
210 ; X64-LABEL: combine_vperm2f128_vpermilvar_as_vpblendpd:
212 ; X64-NEXT: vxorpd %xmm1, %xmm1, %xmm1
213 ; X64-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
215 %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>)
216 %2 = shufflevector <4 x double> %1, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
217 %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>)
221 define <8 x float> @combine_vpermilvar_8f32_movddup(<8 x float> %a0) {
222 ; X32-LABEL: combine_vpermilvar_8f32_movddup:
224 ; X32-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
227 ; X64-LABEL: combine_vpermilvar_8f32_movddup:
229 ; X64-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
231 %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>)
234 define <8 x float> @combine_vpermilvar_8f32_movddup_load(<8 x float> *%a0) {
235 ; X32-LABEL: combine_vpermilvar_8f32_movddup_load:
237 ; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
238 ; X32-NEXT: vmovddup {{.*#+}} ymm0 = mem[0,0,2,2]
241 ; X64-LABEL: combine_vpermilvar_8f32_movddup_load:
243 ; X64-NEXT: vmovddup {{.*#+}} ymm0 = mem[0,0,2,2]
245 %1 = load <8 x float>, <8 x float> *%a0
246 %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>)
250 define <8 x float> @combine_vpermilvar_8f32_movshdup(<8 x float> %a0) {
251 ; X32-LABEL: combine_vpermilvar_8f32_movshdup:
253 ; X32-NEXT: vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
256 ; X64-LABEL: combine_vpermilvar_8f32_movshdup:
258 ; X64-NEXT: vmovshdup {{.*#+}} ymm0 = ymm0[1,1,3,3,5,5,7,7]
260 %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>)
264 define <8 x float> @combine_vpermilvar_8f32_movsldup(<8 x float> %a0) {
265 ; X32-LABEL: combine_vpermilvar_8f32_movsldup:
267 ; X32-NEXT: vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
270 ; X64-LABEL: combine_vpermilvar_8f32_movsldup:
272 ; X64-NEXT: vmovsldup {{.*#+}} ymm0 = ymm0[0,0,2,2,4,4,6,6]
274 %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>)
278 define <2 x double> @combine_vpermilvar_2f64_identity(<2 x double> %a0) {
279 ; X32-LABEL: combine_vpermilvar_2f64_identity:
283 ; X64-LABEL: combine_vpermilvar_2f64_identity:
286 %1 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %a0, <2 x i64> <i64 2, i64 0>)
287 %2 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %1, <2 x i64> <i64 2, i64 0>)
291 define <2 x double> @combine_vpermilvar_2f64_movddup(<2 x double> %a0) {
292 ; X32-LABEL: combine_vpermilvar_2f64_movddup:
294 ; X32-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
297 ; X64-LABEL: combine_vpermilvar_2f64_movddup:
299 ; X64-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
301 %1 = tail call <2 x double> @llvm.x86.avx.vpermilvar.pd(<2 x double> %a0, <2 x i64> <i64 0, i64 0>)
305 define <4 x double> @combine_vpermilvar_4f64_identity(<4 x double> %a0) {
306 ; X32-LABEL: combine_vpermilvar_4f64_identity:
310 ; X64-LABEL: combine_vpermilvar_4f64_identity:
313 %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>)
314 %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>)
318 define <4 x double> @combine_vpermilvar_4f64_movddup(<4 x double> %a0) {
319 ; X32-LABEL: combine_vpermilvar_4f64_movddup:
321 ; X32-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
324 ; X64-LABEL: combine_vpermilvar_4f64_movddup:
326 ; X64-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
328 %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>)
332 define <4 x float> @combine_vpermilvar_4f32_4stage(<4 x float> %a0) {
333 ; X32-LABEL: combine_vpermilvar_4f32_4stage:
335 ; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,1]
338 ; X64-LABEL: combine_vpermilvar_4f32_4stage:
340 ; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,1]
342 %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>)
343 %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>)
344 %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>)
345 %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>)
349 define <8 x float> @combine_vpermilvar_8f32_4stage(<8 x float> %a0) {
350 ; X32-LABEL: combine_vpermilvar_8f32_4stage:
352 ; X32-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
355 ; X64-LABEL: combine_vpermilvar_8f32_4stage:
357 ; X64-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
359 %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>)
360 %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>)
361 %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>)
362 %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>)
366 define <4 x float> @combine_vpermilvar_4f32_as_insertps(<4 x float> %a0) {
367 ; X32-LABEL: combine_vpermilvar_4f32_as_insertps:
369 ; X32-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[1],zero,xmm0[2],zero
372 ; X64-LABEL: combine_vpermilvar_4f32_as_insertps:
374 ; X64-NEXT: vinsertps {{.*#+}} xmm0 = xmm0[1],zero,xmm0[2],zero
376 %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>)
377 %2 = shufflevector <4 x float> %1, <4 x float> zeroinitializer, <4 x i32> <i32 2, i32 4, i32 1, i32 4>
381 define <2 x double> @constant_fold_vpermilvar_pd() {
382 ; X32-LABEL: constant_fold_vpermilvar_pd:
384 ; X32-NEXT: vmovaps {{.*#+}} xmm0 = [2.000000e+00,1.000000e+00]
387 ; X64-LABEL: constant_fold_vpermilvar_pd:
389 ; X64-NEXT: vmovaps {{.*#+}} xmm0 = [2.000000e+00,1.000000e+00]
391 %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>)
395 define <4 x double> @constant_fold_vpermilvar_pd_256() {
396 ; X32-LABEL: constant_fold_vpermilvar_pd_256:
398 ; X32-NEXT: vmovaps {{.*#+}} ymm0 = [2.000000e+00,1.000000e+00,3.000000e+00,4.000000e+00]
401 ; X64-LABEL: constant_fold_vpermilvar_pd_256:
403 ; X64-NEXT: vmovaps {{.*#+}} ymm0 = [2.000000e+00,1.000000e+00,3.000000e+00,4.000000e+00]
405 %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>)
409 define <4 x float> @constant_fold_vpermilvar_ps() {
410 ; X32-LABEL: constant_fold_vpermilvar_ps:
412 ; X32-NEXT: vmovaps {{.*#+}} xmm0 = [4.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00]
415 ; X64-LABEL: constant_fold_vpermilvar_ps:
417 ; X64-NEXT: vmovaps {{.*#+}} xmm0 = [4.000000e+00,1.000000e+00,3.000000e+00,2.000000e+00]
419 %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>)
423 define <8 x float> @constant_fold_vpermilvar_ps_256() {
424 ; X32-LABEL: constant_fold_vpermilvar_ps_256:
426 ; 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]
429 ; X64-LABEL: constant_fold_vpermilvar_ps_256:
431 ; 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]
433 %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>)