159 changes: 61 additions & 98 deletions llvm/test/CodeGen/X86/avx512-shuffles/partial_permute.ll
View file

Large diffs are not rendered by default.

10 changes: 4 additions & 6 deletions llvm/test/CodeGen/X86/pr29112.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,16 +11,14 @@ define <4 x float> @bar(<4 x float>* %a1p, <4 x float>* %a2p, <4 x float> %a3, <
; CHECK-NEXT: subq $72, %rsp
; CHECK-NEXT: .cfi_def_cfa_offset 80
; CHECK-NEXT: vmovaps %xmm1, %xmm9
; CHECK-NEXT: vbroadcasti32x4 {{.*#+}} zmm14 = [4,22,1,17,4,22,1,17,4,22,1,17,4,22,1,17]
; CHECK-NEXT: # zmm14 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
; CHECK-NEXT: vmovaps {{.*#+}} xmm14 = [4,22,1,17]
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm14
; CHECK-NEXT: vbroadcasti32x4 {{.*#+}} zmm10 = [4,30,1,22,4,30,1,22,4,30,1,22,4,30,1,22]
; CHECK-NEXT: # zmm10 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
; CHECK-NEXT: vmovaps {{.*#+}} xmm10 = [4,30,1,22]
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm10
; CHECK-NEXT: vbroadcastsd {{.*#+}} zmm7 = [85899345925,85899345925,85899345925,85899345925,85899345925,85899345925,85899345925,85899345925]
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm7
; CHECK-NEXT: vmovaps {{.*#+}} xmm8 = [4,28,1,29]
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm8
; CHECK-NEXT: vmovaps {{.*#+}} xmm7 = <5,20,u,u>
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm7
; CHECK-NEXT: vmovaps {{.*#+}} xmm4 = [4,21,1,7]
; CHECK-NEXT: vpermi2ps %zmm3, %zmm2, %zmm4
; CHECK-NEXT: vextractf128 $1, %ymm3, %xmm5
Expand Down
2 changes: 1 addition & 1 deletion llvm/test/CodeGen/X86/pr46532.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@ define void @WhileWithLoopInvariantOperation.21() {
; CHECK-NEXT: movq (%rax), %rax
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT: vmovaps %xmm0, 32(%rax)
; CHECK-NEXT: vmovaps {{.*#+}} ymm0 = [4294967295,4294967295,0,0,0,0,0,0]
; CHECK-NEXT: vmovaps {{.*#+}} xmm0 = [4294967295,4294967295,0,0]
; CHECK-NEXT: vmaskmovps %ymm0, %ymm0, (%rax)
while.1.body.preheader:
%0 = load i8*, i8** undef, align 8, !invariant.load !0, !dereferenceable !1, !align !2
Expand Down
12 changes: 6 additions & 6 deletions llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -465,9 +465,9 @@ define <4 x double> @PR34175(<32 x i16>* %p) {
;
; AVX512BWVL-LABEL: PR34175:
; AVX512BWVL: # %bb.0:
; AVX512BWVL-NEXT: vmovdqu (%rdi), %ymm0
; AVX512BWVL-NEXT: vpbroadcastq {{.*#+}} ymm1 = [6755468161056768,6755468161056768,6755468161056768,6755468161056768]
; AVX512BWVL-NEXT: vpermi2w 32(%rdi), %ymm0, %ymm1
; AVX512BWVL-NEXT: vmovdqa {{.*#+}} xmm0 = <0,8,16,24,u,u,u,u>
; AVX512BWVL-NEXT: vmovdqu (%rdi), %ymm1
; AVX512BWVL-NEXT: vpermt2w 32(%rdi), %ymm0, %ymm1
; AVX512BWVL-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX512BWVL-NEXT: vcvtdq2pd %xmm0, %ymm0
; AVX512BWVL-NEXT: retq
Expand All @@ -484,9 +484,9 @@ define <4 x double> @PR34175(<32 x i16>* %p) {
;
; AVX512VBMIVL-LABEL: PR34175:
; AVX512VBMIVL: # %bb.0:
; AVX512VBMIVL-NEXT: vmovdqu (%rdi), %ymm0
; AVX512VBMIVL-NEXT: vpbroadcastq {{.*#+}} ymm1 = [6755468161056768,6755468161056768,6755468161056768,6755468161056768]
; AVX512VBMIVL-NEXT: vpermi2w 32(%rdi), %ymm0, %ymm1
; AVX512VBMIVL-NEXT: vmovdqa {{.*#+}} xmm0 = <0,8,16,24,u,u,u,u>
; AVX512VBMIVL-NEXT: vmovdqu (%rdi), %ymm1
; AVX512VBMIVL-NEXT: vpermt2w 32(%rdi), %ymm0, %ymm1
; AVX512VBMIVL-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX512VBMIVL-NEXT: vcvtdq2pd %xmm0, %ymm0
; AVX512VBMIVL-NEXT: retq
Expand Down
4 changes: 2 additions & 2 deletions llvm/test/CodeGen/X86/trunc-subvector.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ define <2 x i32> @test5(<8 x i32> %v) {
;
; AVX2-LABEL: test5:
; AVX2: # %bb.0:
; AVX2-NEXT: vbroadcastsd {{.*#+}} ymm1 = [17179869187,17179869187,17179869187,17179869187]
; AVX2-NEXT: vmovaps {{.*#+}} xmm1 = [3,4,4,4]
; AVX2-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX2-NEXT: vzeroupper
Expand Down Expand Up @@ -177,7 +177,7 @@ define <2 x i32> @test10(<8 x i32> %v) {
;
; AVX2-LABEL: test10:
; AVX2: # %bb.0:
; AVX2-NEXT: vbroadcastsd {{.*#+}} ymm1 = [17179869187,17179869187,17179869187,17179869187]
; AVX2-NEXT: vmovaps {{.*#+}} xmm1 = [3,4,4,4]
; AVX2-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX2-NEXT: vzeroupper
Expand Down
14 changes: 7 additions & 7 deletions llvm/test/CodeGen/X86/vector-shuffle-256-v16.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -354,7 +354,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,8]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -396,7 +396,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,9,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,9,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -437,7 +437,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,10,0,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,10,0,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -478,7 +478,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,11,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,11,0,0,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -519,7 +519,7 @@ define <16 x i16> @shuffle_v16i16_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,12,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,12,0,0,0,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -560,7 +560,7 @@ define <16 x i16> @shuffle_v16i16_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,13,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,13,0,0,0,0,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down Expand Up @@ -601,7 +601,7 @@ define <16 x i16> @shuffle_v16i16_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_0
;
; AVX512VL-LABEL: shuffle_v16i16_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm1 = [0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm1 = [0,14,0,0,0,0,0,0]
; AVX512VL-NEXT: vpermw %ymm0, %ymm1, %ymm0
; AVX512VL-NEXT: retq
;
Expand Down
30 changes: 15 additions & 15 deletions llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -812,7 +812,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_16_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,16]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -859,7 +859,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_17_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,17,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,17,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -906,7 +906,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_18_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,18,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,18,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -953,7 +953,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_19_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,19,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,0,0,19,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1000,7 +1000,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_20_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,20,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,0,20,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1047,7 +1047,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_21_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,21,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,0,21,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1094,7 +1094,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_22_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,22,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,0,22,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1141,7 +1141,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_23_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,23,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,23,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1188,7 +1188,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_24_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,24,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,24,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1235,7 +1235,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_00_25_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,25,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,25,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1282,7 +1282,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_00_26_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,26,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,26,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1329,7 +1329,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_00_27_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,27,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,27,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1376,7 +1376,7 @@ define <32 x i8> @shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_00_28_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,28,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,28,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1423,7 +1423,7 @@ define <32 x i8> @shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_00_29_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,29,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,29,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down Expand Up @@ -1470,7 +1470,7 @@ define <32 x i8> @shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_
;
; AVX512VLVBMI-LABEL: shuffle_v32i8_00_30_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
; AVX512VLVBMI: # %bb.0:
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} ymm1 = [0,30,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vmovdqa {{.*#+}} xmm1 = [0,30,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512VLVBMI-NEXT: vpermb %ymm0, %ymm1, %ymm0
; AVX512VLVBMI-NEXT: retq
;
Expand Down
24 changes: 10 additions & 14 deletions llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -136,7 +136,7 @@ define <8 x float> @shuffle_v8f32_00040000(<8 x float> %a, <8 x float> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8f32_00040000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,4,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,0,4]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -153,7 +153,7 @@ define <8 x float> @shuffle_v8f32_00500000(<8 x float> %a, <8 x float> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8f32_00500000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,5,0,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,5,0]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -170,7 +170,7 @@ define <8 x float> @shuffle_v8f32_06000000(<8 x float> %a, <8 x float> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8f32_06000000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,6,0,0,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,6,0,0]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand Down Expand Up @@ -1470,7 +1470,7 @@ define <8 x i32> @shuffle_v8i32_00040000(<8 x i32> %a, <8 x i32> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8i32_00040000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,4,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,0,4]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -1487,7 +1487,7 @@ define <8 x i32> @shuffle_v8i32_00500000(<8 x i32> %a, <8 x i32> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8i32_00500000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,5,0,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,5,0]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -1504,7 +1504,7 @@ define <8 x i32> @shuffle_v8i32_06000000(<8 x i32> %a, <8 x i32> %b) {
;
; AVX2OR512VL-LABEL: shuffle_v8i32_06000000:
; AVX2OR512VL: # %bb.0:
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} ymm1 = [0,6,0,0,0,0,0,0]
; AVX2OR512VL-NEXT: vmovaps {{.*#+}} xmm1 = [0,6,0,0]
; AVX2OR512VL-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2OR512VL-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand Down Expand Up @@ -3280,15 +3280,13 @@ define <8 x i32> @lowhalf_v8i32(<8 x i32> %x, <8 x i32> %y) {
; AVX2-LABEL: lowhalf_v8i32:
; AVX2: # %bb.0:
; AVX2-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
; AVX2-NEXT: vbroadcastf128 {{.*#+}} ymm1 = [2,6,3,6,2,6,3,6]
; AVX2-NEXT: # ymm1 = mem[0,1,0,1]
; AVX2-NEXT: vmovaps {{.*#+}} xmm1 = [2,6,3,6]
; AVX2-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2-NEXT: retq
;
; AVX512VL-LABEL: lowhalf_v8i32:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm2 = [2,14,3,14,2,14,3,14]
; AVX512VL-NEXT: # ymm2 = mem[0,1,0,1]
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm2 = [2,14,3,14]
; AVX512VL-NEXT: vpermt2d %ymm1, %ymm2, %ymm0
; AVX512VL-NEXT: retq
%r = shufflevector <8 x i32> %x, <8 x i32> %y, <8 x i32> <i32 2, i32 14, i32 3, i32 14, i32 undef, i32 undef, i32 undef, i32 undef>
Expand All @@ -3308,15 +3306,13 @@ define <8 x float> @lowhalf_v8f32(<8 x float> %x, <8 x float> %y) {
; AVX2-LABEL: lowhalf_v8f32:
; AVX2: # %bb.0:
; AVX2-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
; AVX2-NEXT: vbroadcastf128 {{.*#+}} ymm1 = [2,6,3,6,2,6,3,6]
; AVX2-NEXT: # ymm1 = mem[0,1,0,1]
; AVX2-NEXT: vmovaps {{.*#+}} xmm1 = [2,6,3,6]
; AVX2-NEXT: vpermps %ymm0, %ymm1, %ymm0
; AVX2-NEXT: retq
;
; AVX512VL-LABEL: lowhalf_v8f32:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm2 = [2,14,3,14,2,14,3,14]
; AVX512VL-NEXT: # ymm2 = mem[0,1,0,1]
; AVX512VL-NEXT: vmovaps {{.*#+}} xmm2 = [2,14,3,14]
; AVX512VL-NEXT: vpermt2ps %ymm1, %ymm2, %ymm0
; AVX512VL-NEXT: retq
%r = shufflevector <8 x float> %x, <8 x float> %y, <8 x i32> <i32 2, i32 14, i32 3, i32 14, i32 undef, i32 undef, i32 undef, i32 undef>
Expand Down
10 changes: 4 additions & 6 deletions llvm/test/CodeGen/X86/vector-shuffle-512-v16.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -388,9 +388,8 @@ define <8 x float> @shuffle_v16f32_extract_256(float* %RET, float* %a) {
define <8 x float> @test_v16f32_0_1_2_3_4_6_7_10 (<16 x float> %v) {
; ALL-LABEL: test_v16f32_0_1_2_3_4_6_7_10:
; ALL: # %bb.0:
; ALL-NEXT: vbroadcasti64x4 {{.*#+}} zmm1 = [0,1,2,3,4,6,7,10,0,1,2,3,4,6,7,10]
; ALL-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3]
; ALL-NEXT: vpermd %zmm0, %zmm1, %zmm0
; ALL-NEXT: vmovaps {{.*#+}} ymm1 = [0,1,2,3,4,6,7,10]
; ALL-NEXT: vpermps %zmm0, %zmm1, %zmm0
; ALL-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; ALL-NEXT: retq
%res = shufflevector <16 x float> %v, <16 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 6, i32 7, i32 10>
Expand All @@ -401,9 +400,8 @@ define <8 x float> @test_v16f32_0_1_2_3_4_6_7_10 (<16 x float> %v) {
define <4 x float> @test_v16f32_0_1_3_6 (<16 x float> %v) {
; ALL-LABEL: test_v16f32_0_1_3_6:
; ALL: # %bb.0:
; ALL-NEXT: vbroadcasti32x4 {{.*#+}} zmm1 = [0,1,3,6,0,1,3,6,0,1,3,6,0,1,3,6]
; ALL-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]
; ALL-NEXT: vpermd %zmm0, %zmm1, %zmm0
; ALL-NEXT: vmovaps {{.*#+}} xmm1 = [0,1,3,6]
; ALL-NEXT: vpermps %zmm0, %zmm1, %zmm0
; ALL-NEXT: # kill: def $xmm0 killed $xmm0 killed $zmm0
; ALL-NEXT: vzeroupper
; ALL-NEXT: retq
Expand Down
44 changes: 20 additions & 24 deletions llvm/test/CodeGen/X86/vector-shuffle-512-v8.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -94,13 +94,13 @@ define <8 x double> @shuffle_v8f64_00003000(<8 x double> %a, <8 x double> %b) {
define <8 x double> @shuffle_v8f64_00040000(<8 x double> %a, <8 x double> %b) {
; AVX512F-LABEL: shuffle_v8f64_00040000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,4,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,4]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8f64_00040000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,0,0,0,4,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,4,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -110,13 +110,13 @@ define <8 x double> @shuffle_v8f64_00040000(<8 x double> %a, <8 x double> %b) {
define <8 x double> @shuffle_v8f64_00500000(<8 x double> %a, <8 x double> %b) {
; AVX512F-LABEL: shuffle_v8f64_00500000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,5,0,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,5,0]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8f64_00500000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,0,5,0,0,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,5,0,0,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -126,13 +126,13 @@ define <8 x double> @shuffle_v8f64_00500000(<8 x double> %a, <8 x double> %b) {
define <8 x double> @shuffle_v8f64_06000000(<8 x double> %a, <8 x double> %b) {
; AVX512F-LABEL: shuffle_v8f64_06000000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,6,0,0,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,6,0,0,0,0,0,0,0]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8f64_06000000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,6,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,6,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x double> %a, <8 x double> %b, <8 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand Down Expand Up @@ -912,13 +912,13 @@ define <8 x i64> @shuffle_v8i64_00003000(<8 x i64> %a, <8 x i64> %b) {
define <8 x i64> @shuffle_v8i64_00040000(<8 x i64> %a, <8 x i64> %b) {
; AVX512F-LABEL: shuffle_v8i64_00040000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,4,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,4]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8i64_00040000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,0,0,0,4,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,4,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -928,13 +928,13 @@ define <8 x i64> @shuffle_v8i64_00040000(<8 x i64> %a, <8 x i64> %b) {
define <8 x i64> @shuffle_v8i64_00500000(<8 x i64> %a, <8 x i64> %b) {
; AVX512F-LABEL: shuffle_v8i64_00500000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,5,0,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,5,0]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8i64_00500000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,0,0,5,0,0,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,5,0,0,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand All @@ -944,13 +944,13 @@ define <8 x i64> @shuffle_v8i64_00500000(<8 x i64> %a, <8 x i64> %b) {
define <8 x i64> @shuffle_v8i64_06000000(<8 x i64> %a, <8 x i64> %b) {
; AVX512F-LABEL: shuffle_v8i64_06000000:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovaps {{.*#+}} zmm1 = [0,6,0,0,0,0,0,0]
; AVX512F-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,6,0,0,0,0,0,0,0]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: shuffle_v8i64_06000000:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vmovaps {{.*#+}} zmm1 = [0,0,6,0,0,0,0,0,0,0,0,0,0,0,0,0]
; AVX512F-32-NEXT: vmovaps {{.*#+}} xmm1 = [0,0,6,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: retl
%shuffle = shufflevector <8 x i64> %a, <8 x i64> %b, <8 x i32> <i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
Expand Down Expand Up @@ -2187,17 +2187,15 @@ define <8 x double> @shuffle_v2f64_v8f64_01010101(<2 x double> %a) {
define <4 x double> @test_v8f64_2346 (<8 x double> %v) {
; AVX512F-LABEL: test_v8f64_2346:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vbroadcasti64x4 {{.*#+}} zmm1 = [2,3,4,6,2,3,4,6]
; AVX512F-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3]
; AVX512F-NEXT: vpermq %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [2,3,4,6]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: test_v8f64_2346:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vbroadcasti64x4 {{.*#+}} zmm1 = [2,0,3,0,4,0,6,0,2,0,3,0,4,0,6,0]
; AVX512F-32-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3]
; AVX512F-32-NEXT: vpermq %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [2,0,3,0,4,0,6,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512F-32-NEXT: retl
%res = shufflevector <8 x double> %v, <8 x double> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 6>
Expand All @@ -2221,17 +2219,15 @@ define <2 x double> @test_v8f64_34 (<8 x double> %v) {
define <4 x i64> @test_v8i64_1257 (<8 x i64> %v) {
; AVX512F-LABEL: test_v8i64_1257:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vbroadcasti64x4 {{.*#+}} zmm1 = [1,2,5,7,1,2,5,7]
; AVX512F-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3]
; AVX512F-NEXT: vpermq %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: vmovaps {{.*#+}} ymm1 = [1,2,5,7]
; AVX512F-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512F-NEXT: retq
;
; AVX512F-32-LABEL: test_v8i64_1257:
; AVX512F-32: # %bb.0:
; AVX512F-32-NEXT: vbroadcasti64x4 {{.*#+}} zmm1 = [1,0,2,0,5,0,7,0,1,0,2,0,5,0,7,0]
; AVX512F-32-NEXT: # zmm1 = mem[0,1,2,3,0,1,2,3]
; AVX512F-32-NEXT: vpermq %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: vmovaps {{.*#+}} ymm1 = [1,0,2,0,5,0,7,0]
; AVX512F-32-NEXT: vpermpd %zmm0, %zmm1, %zmm0
; AVX512F-32-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512F-32-NEXT: retl
%res = shufflevector <8 x i64> %v, <8 x i64> undef, <4 x i32> <i32 1, i32 2, i32 5, i32 7>
Expand Down