diff --git a/llvm/test/CodeGen/X86/tuning-shuffle-permilps.ll b/llvm/test/CodeGen/X86/tuning-shuffle-permilps.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/X86/tuning-shuffle-permilps.ll
@@ -0,0 +1,568 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown --mattr=+avx,+no-bypass-delay-shuffle  | FileCheck %s --check-prefixes=CHECK,CHECK-AV,CHECK-AVX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown --mattr=+avx2,+no-bypass-delay-shuffle  | FileCheck %s --check-prefixes=CHECK,CHECK-AV,CHECK-AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-512,CHECK-ICX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v4  | FileCheck %s --check-prefixes=CHECK,CHECK-512,CHECK-V4
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=znver4  | FileCheck %s --check-prefixes=CHECK,CHECK-512,CHECK-ZNVER4
+
+define <16 x float> @transform_VPERMILPSZrr(<16 x float> %a) nounwind {
+; CHECK-AV-LABEL: transform_VPERMILPSZrr:
+; CHECK-AV:       # %bb.0:
+; CHECK-AV-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AV-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
+; CHECK-AV-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrr:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm0 = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    retq
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  ret <16 x float> %shufp
+}
+
+define <8 x float> @transform_VPERMILPSYrr(<8 x float> %a) nounwind {
+; CHECK-LABEL: transform_VPERMILPSYrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VPERMILPSrr(<4 x float> %a) nounwind {
+; CHECK-LABEL: transform_VPERMILPSrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  ret <4 x float> %shufp
+}
+
+define <16 x float> @transform_VPERMILPSZrrkz(<16 x float> %a, i16 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSZrrkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm4
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm3, %xmm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm3, %ymm4, %ymm3
+; CHECK-AVX-NEXT:    vandps %ymm0, %ymm3, %ymm0
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm3
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm3, %ymm2
+; CHECK-AVX-NEXT:    vandps %ymm1, %ymm2, %ymm1
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSZrrkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastw %xmm2, %ymm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm4
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm4, %ymm3
+; CHECK-AVX2-NEXT:    vpand %ymm0, %ymm3, %ymm0
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [256,512,1024,2048,4096,8192,16384,32768]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpand %ymm1, %ymm2, %ymm1
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrrkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} {z} = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VPERMILPSYrrkz(<8 x float> %a, i8 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSYrrkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm1
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm1, %ymm1
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm1
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm2
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm1, %xmm1
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm2, %ymm1
+; CHECK-AVX-NEXT:    vandps %ymm0, %ymm1, %ymm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSYrrkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm1
+; CHECK-AVX2-NEXT:    vpbroadcastb %xmm1, %ymm1
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm2 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm2, %ymm1, %ymm1
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm2, %ymm1, %ymm1
+; CHECK-AVX2-NEXT:    vpand %ymm0, %ymm1, %ymm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSYrrkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} ymm0 {%k1} {z} = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VPERMILPSrrkz(<4 x float> %a, i4 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSrrkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm1
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; CHECK-AVX-NEXT:    vmovdqa {{.*#+}} xmm2 = [1,2,4,8]
+; CHECK-AVX-NEXT:    vpand %xmm2, %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vpcmpeqd %xmm2, %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSrrkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm1
+; CHECK-AVX2-NEXT:    vpbroadcastd %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} xmm2 = [1,2,4,8]
+; CHECK-AVX2-NEXT:    vpand %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vpcmpeqd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vpand %xmm0, %xmm1, %xmm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSrrkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} xmm0 {%k1} {z} = xmm0[3,2,1,0]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VPERMILPSZrrk(<16 x float> %a, <16 x float> %b, i16 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSZrrk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm4
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm4 = xmm4[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm4, %ymm4, %ymm4
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm4, %ymm5
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm5, %xmm6
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm5, %xmm5
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm5, %xmm5
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm5, %ymm6, %ymm5
+; CHECK-AVX-NEXT:    vblendvps %ymm5, %ymm0, %ymm2, %ymm0
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm4, %ymm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm4
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm4, %ymm2
+; CHECK-AVX-NEXT:    vblendvps %ymm2, %ymm1, %ymm3, %ymm1
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSZrrk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm4
+; CHECK-AVX2-NEXT:    vpbroadcastw %xmm4, %ymm4
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm5 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm5, %ymm4, %ymm6
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm5, %ymm6, %ymm5
+; CHECK-AVX2-NEXT:    vblendvps %ymm5, %ymm0, %ymm2, %ymm0
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm2 = [256,512,1024,2048,4096,8192,16384,32768]
+; CHECK-AVX2-NEXT:    vpand %ymm2, %ymm4, %ymm4
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm2, %ymm4, %ymm2
+; CHECK-AVX2-NEXT:    vblendvps %ymm2, %ymm1, %ymm3, %ymm1
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrrk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm1 {%k1} = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    vmovaps %zmm1, %zmm0
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %b
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VPERMILPSYrrk(<8 x float> %a, <8 x float> %b, i8 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSYrrk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm3
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm3, %ymm2
+; CHECK-AVX-NEXT:    vblendvps %ymm2, %ymm0, %ymm1, %ymm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSYrrk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastb %xmm2, %ymm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vblendvps %ymm2, %ymm0, %ymm1, %ymm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSYrrk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} ymm1 {%k1} = ymm0[3,2,1,0,7,6,5,4]
+; CHECK-512-NEXT:    vmovaps %ymm1, %ymm0
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %b
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VPERMILPSrrk(<4 x float> %a, <4 x float> %b, i4 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSrrk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-AVX-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vmovdqa {{.*#+}} xmm3 = [1,2,4,8]
+; CHECK-AVX-NEXT:    vpand %xmm3, %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd %xmm3, %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vblendvps %xmm2, %xmm0, %xmm1, %xmm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSrrk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; CHECK-AVX2-NEXT:    vmovd %edi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastd %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} xmm3 = [1,2,4,8]
+; CHECK-AVX2-NEXT:    vpand %xmm3, %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %xmm3, %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vblendvps %xmm2, %xmm0, %xmm1, %xmm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSrrk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %edi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} xmm1 {%k1} = xmm0[3,2,1,0]
+; CHECK-512-NEXT:    vmovaps %xmm1, %xmm0
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %b
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VPERMILPSZrm(ptr %ap) nounwind {
+; CHECK-AV-LABEL: transform_VPERMILPSZrm:
+; CHECK-AV:       # %bb.0:
+; CHECK-AV-NEXT:    vpermilps {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AV-NEXT:    vpermilps {{.*#+}} ymm1 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AV-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrm:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    retq
+  %a = load <16 x float>, ptr %ap
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  ret <16 x float> %shufp
+}
+
+define <8 x float> @transform_VPERMILPSYrm(ptr %ap) nounwind {
+; CHECK-LABEL: transform_VPERMILPSYrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vpermilps {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-NEXT:    retq
+  %a = load <8 x float>, ptr %ap
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VPERMILPSrm(ptr %ap) nounwind {
+; CHECK-LABEL: transform_VPERMILPSrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vpermilps {{.*#+}} xmm0 = mem[3,2,1,0]
+; CHECK-NEXT:    retq
+  %a = load <4 x float>, ptr %ap
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  ret <4 x float> %shufp
+}
+
+define <16 x float> @transform_VPERMILPSZrmkz(ptr %ap, i16 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSZrmkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm1 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm4
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm3, %xmm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm3, %ymm4, %ymm3
+; CHECK-AVX-NEXT:    vandps %ymm0, %ymm3, %ymm0
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm3
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm3, %ymm2
+; CHECK-AVX-NEXT:    vandps %ymm1, %ymm2, %ymm1
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSZrmkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm1 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastw %xmm2, %ymm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm4
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm4, %ymm3
+; CHECK-AVX2-NEXT:    vpand %ymm0, %ymm3, %ymm0
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [256,512,1024,2048,4096,8192,16384,32768]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpand %ymm1, %ymm2, %ymm1
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrmkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %a = load <16 x float>, ptr %ap
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VPERMILPSYrmkz(ptr %ap, i8 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSYrmkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm1
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm1, %ymm1
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm1
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm2
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm1, %xmm1
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm2, %ymm1
+; CHECK-AVX-NEXT:    vandps %ymm0, %ymm1, %ymm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSYrmkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm1
+; CHECK-AVX2-NEXT:    vpbroadcastb %xmm1, %ymm1
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm2 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm2, %ymm1, %ymm1
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm2, %ymm1, %ymm1
+; CHECK-AVX2-NEXT:    vpand %ymm0, %ymm1, %ymm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSYrmkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} ymm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %a = load <8 x float>, ptr %ap
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VPERMILPSrmkz(ptr %ap, i4 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSrmkz:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,2,1,0]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm1
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; CHECK-AVX-NEXT:    vmovdqa {{.*#+}} xmm2 = [1,2,4,8]
+; CHECK-AVX-NEXT:    vpand %xmm2, %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vpcmpeqd %xmm2, %xmm1, %xmm1
+; CHECK-AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSrmkz:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,2,1,0]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm1
+; CHECK-AVX2-NEXT:    vpbroadcastd %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} xmm2 = [1,2,4,8]
+; CHECK-AVX2-NEXT:    vpand %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vpcmpeqd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT:    vpand %xmm0, %xmm1, %xmm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSrmkz:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} xmm0 {%k1} {z} = mem[3,2,1,0]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %a = load <4 x float>, ptr %ap
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VPERMILPSZrmk(ptr %ap, <16 x float> %b, i16 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSZrmk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm2 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm3 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm4
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm4 = xmm4[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm4, %ymm4, %ymm4
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm4, %ymm5
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm5, %xmm6
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm5, %xmm5
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm5, %xmm5
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm5, %ymm6, %ymm5
+; CHECK-AVX-NEXT:    vblendvps %ymm5, %ymm3, %ymm0, %ymm0
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm4, %ymm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm4
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm3, %xmm3
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm3, %ymm4, %ymm3
+; CHECK-AVX-NEXT:    vblendvps %ymm3, %ymm2, %ymm1, %ymm1
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSZrmk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm2 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm3 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm4
+; CHECK-AVX2-NEXT:    vpbroadcastw %xmm4, %ymm4
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm5 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm5, %ymm4, %ymm6
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm5, %ymm6, %ymm5
+; CHECK-AVX2-NEXT:    vblendvps %ymm5, %ymm3, %ymm0, %ymm0
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [256,512,1024,2048,4096,8192,16384,32768]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm4, %ymm4
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm4, %ymm3
+; CHECK-AVX2-NEXT:    vblendvps %ymm3, %ymm2, %ymm1, %ymm1
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSZrmk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %a = load <16 x float>, ptr %ap
+  %shufp = shufflevector <16 x float> %a, <16 x float> poison, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %b
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VPERMILPSYrmk(ptr %ap, <8 x float> %b, i8 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSYrmk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} ymm1 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm3
+; CHECK-AVX-NEXT:    vextractf128 $1, %ymm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vinsertf128 $1, %xmm2, %ymm3, %ymm2
+; CHECK-AVX-NEXT:    vblendvps %ymm2, %ymm1, %ymm0, %ymm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSYrmk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} ymm1 = mem[3,2,1,0,7,6,5,4]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastb %xmm2, %ymm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128]
+; CHECK-AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %ymm3, %ymm2, %ymm2
+; CHECK-AVX2-NEXT:    vblendvps %ymm2, %ymm1, %ymm0, %ymm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSYrmk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} ymm0 {%k1} = mem[3,2,1,0,7,6,5,4]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %a = load <8 x float>, ptr %ap
+  %shufp = shufflevector <8 x float> %a, <8 x float> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %b
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VPERMILPSrmk(ptr %ap, <4 x float> %b, i4 %mask_int) nounwind {
+; CHECK-AVX-LABEL: transform_VPERMILPSrmk:
+; CHECK-AVX:       # %bb.0:
+; CHECK-AVX-NEXT:    vpermilps {{.*#+}} xmm1 = mem[3,2,1,0]
+; CHECK-AVX-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; CHECK-AVX-NEXT:    vmovdqa {{.*#+}} xmm3 = [1,2,4,8]
+; CHECK-AVX-NEXT:    vpand %xmm3, %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vpcmpeqd %xmm3, %xmm2, %xmm2
+; CHECK-AVX-NEXT:    vblendvps %xmm2, %xmm1, %xmm0, %xmm0
+; CHECK-AVX-NEXT:    retq
+;
+; CHECK-AVX2-LABEL: transform_VPERMILPSrmk:
+; CHECK-AVX2:       # %bb.0:
+; CHECK-AVX2-NEXT:    vpermilps {{.*#+}} xmm1 = mem[3,2,1,0]
+; CHECK-AVX2-NEXT:    vmovd %esi, %xmm2
+; CHECK-AVX2-NEXT:    vpbroadcastd %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vmovdqa {{.*#+}} xmm3 = [1,2,4,8]
+; CHECK-AVX2-NEXT:    vpand %xmm3, %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vpcmpeqd %xmm3, %xmm2, %xmm2
+; CHECK-AVX2-NEXT:    vblendvps %xmm2, %xmm1, %xmm0, %xmm0
+; CHECK-AVX2-NEXT:    retq
+;
+; CHECK-512-LABEL: transform_VPERMILPSrmk:
+; CHECK-512:       # %bb.0:
+; CHECK-512-NEXT:    kmovd %esi, %k1
+; CHECK-512-NEXT:    vpermilps {{.*#+}} xmm0 {%k1} = mem[3,2,1,0]
+; CHECK-512-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %a = load <4 x float>, ptr %ap
+  %shufp = shufflevector <4 x float> %a, <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %b
+  ret <4 x float> %res
+}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-ICX: {{.*}}
+; CHECK-V4: {{.*}}
+; CHECK-ZNVER4: {{.*}}