[X86] Improve inst tuning tests for X86FixupInstTuning Pass; NFC

[goldstein.w.n]

Summary

1. Add tests for unpckps.
2. Add explicit test for fast shuffles (ICX+) but WITH bypass delay.

[goldstein.w.n]

Rebase

[RKSimon]

LGTM

[goldstein.w.n]

Add SnB test cases

[goldstein.w.n]

Add SnB testcases

[goldstein.w.n]

Rebase

llvm/test/CodeGen/X86/tuning-shuffle-permilps-avx512.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v4  | FileCheck %s --check-prefixes=CHECK,CHECK-V4
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl,+avx512bw,+avx512dq  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX512
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=znver4  | FileCheck %s --check-prefixes=CHECK,CHECK-ZNVER4
 
 define <16 x float> @transform_VPERMILPSZrr(<16 x float> %a) nounwind {
 ; CHECK-LABEL: transform_VPERMILPSZrr:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vshufps {{.*#+}} zmm0 = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]

 ; CHECK-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-ICX-LABEL: transform_VPERMILPSZrm:
-; CHECK-ICX:       # %bb.0:
-; CHECK-ICX-NEXT:    vpshufd {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
-; CHECK-ICX-NEXT:    retq
+; CHECK-ICX-NO-BYPASS-DELAY-LABEL: transform_VPERMILPSZrm:
+; CHECK-ICX-NO-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    vpshufd {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    retq
+;
+; CHECK-ICX-BYPASS-DELAY-LABEL: transform_VPERMILPSZrm:
+; CHECK-ICX-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-BYPASS-DELAY-NEXT:    vpermilps {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-BYPASS-DELAY-NEXT:    retq
 ;
 ; CHECK-V4-LABEL: transform_VPERMILPSZrm:
 ; CHECK-V4:       # %bb.0:
 ; CHECK-V4-NEXT:    vpermilps {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
 ; CHECK-V4-NEXT:    retq
 ;
 ; CHECK-AVX512-LABEL: transform_VPERMILPSZrm:
 ; CHECK-AVX512:       # %bb.0:

 ; CHECK-ZNVER4-NEXT:    vpermilps {{.*#+}} xmm0 = mem[3,2,1,0]
 ; CHECK-ZNVER4-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-ICX-LABEL: transform_VPERMILPSZrmkz:
-; CHECK-ICX:       # %bb.0:
-; CHECK-ICX-NEXT:    kmovd %esi, %k1
-; CHECK-ICX-NEXT:    vpshufd {{.*#+}} zmm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
-; CHECK-ICX-NEXT:    retq
+; CHECK-ICX-NO-BYPASS-DELAY-LABEL: transform_VPERMILPSZrmkz:
+; CHECK-ICX-NO-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    kmovd %esi, %k1
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    vpshufd {{.*#+}} zmm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    retq
+;
+; CHECK-ICX-BYPASS-DELAY-LABEL: transform_VPERMILPSZrmkz:
+; CHECK-ICX-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-BYPASS-DELAY-NEXT:    kmovd %esi, %k1
+; CHECK-ICX-BYPASS-DELAY-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-BYPASS-DELAY-NEXT:    retq
 ;
 ; CHECK-V4-LABEL: transform_VPERMILPSZrmkz:
 ; CHECK-V4:       # %bb.0:
 ; CHECK-V4-NEXT:    kmovd %esi, %k1
 ; CHECK-V4-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
 ; CHECK-V4-NEXT:    retq
 ;
 ; CHECK-AVX512-LABEL: transform_VPERMILPSZrmkz:
; CHECK-ZNVER4-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-ICX-LABEL: transform_VPERMILPSZrmk:
-; CHECK-ICX:       # %bb.0:
-; CHECK-ICX-NEXT:    kmovd %esi, %k1
-; CHECK-ICX-NEXT:    vpshufd {{.*#+}} zmm0 {%k1} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
-; CHECK-ICX-NEXT:    retq
+; CHECK-ICX-NO-BYPASS-DELAY-LABEL: transform_VPERMILPSZrmk:
+; CHECK-ICX-NO-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    kmovd %esi, %k1
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    vpshufd {{.*#+}} zmm0 {%k1} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-NO-BYPASS-DELAY-NEXT:    retq
+;
+; CHECK-ICX-BYPASS-DELAY-LABEL: transform_VPERMILPSZrmk:
+; CHECK-ICX-BYPASS-DELAY:       # %bb.0:
+; CHECK-ICX-BYPASS-DELAY-NEXT:    kmovd %esi, %k1
+; CHECK-ICX-BYPASS-DELAY-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-ICX-BYPASS-DELAY-NEXT:    retq
 ;
 ; CHECK-V4-LABEL: transform_VPERMILPSZrmk:
 ; CHECK-V4:       # %bb.0:
 ; CHECK-V4-NEXT:    kmovd %esi, %k1
 ; CHECK-V4-NEXT:    vpermilps {{.*#+}} zmm0 {%k1} = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
 ; CHECK-V4-NEXT:    retq
 ;
 ; CHECK-AVX512-LABEL: transform_VPERMILPSZrmk:

llvm/test/CodeGen/X86/tuning-shuffle-permilps.ll

 ; 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-AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown --mattr=+avx  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX1-DELAY
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown --mattr=+avx2,+no-bypass-delay-shuffle  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown --mattr=+avx2  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
 
 define <8 x float> @transform_VPERMILPSYrr(<8 x float> %a) nounwind {
 ; CHECK-LABEL: transform_VPERMILPSYrr:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vshufps {{.*#+}} 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

 ; CHECK-AVX2:       # %bb.0:
 ; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
 ; CHECK-AVX2-NEXT:    retq
 ;
 ; CHECK-AVX2-DELAY-LABEL: transform_VPERMILPSYrm:
 ; CHECK-AVX2-DELAY:       # %bb.0:
 ; CHECK-AVX2-DELAY-NEXT:    vpermilps {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
 ; CHECK-AVX2-DELAY-NEXT:    retq
+;
+; CHECK-ICX-LABEL: transform_VPERMILPSYrm:
+; CHECK-ICX:       # %bb.0:
+; CHECK-ICX-NEXT:    vpshufd {{.*#+}} ymm0 = mem[3,2,1,0,7,6,5,4]
+; CHECK-ICX-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-AVX1-LABEL: transform_VPERMILPSrm:
 ; CHECK-AVX1:       # %bb.0:

 ; CHECK-AVX2:       # %bb.0:
 ; CHECK-AVX2-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,2,1,0]
 ; CHECK-AVX2-NEXT:    retq
 ;
 ; CHECK-AVX2-DELAY-LABEL: transform_VPERMILPSrm:
 ; CHECK-AVX2-DELAY:       # %bb.0:
 ; CHECK-AVX2-DELAY-NEXT:    vpermilps {{.*#+}} xmm0 = mem[3,2,1,0]
 ; CHECK-AVX2-DELAY-NEXT:    retq
+;
+; CHECK-ICX-LABEL: transform_VPERMILPSrm:
+; CHECK-ICX:       # %bb.0:
+; CHECK-ICX-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,2,1,0]
+; CHECK-ICX-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
 }
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-ICX-BYPASS-DELAY: {{.*}}
+; CHECK-ICX-NO-BYPASS-DELAY: {{.*}}

llvm/test/CodeGen/X86/tuning-shuffle-unpckpd-avx512.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=skylake-avx512  | FileCheck %s --check-prefixes=CHECK,CHECK-SKX
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v4  | FileCheck %s --check-prefixes=CHECK,CHECK-V4
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl,+avx512bw,+avx512dq  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX512
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=znver4  | FileCheck %s --check-prefixes=CHECK,CHECK-ZNVER4
 
 
 define <16 x float> @transform_VUNPCKLPDZrr(<16 x float> %a, <16 x float> %b) nounwind {
 ; CHECK-LABEL: transform_VUNPCKLPDZrr:
 ; CHECK:       # %bb.0:

 ; CHECK-NEXT:    vmovapd %xmm1, %xmm0
 ; CHECK-NEXT:    retq
   %mask = bitcast i2 %mask_int to <2 x i1>
   %b = load <2 x double>, ptr %pb
   %shufp = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 1, i32 3>
   %res = select <2 x i1> %mask, <2 x double> %shufp, <2 x double> %c
   ret <2 x double> %res
 }
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-ICX-BYPASS-DELAY: {{.*}}
+; CHECK-ICX-NO-BYPASS-DELAY: {{.*}}

llvm/test/CodeGen/X86/tuning-shuffle-unpckpd.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=skylake  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2,CHECK-SKL
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v3  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2,CHECK-V3
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
 
 define <8 x float> @transform_VUNPCKLPDYrr(<8 x float> %a, <8 x float> %b) nounwind {
 ; CHECK-AVX2-LABEL: transform_VUNPCKLPDYrr:
 ; CHECK-AVX2:       # %bb.0:
 ; CHECK-AVX2-NEXT:    vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
 ; CHECK-AVX2-NEXT:    retq
 ;
 ; CHECK-ICX-LABEL: transform_VUNPCKLPDYrr:

 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vunpckhpd {{.*#+}} xmm0 = xmm0[1],mem[1]
 ; CHECK-NEXT:    retq
   %b = load <4 x float>, ptr %pb
   %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 3, i32 6, i32 7>
   ret <4 x float> %shufp
 }
 ;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-ICX-BYPASS-DELAY: {{.*}}
+; CHECK-ICX-NO-BYPASS-DELAY: {{.*}}
 ; CHECK-SKL: {{.*}}
 ; CHECK-V3: {{.*}}

llvm/test/CodeGen/X86/tuning-shuffle-unpckps-avx512.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=skylake-avx512  | FileCheck %s --check-prefixes=CHECK,CHECK-SKX
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v4  | FileCheck %s --check-prefixes=CHECK,CHECK-V4
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl,+avx512bw,+avx512dq  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX512
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=znver4  | FileCheck %s --check-prefixes=CHECK,CHECK-ZNVER4
+
+define <16 x float> @transform_VUNPCKLPSZrr(<16 x float> %a, <16 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  ret <16 x float> %shufp
+}
+
+define <16 x float> @transform_VUNPCKHPSZrr(<16 x float> %a, <16 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  ret <16 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKLPSYrr(<8 x float> %a, <8 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  ret <8 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKHPSYrr(<8 x float> %a, <8 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKLPSrr(<4 x float> %a, <4 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  ret <4 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKHPSrr(<4 x float> %a, <4 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  ret <4 x float> %shufp
+}
+
+define <16 x float> @transform_VUNPCKLPSZrrkz(<16 x float> %a, <16 x float> %b, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKHPSZrrkz(<16 x float> %a, <16 x float> %b, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm0 {%k1} {z} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKLPSYrrkz(<8 x float> %a, <8 x float> %b, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 {%k1} {z} = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKHPSYrrkz(<8 x float> %a, <8 x float> %b, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 {%k1} {z} = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKLPSrrkz(<4 x float> %a, <4 x float> %b, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 {%k1} {z} = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKHPSrrkz(<4 x float> %a, <4 x float> %b, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrrkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 {%k1} {z} = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKLPSZrrk(<16 x float> %a, <16 x float> %b, <16 x float> %c, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm2 {%k1} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %c
+  ret <16 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKHPSZrrk(<16 x float> %a, <16 x float> %b, <16 x float> %c, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm2 {%k1} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %c
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKLPSYrrk(<8 x float> %a, <8 x float> %b, <8 x float> %c, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm2 {%k1} = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %c
+  ret <8 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKHPSYrrk(<8 x float> %a, <8 x float> %b, <8 x float> %c, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm2 {%k1} = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %c
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKLPSrrk(<4 x float> %a, <4 x float> %b, <4 x float> %c, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm2 {%k1} = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %c
+  ret <4 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKHPSrrk(<4 x float> %a, <4 x float> %b, <4 x float> %c, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrrk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm2 {%k1} = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %c
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKLPSZrm(<16 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm0 = zmm0[0],mem[0],zmm0[1],mem[1],zmm0[4],mem[4],zmm0[5],mem[5],zmm0[8],mem[8],zmm0[9],mem[9],zmm0[12],mem[12],zmm0[13],mem[13]
+; CHECK-NEXT:    retq
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  ret <16 x float> %shufp
+}
+
+define <16 x float> @transform_VUNPCKHPSZrm(<16 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm0 = zmm0[2],mem[2],zmm0[3],mem[3],zmm0[6],mem[6],zmm0[7],mem[7],zmm0[10],mem[10],zmm0[11],mem[11],zmm0[14],mem[14],zmm0[15],mem[15]
+; CHECK-NEXT:    retq
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  ret <16 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKLPSYrm(<8 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 = ymm0[0],mem[0],ymm0[1],mem[1],ymm0[4],mem[4],ymm0[5],mem[5]
+; CHECK-NEXT:    retq
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  ret <8 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKHPSYrm(<8 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 = ymm0[2],mem[2],ymm0[3],mem[3],ymm0[6],mem[6],ymm0[7],mem[7]
+; CHECK-NEXT:    retq
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKLPSrm(<4 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    retq
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  ret <4 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKHPSrm(<4 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2],mem[2],xmm0[3],mem[3]
+; CHECK-NEXT:    retq
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  ret <4 x float> %shufp
+}
+
+define <16 x float> @transform_VUNPCKLPSZrmkz(<16 x float> %a, ptr %pb, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm0 {%k1} {z} = zmm0[0],mem[0],zmm0[1],mem[1],zmm0[4],mem[4],zmm0[5],mem[5],zmm0[8],mem[8],zmm0[9],mem[9],zmm0[12],mem[12],zmm0[13],mem[13]
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKHPSZrmkz(<16 x float> %a, ptr %pb, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm0 {%k1} {z} = zmm0[2],mem[2],zmm0[3],mem[3],zmm0[6],mem[6],zmm0[7],mem[7],zmm0[10],mem[10],zmm0[11],mem[11],zmm0[14],mem[14],zmm0[15],mem[15]
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> zeroinitializer
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKLPSYrmkz(<8 x float> %a, ptr %pb, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 {%k1} {z} = ymm0[0],mem[0],ymm0[1],mem[1],ymm0[4],mem[4],ymm0[5],mem[5]
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKHPSYrmkz(<8 x float> %a, ptr %pb, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 {%k1} {z} = ymm0[2],mem[2],ymm0[3],mem[3],ymm0[6],mem[6],ymm0[7],mem[7]
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> zeroinitializer
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKLPSrmkz(<4 x float> %a, ptr %pb, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 {%k1} {z} = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKHPSrmkz(<4 x float> %a, ptr %pb, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrmkz:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 {%k1} {z} = xmm0[2],mem[2],xmm0[3],mem[3]
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> zeroinitializer
+  ret <4 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKLPSZrmk(<16 x float> %a, ptr %pb, <16 x float> %c, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSZrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} zmm1 {%k1} = zmm0[0],mem[0],zmm0[1],mem[1],zmm0[4],mem[4],zmm0[5],mem[5],zmm0[8],mem[8],zmm0[9],mem[9],zmm0[12],mem[12],zmm0[13],mem[13]
+; CHECK-NEXT:    vmovaps %zmm1, %zmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %c
+  ret <16 x float> %res
+}
+
+define <16 x float> @transform_VUNPCKHPSZrmk(<16 x float> %a, ptr %pb, <16 x float> %c, i16 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSZrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} zmm1 {%k1} = zmm0[2],mem[2],zmm0[3],mem[3],zmm0[6],mem[6],zmm0[7],mem[7],zmm0[10],mem[10],zmm0[11],mem[11],zmm0[14],mem[14],zmm0[15],mem[15]
+; CHECK-NEXT:    vmovaps %zmm1, %zmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i16 %mask_int to <16 x i1>
+  %b = load <16 x float>, ptr %pb
+  %shufp = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
+  %res = select <16 x i1> %mask, <16 x float> %shufp, <16 x float> %c
+  ret <16 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKLPSYrmk(<8 x float> %a, ptr %pb, <8 x float> %c, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm1 {%k1} = ymm0[0],mem[0],ymm0[1],mem[1],ymm0[4],mem[4],ymm0[5],mem[5]
+; CHECK-NEXT:    vmovaps %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %c
+  ret <8 x float> %res
+}
+
+define <8 x float> @transform_VUNPCKHPSYrmk(<8 x float> %a, ptr %pb, <8 x float> %c, i8 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm1 {%k1} = ymm0[2],mem[2],ymm0[3],mem[3],ymm0[6],mem[6],ymm0[7],mem[7]
+; CHECK-NEXT:    vmovaps %ymm1, %ymm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i8 %mask_int to <8 x i1>
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  %res = select <8 x i1> %mask, <8 x float> %shufp, <8 x float> %c
+  ret <8 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKLPSrmk(<4 x float> %a, ptr %pb, <4 x float> %c, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm1 {%k1} = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    vmovaps %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %c
+  ret <4 x float> %res
+}
+
+define <4 x float> @transform_VUNPCKHPSrmk(<4 x float> %a, ptr %pb, <4 x float> %c, i4 %mask_int) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrmk:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    kmovd %esi, %k1
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm1 {%k1} = xmm0[2],mem[2],xmm0[3],mem[3]
+; CHECK-NEXT:    vmovaps %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %mask = bitcast i4 %mask_int to <4 x i1>
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  %res = select <4 x i1> %mask, <4 x float> %shufp, <4 x float> %c
+  ret <4 x float> %res
+}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-AVX512: {{.*}}
+; CHECK-ICX: {{.*}}
+; CHECK-ICX-BYPASS-DELAY: {{.*}}
+; CHECK-ICX-NO-BYPASS-DELAY: {{.*}}
+; CHECK-SKX: {{.*}}
+; CHECK-V4: {{.*}}
+; CHECK-ZNVER4: {{.*}}

llvm/test/CodeGen/X86/tuning-shuffle-unpckps.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=skylake  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2,CHECK-SKL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64-v3  | FileCheck %s --check-prefixes=CHECK,CHECK-AVX2,CHECK-V3
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server  | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-NO-BYPASS-DELAY
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=icelake-server -mattr=-no-bypass-delay-shuffle | FileCheck %s --check-prefixes=CHECK,CHECK-ICX,CHECK-ICX-BYPASS-DELAY
+
+define <8 x float> @transform_VUNPCKLPSYrr(<8 x float> %a, <8 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  ret <8 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKHPSYrr(<8 x float> %a, <8 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKLPSrr(<4 x float> %a, <4 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  ret <4 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKHPSrr(<4 x float> %a, <4 x float> %b) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-NEXT:    retq
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  ret <4 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKLPSYrm(<8 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSYrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} ymm0 = ymm0[0],mem[0],ymm0[1],mem[1],ymm0[4],mem[4],ymm0[5],mem[5]
+; CHECK-NEXT:    retq
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 4, i32 12, i32 5, i32 13>
+  ret <8 x float> %shufp
+}
+
+define <8 x float> @transform_VUNPCKHPSYrm(<8 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSYrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} ymm0 = ymm0[2],mem[2],ymm0[3],mem[3],ymm0[6],mem[6],ymm0[7],mem[7]
+; CHECK-NEXT:    retq
+  %b = load <8 x float>, ptr %pb
+  %shufp = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15>
+  ret <8 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKLPSrm(<4 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKLPSrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    retq
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  ret <4 x float> %shufp
+}
+
+define <4 x float> @transform_VUNPCKHPSrm(<4 x float> %a, ptr %pb) nounwind {
+; CHECK-LABEL: transform_VUNPCKHPSrm:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2],mem[2],xmm0[3],mem[3]
+; CHECK-NEXT:    retq
+  %b = load <4 x float>, ptr %pb
+  %shufp = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  ret <4 x float> %shufp
+}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; CHECK-AVX2: {{.*}}
+; CHECK-ICX: {{.*}}
+; CHECK-ICX-BYPASS-DELAY: {{.*}}
+; CHECK-ICX-NO-BYPASS-DELAY: {{.*}}
+; CHECK-SKL: {{.*}}
+; CHECK-V3: {{.*}}