[X86][SSE] Check vec_set BUILD_VECTOR tests on both 32 and 64-bit targets

llvm-svn: 302683

10 files changed, 204 insertions, 99 deletions

diff --git a/llvm/test/CodeGen/X86/vec_set-2.ll b/llvm/test/CodeGen/X86/vec_set-2.ll
index 443264cdffd..51c8b211110 100644
--- a/llvm/test/CodeGen/X86/vec_set-2.ll
+++ b/llvm/test/CodeGen/X86/vec_set-2.ll
@@ -1,11 +1,19 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,-sse4.1 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X64
 
 define <4 x float> @test1(float %a) nounwind {
-; CHECK-LABEL: test1:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    retl
+; X86-LABEL: test1:
+; X86:       # BB#0:
+; X86-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test1:
+; X64:       # BB#0:
+; X64-NEXT:    xorps %xmm1, %xmm1
+; X64-NEXT:    movss {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
+; X64-NEXT:    movaps %xmm1, %xmm0
+; X64-NEXT:    retq
   %tmp = insertelement <4 x float> zeroinitializer, float %a, i32 0
   %tmp5 = insertelement <4 x float> %tmp, float 0.000000e+00, i32 1
   %tmp6 = insertelement <4 x float> %tmp5, float 0.000000e+00, i32 2
@@ -14,10 +22,15 @@ define <4 x float> @test1(float %a) nounwind {
 }
 
 define <2 x i64> @test(i32 %a) nounwind {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    movd %edi, %xmm0
+; X64-NEXT:    retq
   %tmp = insertelement <4 x i32> zeroinitializer, i32 %a, i32 0
   %tmp6 = insertelement <4 x i32> %tmp, i32 0, i32 1
   %tmp8 = insertelement <4 x i32> %tmp6, i32 0, i32 2
diff --git a/llvm/test/CodeGen/X86/vec_set-3.ll b/llvm/test/CodeGen/X86/vec_set-3.ll
index ee4a0859996..b34f30924a8 100644
--- a/llvm/test/CodeGen/X86/vec_set-3.ll
+++ b/llvm/test/CodeGen/X86/vec_set-3.ll
@@ -1,11 +1,17 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,+sse4.1 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,+sse4.1 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,+sse4.1 | FileCheck %s --check-prefix=X64
 
 define <4 x float> @test(float %a) {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    insertps {{.*#+}} xmm0 = zero,mem[0],zero,zero
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    insertps {{.*#+}} xmm0 = zero,mem[0],zero,zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
+; X64-NEXT:    retq
   %tmp = insertelement <4 x float> zeroinitializer, float %a, i32 1
   %tmp5 = insertelement <4 x float> %tmp, float 0.000000e+00, i32 2
   %tmp6 = insertelement <4 x float> %tmp5, float 0.000000e+00, i32 3
@@ -13,11 +19,17 @@ define <4 x float> @test(float %a) {
 }
 
 define <2 x i64> @test2(i32 %a) {
-; CHECK-LABEL: test2:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,0,1]
-; CHECK-NEXT:    retl
+; X86-LABEL: test2:
+; X86:       # BB#0:
+; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,0,1]
+; X86-NEXT:    retl
+;
+; X64-LABEL: test2:
+; X64:       # BB#0:
+; X64-NEXT:    movd %edi, %xmm0
+; X64-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,0,1]
+; X64-NEXT:    retq
   %tmp7 = insertelement <4 x i32> zeroinitializer, i32 %a, i32 2
   %tmp9 = insertelement <4 x i32> %tmp7, i32 0, i32 3
   %tmp10 = bitcast <4 x i32> %tmp9 to <2 x i64>
@@ -25,10 +37,15 @@ define <2 x i64> @test2(i32 %a) {
 }
 
 define <4 x float> @test3(<4 x float> %A) {
-; CHECK-LABEL: test3:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
-; CHECK-NEXT:    retl
+; X86-LABEL: test3:
+; X86:       # BB#0:
+; X86-NEXT:    insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test3:
+; X64:       # BB#0:
+; X64-NEXT:    insertps {{.*#+}} xmm0 = zero,xmm0[0],zero,zero
+; X64-NEXT:    retq
   %tmp0 = extractelement <4 x float> %A, i32 0
   %tmp1 = insertelement <4 x float> <float 0.000000e+00, float undef, float undef, float undef >, float %tmp0, i32 1
   %tmp2 = insertelement <4 x float> %tmp1, float 0.000000e+00, i32 2
diff --git a/llvm/test/CodeGen/X86/vec_set-4.ll b/llvm/test/CodeGen/X86/vec_set-4.ll
index 8f35529d61b..09142e16aa6 100644
--- a/llvm/test/CodeGen/X86/vec_set-4.ll
+++ b/llvm/test/CodeGen/X86/vec_set-4.ll
@@ -1,12 +1,19 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
 
 define <2 x i64> @test(i16 %a) nounwind {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    pxor %xmm0, %xmm0
-; CHECK-NEXT:    pinsrw $3, {{[0-9]+}}(%esp), %xmm0
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    pxor %xmm0, %xmm0
+; X86-NEXT:    pinsrw $3, {{[0-9]+}}(%esp), %xmm0
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    pxor %xmm0, %xmm0
+; X64-NEXT:    pinsrw $3, %edi, %xmm0
+; X64-NEXT:    retq
   %tmp10 = insertelement <8 x i16> zeroinitializer, i16 %a, i32 3
   %tmp12 = insertelement <8 x i16> %tmp10, i16 0, i32 4
   %tmp14 = insertelement <8 x i16> %tmp12, i16 0, i32 5
@@ -17,12 +24,19 @@ define <2 x i64> @test(i16 %a) nounwind {
 }
 
 define <2 x i64> @test2(i8 %a) nounwind {
-; CHECK-LABEL: test2:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT:    pxor %xmm0, %xmm0
-; CHECK-NEXT:    pinsrw $5, %eax, %xmm0
-; CHECK-NEXT:    retl
+; X86-LABEL: test2:
+; X86:       # BB#0:
+; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    pxor %xmm0, %xmm0
+; X86-NEXT:    pinsrw $5, %eax, %xmm0
+; X86-NEXT:    retl
+;
+; X64-LABEL: test2:
+; X64:       # BB#0:
+; X64-NEXT:    movzbl %dil, %eax
+; X64-NEXT:    pxor %xmm0, %xmm0
+; X64-NEXT:    pinsrw $5, %eax, %xmm0
+; X64-NEXT:    retq
   %tmp24 = insertelement <16 x i8> zeroinitializer, i8 %a, i32 10
   %tmp26 = insertelement <16 x i8> %tmp24, i8 0, i32 11
   %tmp28 = insertelement <16 x i8> %tmp26, i8 0, i32 12
diff --git a/llvm/test/CodeGen/X86/vec_set-6.ll b/llvm/test/CodeGen/X86/vec_set-6.ll
index 4429834b8ef..3c9aca3a02d 100644
--- a/llvm/test/CodeGen/X86/vec_set-6.ll
+++ b/llvm/test/CodeGen/X86/vec_set-6.ll
@@ -1,13 +1,22 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,+sse4.1 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,+sse4.1 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,+sse4.1 | FileCheck %s --check-prefix=X64
 
 define <4 x float> @test(float %a, float %b, float %c) nounwind {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movsd {{.*#+}} xmm1 = mem[0],zero
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,1]
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    movsd {{.*#+}} xmm1 = mem[0],zero
+; X86-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,1]
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
+; X64-NEXT:    xorps %xmm2, %xmm2
+; X64-NEXT:    blendps {{.*#+}} xmm0 = xmm0[0],xmm2[1,2,3]
+; X64-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,1]
+; X64-NEXT:    retq
   %tmp = insertelement <4 x float> zeroinitializer, float %a, i32 1
   %tmp8 = insertelement <4 x float> %tmp, float %b, i32 2
   %tmp10 = insertelement <4 x float> %tmp8, float %c, i32 3
diff --git a/llvm/test/CodeGen/X86/vec_set-7.ll b/llvm/test/CodeGen/X86/vec_set-7.ll
index e8fe6debb14..757a0d44cd4 100644
--- a/llvm/test/CodeGen/X86/vec_set-7.ll
+++ b/llvm/test/CodeGen/X86/vec_set-7.ll
@@ -1,12 +1,18 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
 
 define <2 x i64> @test(<2 x i64>* %p) nounwind {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X64-NEXT:    retq
   %tmp = bitcast <2 x i64>* %p to double*
   %tmp.upgrd.1 = load double, double* %tmp
   %tmp.upgrd.2 = insertelement <2 x double> undef, double %tmp.upgrd.1, i32 0
diff --git a/llvm/test/CodeGen/X86/vec_set-8.ll b/llvm/test/CodeGen/X86/vec_set-8.ll
index 7a4326c01bb..a9dceb90855 100644
--- a/llvm/test/CodeGen/X86/vec_set-8.ll
+++ b/llvm/test/CodeGen/X86/vec_set-8.ll
@@ -1,11 +1,17 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.2 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=X64
 
 define <2 x i64> @test(i64 %i) nounwind  {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movq %rdi, %xmm0
-; CHECK-NEXT:    retq
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    movq %rdi, %xmm0
+; X64-NEXT:    retq
   %tmp10 = insertelement <2 x i64> undef, i64 %i, i32 0
   %tmp11 = insertelement <2 x i64> %tmp10, i64 0, i32 1
   ret <2 x i64> %tmp11
diff --git a/llvm/test/CodeGen/X86/vec_set-A.ll b/llvm/test/CodeGen/X86/vec_set-A.ll
index cae39a3d775..259ace98d36 100644
--- a/llvm/test/CodeGen/X86/vec_set-A.ll
+++ b/llvm/test/CodeGen/X86/vec_set-A.ll
@@ -1,12 +1,19 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
 
 define <2 x i64> @test1() nounwind {
-; CHECK-LABEL: test1:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movl $1, %eax
-; CHECK-NEXT:    movd %eax, %xmm0
-; CHECK-NEXT:    retl
+; X86-LABEL: test1:
+; X86:       # BB#0:
+; X86-NEXT:    movl $1, %eax
+; X86-NEXT:    movd %eax, %xmm0
+; X86-NEXT:    retl
+;
+; X64-LABEL: test1:
+; X64:       # BB#0:
+; X64-NEXT:    movl $1, %eax
+; X64-NEXT:    movq %rax, %xmm0
+; X64-NEXT:    retq
   ret <2 x i64> < i64 1, i64 0 >
 }
 
diff --git a/llvm/test/CodeGen/X86/vec_set-B.ll b/llvm/test/CodeGen/X86/vec_set-B.ll
index 0580a337665..ecd9b57cfd0 100644
--- a/llvm/test/CodeGen/X86/vec_set-B.ll
+++ b/llvm/test/CodeGen/X86/vec_set-B.ll
@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
 
 ; These should both generate something like this:
 ;_test3:
@@ -9,26 +10,37 @@
 ;	ret
 
 define <2 x i64> @test3(i64 %arg) nounwind {
-; CHECK-LABEL: test3:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movl $1234567, %eax # imm = 0x12D687
-; CHECK-NEXT:    andl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT:    movd %eax, %xmm0
-; CHECK-NEXT:    retl
+; X86-LABEL: test3:
+; X86:       # BB#0:
+; X86-NEXT:    movl $1234567, %eax # imm = 0x12D687
+; X86-NEXT:    andl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movd %eax, %xmm0
+; X86-NEXT:    retl
+;
+; X64-LABEL: test3:
+; X64:       # BB#0:
+; X64-NEXT:    andl $1234567, %edi # imm = 0x12D687
+; X64-NEXT:    movq %rdi, %xmm0
+; X64-NEXT:    retq
   %A = and i64 %arg, 1234567
   %B = insertelement <2 x i64> zeroinitializer, i64 %A, i32 0
   ret <2 x i64> %B
 }
 
 define <2 x i64> @test2(i64 %arg) nounwind {
-; CHECK-LABEL: test2:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movl $1234567, %eax # imm = 0x12D687
-; CHECK-NEXT:    andl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT:    movd %eax, %xmm0
-; CHECK-NEXT:    retl
+; X86-LABEL: test2:
+; X86:       # BB#0:
+; X86-NEXT:    movl $1234567, %eax # imm = 0x12D687
+; X86-NEXT:    andl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movd %eax, %xmm0
+; X86-NEXT:    retl
+;
+; X64-LABEL: test2:
+; X64:       # BB#0:
+; X64-NEXT:    andl $1234567, %edi # imm = 0x12D687
+; X64-NEXT:    movq %rdi, %xmm0
+; X64-NEXT:    retq
   %A = and i64 %arg, 1234567
   %B = insertelement <2 x i64> undef, i64 %A, i32 0
   ret <2 x i64> %B
 }
-
diff --git a/llvm/test/CodeGen/X86/vec_set-C.ll b/llvm/test/CodeGen/X86/vec_set-C.ll
index 994bc2b3056..865e2fb83f1 100644
--- a/llvm/test/CodeGen/X86/vec_set-C.ll
+++ b/llvm/test/CodeGen/X86/vec_set-C.ll
@@ -1,12 +1,12 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-linux-gnu -mattr=+sse2,-avx | FileCheck %s --check-prefix=X32
+; RUN: llc < %s -mtriple=i386-linux-gnu -mattr=+sse2,-avx | FileCheck %s --check-prefix=X86
 ; RUN: llc < %s -mtriple=x86_64-pc-linux -mattr=+sse2,-avx | FileCheck %s --check-prefix=X64
 
 define <2 x i64> @t1(i64 %x) nounwind  {
-; X32-LABEL: t1:
-; X32:       # BB#0:
-; X32-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
-; X32-NEXT:    retl
+; X86-LABEL: t1:
+; X86:       # BB#0:
+; X86-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-NEXT:    retl
 ;
 ; X64-LABEL: t1:
 ; X64:       # BB#0:
diff --git a/llvm/test/CodeGen/X86/vec_set.ll b/llvm/test/CodeGen/X86/vec_set.ll
index 49bd3beef75..6439a6dcb00 100644
--- a/llvm/test/CodeGen/X86/vec_set.ll
+++ b/llvm/test/CodeGen/X86/vec_set.ll
@@ -1,27 +1,48 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,-sse4.1 | FileCheck %s
+; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X86
+; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,-sse4.1 | FileCheck %s --check-prefix=X64
 
 define void @test(<8 x i16>* %b, i16 %a0, i16 %a1, i16 %a2, i16 %a3, i16 %a4, i16 %a5, i16 %a6, i16 %a7) nounwind {
-; CHECK-LABEL: test:
-; CHECK:       # BB#0:
-; CHECK-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; CHECK-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; CHECK-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; CHECK-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; CHECK-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; CHECK-NEXT:    movd {{.*#+}} xmm3 = mem[0],zero,zero,zero
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3]
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
-; CHECK-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
-; CHECK-NEXT:    movdqa %xmm3, (%eax)
-; CHECK-NEXT:    retl
+; X86-LABEL: test:
+; X86:       # BB#0:
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X86-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
+; X86-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
+; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT:    movd {{.*#+}} xmm3 = mem[0],zero,zero,zero
+; X86-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3]
+; X86-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
+; X86-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; X86-NEXT:    movdqa %xmm3, (%eax)
+; X86-NEXT:    retl
+;
+; X64-LABEL: test:
+; X64:       # BB#0:
+; X64-NEXT:    movd %r8d, %xmm0
+; X64-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X64-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; X64-NEXT:    movd %edx, %xmm1
+; X64-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X64-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
+; X64-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-NEXT:    movd %ecx, %xmm0
+; X64-NEXT:    movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
+; X64-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X64-NEXT:    movd %r9d, %xmm2
+; X64-NEXT:    movd %esi, %xmm3
+; X64-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; X64-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3]
+; X64-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
+; X64-NEXT:    movdqa %xmm3, (%rdi)
+; X64-NEXT:    retq
   %tmp = insertelement <8 x i16> zeroinitializer, i16 %a0, i32 0
   %tmp2 = insertelement <8 x i16> %tmp, i16 %a1, i32 1
   %tmp4 = insertelement <8 x i16> %tmp2, i16 %a2, i32 2