summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
diff options
context:
space:
mode:
authorSimon Pilgrim <llvm-dev@redking.me.uk>2017-06-04 20:12:04 +0000
committerSimon Pilgrim <llvm-dev@redking.me.uk>2017-06-04 20:12:04 +0000
commit46dd55f1e1fdbc8cdf1d2ca535864e255ab84cc2 (patch)
treecdabd48e3d4e943603841aa60c1ce85b8e775343 /llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
parent59ed77b813a58e6f81fca3896f61e967a4f5f495 (diff)
downloadbcm5719-llvm-46dd55f1e1fdbc8cdf1d2ca535864e255ab84cc2.tar.gz
bcm5719-llvm-46dd55f1e1fdbc8cdf1d2ca535864e255ab84cc2.zip
[X86][SSE] Change BUILD_VECTOR interleaving ordering to improve coalescing/combine opportunities
We currently generate BUILD_VECTOR as a tree of UNPCKL shuffles of the same type: e.g. for v4f32: Step 1: unpcklps 0, 2 ==> X: <?, ?, 2, 0> : unpcklps 1, 3 ==> Y: <?, ?, 3, 1> Step 2: unpcklps X, Y ==> <3, 2, 1, 0> The issue is because we are not placing sequential vector elements together early enough, we fail to recognise many combinable patterns - consecutive scalar loads, extractions etc. Instead, this patch unpacks progressively larger sequential vector elements together: e.g. for v4f32: Step 1: unpcklps 0, 2 ==> X: <?, ?, 1, 0> : unpcklps 1, 3 ==> Y: <?, ?, 3, 2> Step 2: unpcklpd X, Y ==> <3, 2, 1, 0> This does mean that we are creating UNPCKL shuffle of different value types, but the relevant combines that benefit from this are quite capable of handling the additional BITCASTs that are now included in the shuffle tree. Differential Revision: https://reviews.llvm.org/D33864 llvm-svn: 304688
Diffstat (limited to 'llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll')
-rw-r--r--llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll159
1 files changed, 79 insertions, 80 deletions
diff --git a/llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
index 20387ccd6b7..ff5d624e604 100644
--- a/llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
+++ b/llvm/test/CodeGen/X86/sse2-intrinsics-fast-isel.ll
@@ -2076,7 +2076,7 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm2
; X32-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2087,8 +2087,8 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm1
; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2099,7 +2099,7 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm3
; X32-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2110,27 +2110,27 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_set_epi8:
; X64: # BB#0:
; X64-NEXT: movzbl %dil, %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: movzbl %sil, %eax
; X64-NEXT: movd %eax, %xmm1
; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X64-NEXT: movzbl %r8b, %eax
+; X64-NEXT: movzbl %dl, %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: movzbl %cl, %eax
; X64-NEXT: movd %eax, %xmm2
; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
-; X64-NEXT: movzbl %dl, %eax
+; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
+; X64-NEXT: movzbl %r8b, %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: movzbl %r9b, %eax
; X64-NEXT: movd %eax, %xmm3
; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
@@ -2138,20 +2138,20 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm1
; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
-; X64-NEXT: movzbl %sil, %eax
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm2
; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X64-NEXT: movzbl %r9b, %eax
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm3
; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
-; X64-NEXT: movzbl %cl, %eax
+; X64-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm2
@@ -2161,9 +2161,9 @@ define <2 x i64> @test_mm_set_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %a
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X64-NEXT: retq
%res0 = insertelement <16 x i8> undef, i8 %a15, i32 0
%res1 = insertelement <16 x i8> %res0, i8 %a14, i32 1
@@ -2206,11 +2206,11 @@ define <2 x i64> @test_mm_set_epi16(i16 %a0, i16 %a1, i16 %a2, i16 %a3, i16 %a4,
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
; X32-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1],xmm4[2],xmm3[2],xmm4[3],xmm3[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1]
; X32-NEXT: punpcklwd {{.*#+}} xmm6 = xmm6[0],xmm5[0],xmm6[1],xmm5[1],xmm6[2],xmm5[2],xmm6[3],xmm5[3]
; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm4[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_set_epi16:
@@ -2218,20 +2218,20 @@ define <2 x i64> @test_mm_set_epi16(i16 %a0, i16 %a1, i16 %a2, i16 %a3, i16 %a4,
; X64-NEXT: movw {{[0-9]+}}(%rsp), %r10w
; X64-NEXT: movw {{[0-9]+}}(%rsp), %ax
; X64-NEXT: movd %edi, %xmm0
-; X64-NEXT: movd %r8d, %xmm1
+; X64-NEXT: movd %esi, %xmm1
; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
; X64-NEXT: movd %edx, %xmm0
-; X64-NEXT: movd %eax, %xmm2
+; X64-NEXT: movd %ecx, %xmm2
; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; X64-NEXT: movd %esi, %xmm0
+; X64-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; X64-NEXT: movd %r8d, %xmm0
; X64-NEXT: movd %r9d, %xmm1
; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; X64-NEXT: movd %ecx, %xmm3
+; X64-NEXT: movd %eax, %xmm3
; X64-NEXT: movd %r10d, %xmm0
; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
; X64-NEXT: retq
%res0 = insertelement <8 x i16> undef, i16 %a7, i32 0
%res1 = insertelement <8 x i16> %res0, i16 %a6, i32 1
@@ -2254,18 +2254,18 @@ define <2 x i64> @test_mm_set_epi32(i32 %a0, i32 %a1, i32 %a2, i32 %a3) nounwind
; X32-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_set_epi32:
; X64: # BB#0:
; X64-NEXT: movd %edi, %xmm0
-; X64-NEXT: movd %edx, %xmm1
+; X64-NEXT: movd %esi, %xmm1
; X64-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; X64-NEXT: movd %esi, %xmm2
+; X64-NEXT: movd %edx, %xmm2
; X64-NEXT: movd %ecx, %xmm0
; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X64-NEXT: retq
%res0 = insertelement <4 x i32> undef, i32 %a3, i32 0
%res1 = insertelement <4 x i32> %res0, i32 %a2, i32 1
@@ -2282,11 +2282,11 @@ define <2 x i64> @test_mm_set_epi64x(i64 %a0, i64 %a1) nounwind {
; X32: # BB#0:
; X32-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; X32-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_set_epi64x:
@@ -2441,10 +2441,9 @@ define <2 x i64> @test_mm_set1_epi64x(i64 %a0) nounwind {
; X32-LABEL: test_mm_set1_epi64x:
; X32: # BB#0:
; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
; X32-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_set1_epi64x:
@@ -2486,7 +2485,7 @@ define <2 x i64> @test_mm_setr_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm2
; X32-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2497,8 +2496,8 @@ define <2 x i64> @test_mm_setr_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm1
; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2509,7 +2508,7 @@ define <2 x i64> @test_mm_setr_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm3
; X32-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
@@ -2520,9 +2519,9 @@ define <2 x i64> @test_mm_setr_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; X32-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_setr_epi8:
@@ -2534,46 +2533,46 @@ define <2 x i64> @test_mm_setr_epi8(i8 %a0, i8 %a1, i8 %a2, i8 %a3, i8 %a4, i8 %
; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl %cl, %eax
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm2
; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
+; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl %r9b, %eax
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm3
; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl %sil, %eax
+; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm1
; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
; X64-NEXT: movd %eax, %xmm2
; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: movzbl %r9b, %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl %dl, %eax
+; X64-NEXT: movzbl %r8b, %eax
; X64-NEXT: movd %eax, %xmm3
; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3]
+; X64-NEXT: movzbl %cl, %eax
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movzbl %r8b, %eax
+; X64-NEXT: movzbl %dl, %eax
; X64-NEXT: movd %eax, %xmm2
; X64-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
-; X64-NEXT: movzbl {{[0-9]+}}(%rsp), %eax
+; X64-NEXT: movzbl %sil, %eax
; X64-NEXT: movd %eax, %xmm4
; X64-NEXT: movzbl %dil, %eax
; X64-NEXT: movd %eax, %xmm0
; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
-; X64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X64-NEXT: retq
%res0 = insertelement <16 x i8> undef, i8 %a0 , i32 0
%res1 = insertelement <16 x i8> %res0, i8 %a1 , i32 1
@@ -2616,11 +2615,11 @@ define <2 x i64> @test_mm_setr_epi16(i16 %a0, i16 %a1, i16 %a2, i16 %a3, i16 %a4
; X32-NEXT: movd %eax, %xmm0
; X32-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
; X32-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1],xmm4[2],xmm3[2],xmm4[3],xmm3[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1]
; X32-NEXT: punpcklwd {{.*#+}} xmm6 = xmm6[0],xmm5[0],xmm6[1],xmm5[1],xmm6[2],xmm5[2],xmm6[3],xmm5[3]
; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm7[0],xmm0[1],xmm7[1],xmm0[2],xmm7[2],xmm0[3],xmm7[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3]
-; X32-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3]
+; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm4[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_setr_epi16:
@@ -2628,20 +2627,20 @@ define <2 x i64> @test_mm_setr_epi16(i16 %a0, i16 %a1, i16 %a2, i16 %a3, i16 %a4
; X64-NEXT: movw {{[0-9]+}}(%rsp), %ax
; X64-NEXT: movw {{[0-9]+}}(%rsp), %r10w
; X64-NEXT: movd %eax, %xmm0
-; X64-NEXT: movd %ecx, %xmm1
+; X64-NEXT: movd %r10d, %xmm1
; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
; X64-NEXT: movd %r9d, %xmm0
-; X64-NEXT: movd %esi, %xmm2
+; X64-NEXT: movd %r8d, %xmm2
; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
-; X64-NEXT: movd %r10d, %xmm0
+; X64-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; X64-NEXT: movd %ecx, %xmm0
; X64-NEXT: movd %edx, %xmm1
; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
-; X64-NEXT: movd %r8d, %xmm3
+; X64-NEXT: movd %esi, %xmm3
; X64-NEXT: movd %edi, %xmm0
; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
-; X64-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
+; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
; X64-NEXT: retq
%res0 = insertelement <8 x i16> undef, i16 %a0, i32 0
%res1 = insertelement <8 x i16> %res0, i16 %a1, i32 1
@@ -2664,18 +2663,18 @@ define <2 x i64> @test_mm_setr_epi32(i32 %a0, i32 %a1, i32 %a2, i32 %a3) nounwin
; X32-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_setr_epi32:
; X64: # BB#0:
; X64-NEXT: movd %ecx, %xmm0
-; X64-NEXT: movd %esi, %xmm1
+; X64-NEXT: movd %edx, %xmm1
; X64-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; X64-NEXT: movd %edx, %xmm2
+; X64-NEXT: movd %esi, %xmm2
; X64-NEXT: movd %edi, %xmm0
; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X64-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X64-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X64-NEXT: retq
%res0 = insertelement <4 x i32> undef, i32 %a0, i32 0
%res1 = insertelement <4 x i32> %res0, i32 %a1, i32 1
@@ -2692,11 +2691,11 @@ define <2 x i64> @test_mm_setr_epi64x(i64 %a0, i64 %a1) nounwind {
; X32: # BB#0:
; X32-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X32-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; X32-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
; X32-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero
-; X32-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
; X32-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X32-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; X32-NEXT: retl
;
; X64-LABEL: test_mm_setr_epi64x:
OpenPOWER on IntegriCloud