Mercurial > hg > CbC > CbC_llvm

diff test/CodeGen/X86/scalar_widen_div.ll @ 134:3a76565eade5 LLVM5.0.1

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
update 5.0.1
author mir3636
date Sat, 17 Feb 2018 09:57:20 +0900
parents 803732b1fca8
children c2174574ed3a
line wrap: on
line diff
--- a/test/CodeGen/X86/scalar_widen_div.ll	Fri Feb 16 19:10:49 2018 +0900
+++ b/test/CodeGen/X86/scalar_widen_div.ll	Sat Feb 17 09:57:20 2018 +0900
@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-- -mattr=+sse4.2 |  FileCheck %s
 
 ; Verify when widening a divide/remainder operation, we only generate a
@@ -5,10 +6,29 @@
 
 ; CHECK: vectorDiv
 define void @vectorDiv (<2 x i32> addrspace(1)* %nsource, <2 x i32> addrspace(1)* %dsource, <2 x i32> addrspace(1)* %qdest) nounwind {
-; CHECK: idivq
-; CHECK: idivq
-; CHECK-NOT: idivl
-; CHECK: ret
+; CHECK-LABEL: vectorDiv:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    movq %rdx, %r8
+; CHECK-NEXT:    movq %rdi, -{{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rsi, -{{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, -{{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movslq -{{[0-9]+}}(%rsp), %rcx
+; CHECK-NEXT:    pmovsxdq (%rdi,%rcx,8), %xmm0
+; CHECK-NEXT:    pmovsxdq (%rsi,%rcx,8), %xmm1
+; CHECK-NEXT:    pextrq $1, %xmm0, %rax
+; CHECK-NEXT:    pextrq $1, %xmm1, %rsi
+; CHECK-NEXT:    cqto
+; CHECK-NEXT:    idivq %rsi
+; CHECK-NEXT:    movq %rax, %xmm2
+; CHECK-NEXT:    movq %xmm0, %rax
+; CHECK-NEXT:    movq %xmm1, %rsi
+; CHECK-NEXT:    cqto
+; CHECK-NEXT:    idivq %rsi
+; CHECK-NEXT:    movq %rax, %xmm0
+; CHECK-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
+; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; CHECK-NEXT:    movq %xmm0, (%r8,%rcx,8)
+; CHECK-NEXT:    retq
 entry:
   %nsource.addr = alloca <2 x i32> addrspace(1)*, align 4
   %dsource.addr = alloca <2 x i32> addrspace(1)*, align 4
@@ -35,117 +55,310 @@
 
 ; CHECK: test_char_div
 define <3 x i8> @test_char_div(<3 x i8> %num, <3 x i8> %div) {
-; CHECK: idivb
-; CHECK: idivb
-; CHECK: idivb
-; CHECK-NOT: idivb
-; CHECK: ret
+; CHECK-LABEL: test_char_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movl %edi, %eax
+; CHECK-NEXT:    cbtw
+; CHECK-NEXT:    idivb %cl
+; CHECK-NEXT:    movl %eax, %edi
+; CHECK-NEXT:    movl %esi, %eax
+; CHECK-NEXT:    cbtw
+; CHECK-NEXT:    idivb %r8b
+; CHECK-NEXT:    movl %eax, %esi
+; CHECK-NEXT:    movl %edx, %eax
+; CHECK-NEXT:    cbtw
+; CHECK-NEXT:    idivb %r9b
+; CHECK-NEXT:    movl %eax, %ecx
+; CHECK-NEXT:    movl %edi, %eax
+; CHECK-NEXT:    movl %esi, %edx
+; CHECK-NEXT:    retq
   %div.r = sdiv <3 x i8> %num, %div
   ret <3 x i8>  %div.r
 }
 
 ; CHECK: test_uchar_div
 define <3 x i8> @test_uchar_div(<3 x i8> %num, <3 x i8> %div) {
-; CHECK: divb
-; CHECK: divb
-; CHECK: divb
-; CHECK-NOT: divb
-; CHECK: ret
+; CHECK-LABEL: test_uchar_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movzbl %dil, %eax
+; CHECK-NEXT:    # kill: def $eax killed $eax def $ax
+; CHECK-NEXT:    divb %cl
+; CHECK-NEXT:    movl %eax, %edi
+; CHECK-NEXT:    movzbl %sil, %eax
+; CHECK-NEXT:    # kill: def $eax killed $eax def $ax
+; CHECK-NEXT:    divb %r8b
+; CHECK-NEXT:    movl %eax, %esi
+; CHECK-NEXT:    movzbl %dl, %eax
+; CHECK-NEXT:    # kill: def $eax killed $eax def $ax
+; CHECK-NEXT:    divb %r9b
+; CHECK-NEXT:    movl %eax, %ecx
+; CHECK-NEXT:    movl %edi, %eax
+; CHECK-NEXT:    movl %esi, %edx
+; CHECK-NEXT:    retq
   %div.r = udiv <3 x i8> %num, %div
   ret <3 x i8>  %div.r
 }
 
 ; CHECK: test_short_div
 define <5 x i16> @test_short_div(<5 x i16> %num, <5 x i16> %div) {
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK-NOT: idivw
-; CHECK: ret
+; CHECK-LABEL: test_short_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pextrw $4, %xmm0, %eax
+; CHECK-NEXT:    pextrw $4, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %eax, %r8d
+; CHECK-NEXT:    pextrw $3, %xmm0, %eax
+; CHECK-NEXT:    pextrw $3, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %eax, %r9d
+; CHECK-NEXT:    pextrw $2, %xmm0, %eax
+; CHECK-NEXT:    pextrw $2, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %eax, %edi
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %eax, %ecx
+; CHECK-NEXT:    pextrw $1, %xmm0, %eax
+; CHECK-NEXT:    pextrw $1, %xmm1, %esi
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %si
+; CHECK-NEXT:    # kill: def $ax killed $ax def $eax
+; CHECK-NEXT:    movd %ecx, %xmm0
+; CHECK-NEXT:    pinsrw $1, %eax, %xmm0
+; CHECK-NEXT:    pinsrw $2, %edi, %xmm0
+; CHECK-NEXT:    pinsrw $3, %r9d, %xmm0
+; CHECK-NEXT:    pinsrw $4, %r8d, %xmm0
+; CHECK-NEXT:    retq
   %div.r = sdiv <5 x i16> %num, %div
   ret <5 x i16>  %div.r
 }
 
 ; CHECK: test_ushort_div
 define <4 x i16> @test_ushort_div(<4 x i16> %num, <4 x i16> %div) {
-; CHECK: divl
-; CHECK: divl
-; CHECK: divl
-; CHECK: divl
-; CHECK-NOT: divl
-; CHECK: ret
+; CHECK-LABEL: test_ushort_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pxor %xmm2, %xmm2
+; CHECK-NEXT:    pblendw {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3],xmm1[4],xmm2[5],xmm1[6],xmm2[7]
+; CHECK-NEXT:    pblendw {{.*#+}} xmm0 = xmm0[0],xmm2[1],xmm0[2],xmm2[3],xmm0[4],xmm2[5],xmm0[6],xmm2[7]
+; CHECK-NEXT:    pextrd $1, %xmm0, %eax
+; CHECK-NEXT:    pextrd $1, %xmm1, %ecx
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %ecx
+; CHECK-NEXT:    movl %eax, %ecx
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %esi
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %esi
+; CHECK-NEXT:    movd %eax, %xmm2
+; CHECK-NEXT:    pinsrd $1, %ecx, %xmm2
+; CHECK-NEXT:    pextrd $2, %xmm0, %eax
+; CHECK-NEXT:    pextrd $2, %xmm1, %ecx
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %ecx
+; CHECK-NEXT:    pinsrd $2, %eax, %xmm2
+; CHECK-NEXT:    pextrd $3, %xmm0, %eax
+; CHECK-NEXT:    pextrd $3, %xmm1, %ecx
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %ecx
+; CHECK-NEXT:    pinsrd $3, %eax, %xmm2
+; CHECK-NEXT:    movdqa %xmm2, %xmm0
+; CHECK-NEXT:    retq
   %div.r = udiv <4 x i16> %num, %div
   ret <4 x i16>  %div.r
 }
 
 ; CHECK: test_uint_div
 define <3 x i32> @test_uint_div(<3 x i32> %num, <3 x i32> %div) {
-; CHECK: divl
-; CHECK: divl
-; CHECK: divl
-; CHECK-NOT: divl
-; CHECK: ret
+; CHECK-LABEL: test_uint_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pextrd $2, %xmm0, %eax
+; CHECK-NEXT:    pextrd $2, %xmm1, %ecx
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %ecx
+; CHECK-NEXT:    movl %eax, %ecx
+; CHECK-NEXT:    pextrd $1, %xmm0, %eax
+; CHECK-NEXT:    pextrd $1, %xmm1, %esi
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %esi
+; CHECK-NEXT:    movl %eax, %esi
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %edi
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divl %edi
+; CHECK-NEXT:    movd %eax, %xmm0
+; CHECK-NEXT:    pinsrd $1, %esi, %xmm0
+; CHECK-NEXT:    pinsrd $2, %ecx, %xmm0
+; CHECK-NEXT:    retq
   %div.r = udiv <3 x i32> %num, %div
   ret <3 x i32>  %div.r
 }
 
 ; CHECK: test_long_div
 define <3 x i64> @test_long_div(<3 x i64> %num, <3 x i64> %div) {
-; CHECK: idivq
-; CHECK: idivq
-; CHECK: idivq
-; CHECK-NOT: idivq
-; CHECK: ret
+; CHECK-LABEL: test_long_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movq %rdx, %r10
+; CHECK-NEXT:    movq %rdi, %rax
+; CHECK-NEXT:    cqto
+; CHECK-NEXT:    idivq %rcx
+; CHECK-NEXT:    movq %rax, %rcx
+; CHECK-NEXT:    movq %rsi, %rax
+; CHECK-NEXT:    cqto
+; CHECK-NEXT:    idivq %r8
+; CHECK-NEXT:    movq %rax, %rsi
+; CHECK-NEXT:    movq %r10, %rax
+; CHECK-NEXT:    cqto
+; CHECK-NEXT:    idivq %r9
+; CHECK-NEXT:    movq %rax, %rdi
+; CHECK-NEXT:    movq %rcx, %rax
+; CHECK-NEXT:    movq %rsi, %rdx
+; CHECK-NEXT:    movq %rdi, %rcx
+; CHECK-NEXT:    retq
   %div.r = sdiv <3 x i64> %num, %div
   ret <3 x i64>  %div.r
 }
 
 ; CHECK: test_ulong_div
 define <3 x i64> @test_ulong_div(<3 x i64> %num, <3 x i64> %div) {
-; CHECK: divq
-; CHECK: divq
-; CHECK: divq
-; CHECK-NOT: divq
-; CHECK: ret
+; CHECK-LABEL: test_ulong_div:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movq %rdx, %r10
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %rdi, %rax
+; CHECK-NEXT:    divq %rcx
+; CHECK-NEXT:    movq %rax, %rcx
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %rsi, %rax
+; CHECK-NEXT:    divq %r8
+; CHECK-NEXT:    movq %rax, %rsi
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %r10, %rax
+; CHECK-NEXT:    divq %r9
+; CHECK-NEXT:    movq %rax, %rdi
+; CHECK-NEXT:    movq %rcx, %rax
+; CHECK-NEXT:    movq %rsi, %rdx
+; CHECK-NEXT:    movq %rdi, %rcx
+; CHECK-NEXT:    retq
   %div.r = udiv <3 x i64> %num, %div
   ret <3 x i64>  %div.r
 }
 
 ; CHECK: test_char_rem
 define <4 x i8> @test_char_rem(<4 x i8> %num, <4 x i8> %rem) {
-; CHECK: idivl
-; CHECK: idivl
-; CHECK: idivl
-; CHECK: idivl
-; CHECK-NOT: idivl
-; CHECK: ret
+; CHECK-LABEL: test_char_rem:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pslld $24, %xmm1
+; CHECK-NEXT:    psrad $24, %xmm1
+; CHECK-NEXT:    pslld $24, %xmm0
+; CHECK-NEXT:    psrad $24, %xmm0
+; CHECK-NEXT:    pextrd $1, %xmm0, %eax
+; CHECK-NEXT:    pextrd $1, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    movl %edx, %ecx
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %esi
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %esi
+; CHECK-NEXT:    movd %edx, %xmm2
+; CHECK-NEXT:    pinsrd $1, %ecx, %xmm2
+; CHECK-NEXT:    pextrd $2, %xmm0, %eax
+; CHECK-NEXT:    pextrd $2, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    pinsrd $2, %edx, %xmm2
+; CHECK-NEXT:    pextrd $3, %xmm0, %eax
+; CHECK-NEXT:    pextrd $3, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    pinsrd $3, %edx, %xmm2
+; CHECK-NEXT:    movdqa %xmm2, %xmm0
+; CHECK-NEXT:    retq
   %rem.r = srem <4 x i8> %num, %rem
   ret <4 x i8>  %rem.r
 }
 
 ; CHECK: test_short_rem
 define <5 x i16> @test_short_rem(<5 x i16> %num, <5 x i16> %rem) {
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK: idivw
-; CHECK-NOT: idivw
-; CHECK: ret
+; CHECK-LABEL: test_short_rem:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pextrw $4, %xmm0, %eax
+; CHECK-NEXT:    pextrw $4, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %edx, %r8d
+; CHECK-NEXT:    pextrw $3, %xmm0, %eax
+; CHECK-NEXT:    pextrw $3, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %edx, %r9d
+; CHECK-NEXT:    pextrw $2, %xmm0, %eax
+; CHECK-NEXT:    pextrw $2, %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %edx, %edi
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %ecx
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %cx
+; CHECK-NEXT:    movl %edx, %ecx
+; CHECK-NEXT:    pextrw $1, %xmm0, %eax
+; CHECK-NEXT:    pextrw $1, %xmm1, %esi
+; CHECK-NEXT:    # kill: def $ax killed $ax killed $eax
+; CHECK-NEXT:    cwtd
+; CHECK-NEXT:    idivw %si
+; CHECK-NEXT:    # kill: def $dx killed $dx def $edx
+; CHECK-NEXT:    movd %ecx, %xmm0
+; CHECK-NEXT:    pinsrw $1, %edx, %xmm0
+; CHECK-NEXT:    pinsrw $2, %edi, %xmm0
+; CHECK-NEXT:    pinsrw $3, %r9d, %xmm0
+; CHECK-NEXT:    pinsrw $4, %r8d, %xmm0
+; CHECK-NEXT:    retq
   %rem.r = srem <5 x i16> %num, %rem
   ret <5 x i16>  %rem.r
 }
 
 ; CHECK: test_uint_rem
 define <4 x i32> @test_uint_rem(<4 x i32> %num, <4 x i32> %rem) {
-; CHECK: idivl
-; CHECK: idivl
-; CHECK: idivl
-; CHECK: idivl
-; CHECK-NOT: idivl
-; CHECK: ret
+; CHECK-LABEL: test_uint_rem:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pextrd $1, %xmm0, %eax
+; CHECK-NEXT:    pextrd $1, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    movl %edx, %ecx
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %esi
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %esi
+; CHECK-NEXT:    movd %edx, %xmm2
+; CHECK-NEXT:    pinsrd $1, %ecx, %xmm2
+; CHECK-NEXT:    pextrd $2, %xmm0, %eax
+; CHECK-NEXT:    pextrd $2, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    pinsrd $2, %edx, %xmm2
+; CHECK-NEXT:    pextrd $3, %xmm0, %eax
+; CHECK-NEXT:    pextrd $3, %xmm1, %ecx
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %ecx
+; CHECK-NEXT:    pinsrd $3, %edx, %xmm2
+; CHECK-NEXT:    movdqa %xmm2, %xmm0
+; CHECK-NEXT:    retq
   %rem.r = srem <4 x i32> %num, %rem
   ret <4 x i32>  %rem.r
 }
@@ -153,33 +366,84 @@
 
 ; CHECK: test_ulong_rem
 define <5 x i64> @test_ulong_rem(<5 x i64> %num, <5 x i64> %rem) {
-; CHECK: divq
-; CHECK: divq
-; CHECK: divq
-; CHECK: divq
-; CHECK: divq
-; CHECK-NOT: divq
-; CHECK: ret
+; CHECK-LABEL: test_ulong_rem:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movq %rdx, %rax
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    divq {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, %xmm0
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %rsi, %rax
+; CHECK-NEXT:    divq {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, %xmm1
+; CHECK-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %r8, %rax
+; CHECK-NEXT:    divq {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, %xmm0
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %rcx, %rax
+; CHECK-NEXT:    divq {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, %xmm2
+; CHECK-NEXT:    punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm0[0]
+; CHECK-NEXT:    xorl %edx, %edx
+; CHECK-NEXT:    movq %r9, %rax
+; CHECK-NEXT:    divq {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movq %rdx, 32(%rdi)
+; CHECK-NEXT:    movdqa %xmm2, 16(%rdi)
+; CHECK-NEXT:    movdqa %xmm1, (%rdi)
+; CHECK-NEXT:    movq %rdi, %rax
+; CHECK-NEXT:    retq
   %rem.r = urem <5 x i64> %num, %rem
   ret <5 x i64>  %rem.r
 }
 
 ; CHECK: test_int_div
 define void @test_int_div(<3 x i32>* %dest, <3 x i32>* %old, i32 %n) {
-; CHECK: idivl
-; CHECK: idivl
-; CHECK: idivl
-; CHECK-NOT: idivl
-; CHECK: ret
+; CHECK-LABEL: test_int_div:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    movl %edx, %r9d
+; CHECK-NEXT:    testl %edx, %edx
+; CHECK-NEXT:    jle .LBB12_3
+; CHECK-NEXT:  # %bb.1: # %bb.nph
+; CHECK-NEXT:    xorl %ecx, %ecx
+; CHECK-NEXT:    .p2align 4, 0x90
+; CHECK-NEXT:  .LBB12_2: # %for.body
+; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    movdqa (%rdi,%rcx), %xmm0
+; CHECK-NEXT:    movdqa (%rsi,%rcx), %xmm1
+; CHECK-NEXT:    pextrd $1, %xmm0, %eax
+; CHECK-NEXT:    pextrd $1, %xmm1, %r8d
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %r8d
+; CHECK-NEXT:    movl %eax, %r8d
+; CHECK-NEXT:    movd %xmm0, %eax
+; CHECK-NEXT:    movd %xmm1, %r10d
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %r10d
+; CHECK-NEXT:    movd %eax, %xmm2
+; CHECK-NEXT:    pinsrd $1, %r8d, %xmm2
+; CHECK-NEXT:    pextrd $2, %xmm0, %eax
+; CHECK-NEXT:    pextrd $2, %xmm1, %r8d
+; CHECK-NEXT:    cltd
+; CHECK-NEXT:    idivl %r8d
+; CHECK-NEXT:    pinsrd $2, %eax, %xmm2
+; CHECK-NEXT:    movl %eax, 8(%rdi,%rcx)
+; CHECK-NEXT:    movq %xmm2, (%rdi,%rcx)
+; CHECK-NEXT:    addq $16, %rcx
+; CHECK-NEXT:    decl %r9d
+; CHECK-NEXT:    jne .LBB12_2
+; CHECK-NEXT:  .LBB12_3: # %for.end
+; CHECK-NEXT:    retq
 entry:
   %cmp13 = icmp sgt i32 %n, 0
   br i1 %cmp13, label %bb.nph, label %for.end
 
-bb.nph:  
+bb.nph:
   br label %for.body
 
 for.body:
-  %i.014 = phi i32 [ 0, %bb.nph ], [ %inc, %for.body ] 
+  %i.014 = phi i32 [ 0, %bb.nph ], [ %inc, %for.body ]
   %arrayidx11 = getelementptr <3 x i32>, <3 x i32>* %dest, i32 %i.014
   %tmp4 = load <3 x i32>, <3 x i32>* %arrayidx11 ; <<3 x i32>> [#uses=1]
   %arrayidx7 = getelementptr inbounds <3 x i32>, <3 x i32>* %old, i32 %i.014
@@ -187,7 +451,7 @@
   %div = sdiv <3 x i32> %tmp4, %tmp8
   store <3 x i32> %div, <3 x i32>* %arrayidx11
   %inc = add nsw i32 %i.014, 1
-  %exitcond = icmp eq i32 %inc, %n 
+  %exitcond = icmp eq i32 %inc, %n
   br i1 %exitcond, label %for.end, label %for.body
 
 for.end:                                          ; preds = %for.body, %entry