; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X32-SSE
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=X32-AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64-SSE
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=X64-AVX

define void @f(<4 x float> %A, i8* %B, <2 x double> %C, i32 %D, <2 x i64> %E, <4 x i32> %F, <8 x i16> %G, <16 x i8> %H, i64 %I) nounwind {
; X32-SSE-LABEL: f:
; X32-SSE:       # BB#0:
; X32-SSE-NEXT:    pushl %ebp
; X32-SSE-NEXT:    movl %esp, %ebp
; X32-SSE-NEXT:    andl $-16, %esp
; X32-SSE-NEXT:    subl $16, %esp
; X32-SSE-NEXT:    movl 72(%ebp), %eax
; X32-SSE-NEXT:    movl 76(%ebp), %ecx
; X32-SSE-NEXT:    movdqa 56(%ebp), %xmm3
; X32-SSE-NEXT:    movdqa 40(%ebp), %xmm4
; X32-SSE-NEXT:    movdqa 24(%ebp), %xmm5
; X32-SSE-NEXT:    movl 8(%ebp), %edx
; X32-SSE-NEXT:    addps {{\.LCPI.*}}, %xmm0
; X32-SSE-NEXT:    movntps %xmm0, (%edx)
; X32-SSE-NEXT:    paddq {{\.LCPI.*}}, %xmm2
; X32-SSE-NEXT:    movntdq %xmm2, (%edx)
; X32-SSE-NEXT:    addpd {{\.LCPI.*}}, %xmm1
; X32-SSE-NEXT:    movntpd %xmm1, (%edx)
; X32-SSE-NEXT:    paddd {{\.LCPI.*}}, %xmm5
; X32-SSE-NEXT:    movntdq %xmm5, (%edx)
; X32-SSE-NEXT:    paddw {{\.LCPI.*}}, %xmm4
; X32-SSE-NEXT:    movntdq %xmm4, (%edx)
; X32-SSE-NEXT:    paddb {{\.LCPI.*}}, %xmm3
; X32-SSE-NEXT:    movntdq %xmm3, (%edx)
; X32-SSE-NEXT:    movntil %ecx, 4(%edx)
; X32-SSE-NEXT:    movntil %eax, (%edx)
; X32-SSE-NEXT:    movl %ebp, %esp
; X32-SSE-NEXT:    popl %ebp
; X32-SSE-NEXT:    retl
;
; X32-AVX-LABEL: f:
; X32-AVX:       # BB#0:
; X32-AVX-NEXT:    pushl %ebp
; X32-AVX-NEXT:    movl %esp, %ebp
; X32-AVX-NEXT:    andl $-16, %esp
; X32-AVX-NEXT:    subl $16, %esp
; X32-AVX-NEXT:    movl 72(%ebp), %eax
; X32-AVX-NEXT:    movl 76(%ebp), %ecx
; X32-AVX-NEXT:    vmovdqa 56(%ebp), %xmm3
; X32-AVX-NEXT:    vmovdqa 40(%ebp), %xmm4
; X32-AVX-NEXT:    vmovdqa 24(%ebp), %xmm5
; X32-AVX-NEXT:    movl 8(%ebp), %edx
; X32-AVX-NEXT:    vaddps {{\.LCPI.*}}, %xmm0, %xmm0
; X32-AVX-NEXT:    vmovntps %xmm0, (%edx)
; X32-AVX-NEXT:    vpaddq {{\.LCPI.*}}, %xmm2, %xmm0
; X32-AVX-NEXT:    vmovntdq %xmm0, (%edx)
; X32-AVX-NEXT:    vaddpd {{\.LCPI.*}}, %xmm1, %xmm0
; X32-AVX-NEXT:    vmovntpd %xmm0, (%edx)
; X32-AVX-NEXT:    vpaddd {{\.LCPI.*}}, %xmm5, %xmm0
; X32-AVX-NEXT:    vmovntdq %xmm0, (%edx)
; X32-AVX-NEXT:    vpaddw {{\.LCPI.*}}, %xmm4, %xmm0
; X32-AVX-NEXT:    vmovntdq %xmm0, (%edx)
; X32-AVX-NEXT:    vpaddb {{\.LCPI.*}}, %xmm3, %xmm0
; X32-AVX-NEXT:    vmovntdq %xmm0, (%edx)
; X32-AVX-NEXT:    movntil %ecx, 4(%edx)
; X32-AVX-NEXT:    movntil %eax, (%edx)
; X32-AVX-NEXT:    movl %ebp, %esp
; X32-AVX-NEXT:    popl %ebp
; X32-AVX-NEXT:    retl
;
; X64-SSE-LABEL: f:
; X64-SSE:       # BB#0:
; X64-SSE-NEXT:    addps {{.*}}(%rip), %xmm0
; X64-SSE-NEXT:    movntps %xmm0, (%rdi)
; X64-SSE-NEXT:    paddq {{.*}}(%rip), %xmm2
; X64-SSE-NEXT:    movntdq %xmm2, (%rdi)
; X64-SSE-NEXT:    addpd {{.*}}(%rip), %xmm1
; X64-SSE-NEXT:    movntpd %xmm1, (%rdi)
; X64-SSE-NEXT:    paddd {{.*}}(%rip), %xmm3
; X64-SSE-NEXT:    movntdq %xmm3, (%rdi)
; X64-SSE-NEXT:    paddw {{.*}}(%rip), %xmm4
; X64-SSE-NEXT:    movntdq %xmm4, (%rdi)
; X64-SSE-NEXT:    paddb {{.*}}(%rip), %xmm5
; X64-SSE-NEXT:    movntdq %xmm5, (%rdi)
; X64-SSE-NEXT:    movntil %esi, (%rdi)
; X64-SSE-NEXT:    movntiq %rdx, (%rdi)
; X64-SSE-NEXT:    retq
;
; X64-AVX-LABEL: f:
; X64-AVX:       # BB#0:
; X64-AVX-NEXT:    vaddps {{.*}}(%rip), %xmm0, %xmm0
; X64-AVX-NEXT:    vmovntps %xmm0, (%rdi)
; X64-AVX-NEXT:    vpaddq {{.*}}(%rip), %xmm2, %xmm0
; X64-AVX-NEXT:    vmovntdq %xmm0, (%rdi)
; X64-AVX-NEXT:    vaddpd {{.*}}(%rip), %xmm1, %xmm0
; X64-AVX-NEXT:    vmovntpd %xmm0, (%rdi)
; X64-AVX-NEXT:    vpaddd {{.*}}(%rip), %xmm3, %xmm0
; X64-AVX-NEXT:    vmovntdq %xmm0, (%rdi)
; X64-AVX-NEXT:    vpaddw {{.*}}(%rip), %xmm4, %xmm0
; X64-AVX-NEXT:    vmovntdq %xmm0, (%rdi)
; X64-AVX-NEXT:    vpaddb {{.*}}(%rip), %xmm5, %xmm0
; X64-AVX-NEXT:    vmovntdq %xmm0, (%rdi)
; X64-AVX-NEXT:    movntil %esi, (%rdi)
; X64-AVX-NEXT:    movntiq %rdx, (%rdi)
; X64-AVX-NEXT:    retq
  %cast = bitcast i8* %B to <4 x float>*
  %A2 = fadd <4 x float> %A, <float 1.0, float 2.0, float 3.0, float 4.0>
  store <4 x float> %A2, <4 x float>* %cast, align 16, !nontemporal !0
  %cast1 = bitcast i8* %B to <2 x i64>*
  %E2 = add <2 x i64> %E, <i64 1, i64 2>
  store <2 x i64> %E2, <2 x i64>* %cast1, align 16, !nontemporal !0
  %cast2 = bitcast i8* %B to <2 x double>*
  %C2 = fadd <2 x double> %C, <double 1.0, double 2.0>
  store <2 x double> %C2, <2 x double>* %cast2, align 16, !nontemporal !0
  %cast3 = bitcast i8* %B to <4 x i32>*
  %F2 = add <4 x i32> %F, <i32 1, i32 2, i32 3, i32 4>
  store <4 x i32> %F2, <4 x i32>* %cast3, align 16, !nontemporal !0
  %cast4 = bitcast i8* %B to <8 x i16>*
  %G2 = add <8 x i16> %G, <i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 8>
  store <8 x i16> %G2, <8 x i16>* %cast4, align 16, !nontemporal !0
  %cast5 = bitcast i8* %B to <16 x i8>*
  %H2 = add <16 x i8> %H, <i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8>
  store <16 x i8> %H2, <16 x i8>* %cast5, align 16, !nontemporal !0
  %cast6 = bitcast i8* %B to i32*
  store i32 %D, i32* %cast6, align 1, !nontemporal !0
  %cast7 = bitcast i8* %B to i64*
  store i64 %I, i64* %cast7, align 1, !nontemporal !0
  ret void
}

!0 = !{i32 1}