test/CodeGen/X86/ushl_sat_vec.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s --check-prefix=X64
; RUN: llc < %s -mtriple=x86_64-linux -mattr=+avx2 | FileCheck %s --check-prefix=X64-AVX2
; RUN: llc < %s -mtriple=i686 -mattr=cmov | FileCheck %s --check-prefix=X86

declare <2 x i64> @llvm.ushl.sat.v2i64(<2 x i64>, <2 x i64>)
declare <4 x i32> @llvm.ushl.sat.v4i32(<4 x i32>, <4 x i32>)
declare <8 x i16> @llvm.ushl.sat.v8i16(<8 x i16>, <8 x i16>)
declare <16 x i8> @llvm.ushl.sat.v16i8(<16 x i8>, <16 x i8>)

define <2 x i64> @vec_v2i64(<2 x i64> %x, <2 x i64> %y) nounwind {
; X64-LABEL: vec_v2i64:
; X64:       # %bb.0:
; X64-NEXT:    movdqa %xmm0, %xmm2
; X64-NEXT:    psllq %xmm1, %xmm2
; X64-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[2,3,2,3]
; X64-NEXT:    movdqa %xmm0, %xmm4
; X64-NEXT:    psllq %xmm3, %xmm4
; X64-NEXT:    movdqa %xmm4, %xmm5
; X64-NEXT:    movsd {{.*#+}} xmm4 = xmm2[0],xmm4[1]
; X64-NEXT:    psrlq %xmm1, %xmm2
; X64-NEXT:    psrlq %xmm3, %xmm5
; X64-NEXT:    movsd {{.*#+}} xmm5 = xmm2[0],xmm5[1]
; X64-NEXT:    pcmpeqd %xmm5, %xmm0
; X64-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[1,0,3,2]
; X64-NEXT:    pand %xmm1, %xmm0
; X64-NEXT:    pcmpeqd %xmm1, %xmm1
; X64-NEXT:    pxor %xmm1, %xmm0
; X64-NEXT:    por %xmm4, %xmm0
; X64-NEXT:    retq
;
; X64-AVX2-LABEL: vec_v2i64:
; X64-AVX2:       # %bb.0:
; X64-AVX2-NEXT:    vpsllvq %xmm1, %xmm0, %xmm2
; X64-AVX2-NEXT:    vpsrlvq %xmm1, %xmm2, %xmm1
; X64-AVX2-NEXT:    vpcmpeqq %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
; X64-AVX2-NEXT:    vblendvpd %xmm0, %xmm2, %xmm1, %xmm0
; X64-AVX2-NEXT:    retq
;
; X86-LABEL: vec_v2i64:
; X86:       # %bb.0:
; X86-NEXT:    pushl %ebp
; X86-NEXT:    pushl %ebx
; X86-NEXT:    pushl %edi
; X86-NEXT:    pushl %esi
; X86-NEXT:    subl $16, %esp
; X86-NEXT:    movb {{[0-9]+}}(%esp), %ch
; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movl %esi, %eax
; X86-NEXT:    movb %ch, %cl
; X86-NEXT:    shll %cl, %eax
; X86-NEXT:    shldl %cl, %esi, %edx
; X86-NEXT:    xorl %ebx, %ebx
; X86-NEXT:    testb $32, %ch
; X86-NEXT:    cmovnel %eax, %edx
; X86-NEXT:    cmovnel %ebx, %eax
; X86-NEXT:    movl %eax, (%esp) # 4-byte Spill
; X86-NEXT:    movl %edx, %eax
; X86-NEXT:    movl %edx, %ebp
; X86-NEXT:    movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    shrl %cl, %eax
; X86-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    testb $32, %ch
; X86-NEXT:    cmovnel %ebx, %eax
; X86-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %esi
; X86-NEXT:    shll %cl, %esi
; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    shldl %cl, %eax, %edx
; X86-NEXT:    testb $32, %cl
; X86-NEXT:    cmovnel %esi, %edx
; X86-NEXT:    cmovnel %ebx, %esi
; X86-NEXT:    movl %edx, %edi
; X86-NEXT:    shrl %cl, %edi
; X86-NEXT:    testb $32, %cl
; X86-NEXT:    cmovel %edi, %ebx
; X86-NEXT:    movl (%esp), %eax # 4-byte Reload
; X86-NEXT:    movb %ch, %cl
; X86-NEXT:    shrdl %cl, %ebp, %eax
; X86-NEXT:    testb $32, %ch
; X86-NEXT:    cmovnel {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload
; X86-NEXT:    movl %esi, %ebp
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    shrdl %cl, %edx, %ebp
; X86-NEXT:    testb $32, {{[0-9]+}}(%esp)
; X86-NEXT:    cmovnel %edi, %ebp
; X86-NEXT:    xorl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    xorl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    orl %eax, %ecx
; X86-NEXT:    movl $-1, %ecx
; X86-NEXT:    movl (%esp), %edi # 4-byte Reload
; X86-NEXT:    cmovnel %ecx, %edi
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
; X86-NEXT:    cmovnel %ecx, %eax
; X86-NEXT:    xorl {{[0-9]+}}(%esp), %ebp
; X86-NEXT:    xorl {{[0-9]+}}(%esp), %ebx
; X86-NEXT:    orl %ebp, %ebx
; X86-NEXT:    cmovnel %ecx, %esi
; X86-NEXT:    cmovnel %ecx, %edx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl %edx, 12(%ecx)
; X86-NEXT:    movl %esi, 8(%ecx)
; X86-NEXT:    movl %eax, 4(%ecx)
; X86-NEXT:    movl %edi, (%ecx)
; X86-NEXT:    movl %ecx, %eax
; X86-NEXT:    addl $16, %esp
; X86-NEXT:    popl %esi
; X86-NEXT:    popl %edi
; X86-NEXT:    popl %ebx
; X86-NEXT:    popl %ebp
; X86-NEXT:    retl $4
  %tmp = call <2 x i64> @llvm.ushl.sat.v2i64(<2 x i64> %x, <2 x i64> %y)
  ret <2 x i64> %tmp
}

define <4 x i32> @vec_v4i32(<4 x i32> %x, <4 x i32> %y) nounwind {
; X64-LABEL: vec_v4i32:
; X64:       # %bb.0:
; X64-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[2,3,2,3]
; X64-NEXT:    pshuflw {{.*#+}} xmm3 = xmm1[2,3,3,3,4,5,6,7]
; X64-NEXT:    pshuflw {{.*#+}} xmm4 = xmm1[0,1,1,1,4,5,6,7]
; X64-NEXT:    pslld $23, %xmm1
; X64-NEXT:    paddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
; X64-NEXT:    cvttps2dq %xmm1, %xmm1
; X64-NEXT:    movdqa %xmm0, %xmm5
; X64-NEXT:    pmuludq %xmm1, %xmm5
; X64-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,2,2,3]
; X64-NEXT:    pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3]
; X64-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; X64-NEXT:    pmuludq %xmm7, %xmm1
; X64-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; X64-NEXT:    punpckldq {{.*#+}} xmm6 = xmm6[0],xmm1[0],xmm6[1],xmm1[1]
; X64-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[2,3,3,3,4,5,6,7]
; X64-NEXT:    movdqa %xmm6, %xmm7
; X64-NEXT:    psrld %xmm1, %xmm7
; X64-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[0,1,1,1,4,5,6,7]
; X64-NEXT:    movdqa %xmm5, %xmm2
; X64-NEXT:    psrld %xmm1, %xmm2
; X64-NEXT:    punpckhqdq {{.*#+}} xmm2 = xmm2[1],xmm7[1]
; X64-NEXT:    movdqa %xmm6, %xmm1
; X64-NEXT:    psrld %xmm3, %xmm1
; X64-NEXT:    psrld %xmm4, %xmm5
; X64-NEXT:    punpcklqdq {{.*#+}} xmm5 = xmm5[0],xmm1[0]
; X64-NEXT:    shufps {{.*#+}} xmm5 = xmm5[0,3],xmm2[0,3]
; X64-NEXT:    pcmpeqd %xmm5, %xmm0
; X64-NEXT:    pcmpeqd %xmm1, %xmm1
; X64-NEXT:    pxor %xmm1, %xmm0
; X64-NEXT:    por %xmm6, %xmm0
; X64-NEXT:    retq
;
; X64-AVX2-LABEL: vec_v4i32:
; X64-AVX2:       # %bb.0:
; X64-AVX2-NEXT:    vpsllvd %xmm1, %xmm0, %xmm2
; X64-AVX2-NEXT:    vpsrlvd %xmm1, %xmm2, %xmm1
; X64-AVX2-NEXT:    vpcmpeqd %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
; X64-AVX2-NEXT:    vblendvps %xmm0, %xmm2, %xmm1, %xmm0
; X64-AVX2-NEXT:    retq
;
; X86-LABEL: vec_v4i32:
; X86:       # %bb.0:
; X86-NEXT:    pushl %ebp
; X86-NEXT:    pushl %ebx
; X86-NEXT:    pushl %edi
; X86-NEXT:    pushl %esi
; X86-NEXT:    movb {{[0-9]+}}(%esp), %ch
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %edi
; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebx
; X86-NEXT:    movl %ebx, %esi
; X86-NEXT:    shll %cl, %esi
; X86-NEXT:    movl %esi, %ebp
; X86-NEXT:    shrl %cl, %ebp
; X86-NEXT:    cmpl %ebp, %ebx
; X86-NEXT:    movl $-1, %ebx
; X86-NEXT:    cmovnel %ebx, %esi
; X86-NEXT:    movl %edi, %eax
; X86-NEXT:    movb %dl, %cl
; X86-NEXT:    shll %cl, %eax
; X86-NEXT:    movl %eax, %ebp
; X86-NEXT:    shrl %cl, %ebp
; X86-NEXT:    cmpl %ebp, %edi
; X86-NEXT:    cmovnel %ebx, %eax
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebp
; X86-NEXT:    movl %ebp, %edx
; X86-NEXT:    movb %ch, %cl
; X86-NEXT:    shll %cl, %edx
; X86-NEXT:    movl %edx, %edi
; X86-NEXT:    shrl %cl, %edi
; X86-NEXT:    cmpl %edi, %ebp
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebp
; X86-NEXT:    cmovnel %ebx, %edx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    shll %cl, %ebp
; X86-NEXT:    movl %ebp, %edi
; X86-NEXT:    shrl %cl, %edi
; X86-NEXT:    cmpl %edi, {{[0-9]+}}(%esp)
; X86-NEXT:    cmovnel %ebx, %ebp
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl %ebp, 12(%ecx)
; X86-NEXT:    movl %edx, 8(%ecx)
; X86-NEXT:    movl %eax, 4(%ecx)
; X86-NEXT:    movl %esi, (%ecx)
; X86-NEXT:    movl %ecx, %eax
; X86-NEXT:    popl %esi
; X86-NEXT:    popl %edi
; X86-NEXT:    popl %ebx
; X86-NEXT:    popl %ebp
; X86-NEXT:    retl $4
  %tmp = call <4 x i32> @llvm.ushl.sat.v4i32(<4 x i32> %x, <4 x i32> %y)
  ret <4 x i32> %tmp
}

define <8 x i16> @vec_v8i16(<8 x i16> %x, <8 x i16> %y) nounwind {
; X64-LABEL: vec_v8i16:
; X64:       # %bb.0:
; X64-NEXT:    movdqa %xmm1, %xmm2
; X64-NEXT:    punpckhwd {{.*#+}} xmm2 = xmm2[4,4,5,5,6,6,7,7]
; X64-NEXT:    pslld $23, %xmm2
; X64-NEXT:    movdqa {{.*#+}} xmm3 = [1065353216,1065353216,1065353216,1065353216]
; X64-NEXT:    paddd %xmm3, %xmm2
; X64-NEXT:    cvttps2dq %xmm2, %xmm4
; X64-NEXT:    pslld $16, %xmm4
; X64-NEXT:    psrad $16, %xmm4
; X64-NEXT:    movdqa %xmm1, %xmm2
; X64-NEXT:    punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
; X64-NEXT:    pslld $23, %xmm2
; X64-NEXT:    paddd %xmm3, %xmm2
; X64-NEXT:    cvttps2dq %xmm2, %xmm2
; X64-NEXT:    pslld $16, %xmm2
; X64-NEXT:    psrad $16, %xmm2
; X64-NEXT:    packssdw %xmm4, %xmm2
; X64-NEXT:    pmullw %xmm0, %xmm2
; X64-NEXT:    psllw $12, %xmm1
; X64-NEXT:    movdqa %xmm1, %xmm3
; X64-NEXT:    psraw $15, %xmm3
; X64-NEXT:    movdqa %xmm2, %xmm4
; X64-NEXT:    psrlw $8, %xmm4
; X64-NEXT:    pand %xmm3, %xmm4
; X64-NEXT:    pandn %xmm2, %xmm3
; X64-NEXT:    por %xmm4, %xmm3
; X64-NEXT:    paddw %xmm1, %xmm1
; X64-NEXT:    movdqa %xmm1, %xmm4
; X64-NEXT:    psraw $15, %xmm4
; X64-NEXT:    movdqa %xmm4, %xmm5
; X64-NEXT:    pandn %xmm3, %xmm5
; X64-NEXT:    psrlw $4, %xmm3
; X64-NEXT:    pand %xmm4, %xmm3
; X64-NEXT:    por %xmm5, %xmm3
; X64-NEXT:    paddw %xmm1, %xmm1
; X64-NEXT:    movdqa %xmm1, %xmm4
; X64-NEXT:    psraw $15, %xmm4
; X64-NEXT:    movdqa %xmm4, %xmm5
; X64-NEXT:    pandn %xmm3, %xmm5
; X64-NEXT:    psrlw $2, %xmm3
; X64-NEXT:    pand %xmm4, %xmm3
; X64-NEXT:    por %xmm5, %xmm3
; X64-NEXT:    paddw %xmm1, %xmm1
; X64-NEXT:    psraw $15, %xmm1
; X64-NEXT:    movdqa %xmm1, %xmm4
; X64-NEXT:    pandn %xmm3, %xmm4
; X64-NEXT:    psrlw $1, %xmm3
; X64-NEXT:    pand %xmm1, %xmm3
; X64-NEXT:    por %xmm4, %xmm3
; X64-NEXT:    pcmpeqw %xmm3, %xmm0
; X64-NEXT:    pcmpeqd %xmm1, %xmm1
; X64-NEXT:    pxor %xmm1, %xmm0
; X64-NEXT:    por %xmm2, %xmm0
; X64-NEXT:    retq
;
; X64-AVX2-LABEL: vec_v8i16:
; X64-AVX2:       # %bb.0:
; X64-AVX2-NEXT:    vpmovzxwd {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; X64-AVX2-NEXT:    vpmovzxwd {{.*#+}} ymm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; X64-AVX2-NEXT:    vpsllvd %ymm1, %ymm2, %ymm2
; X64-AVX2-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[0,1,4,5,8,9,12,13,u,u,u,u,u,u,u,u,16,17,20,21,24,25,28,29,u,u,u,u,u,u,u,u]
; X64-AVX2-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,2,2,3]
; X64-AVX2-NEXT:    vpmovzxwd {{.*#+}} ymm3 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
; X64-AVX2-NEXT:    vpsrlvd %ymm1, %ymm3, %ymm1
; X64-AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
; X64-AVX2-NEXT:    vpackusdw %xmm3, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpcmpeqw %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpxor %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT:    vpor %xmm2, %xmm0, %xmm0
; X64-AVX2-NEXT:    vzeroupper
; X64-AVX2-NEXT:    retq
;
; X86-LABEL: vec_v8i16:
; X86:       # %bb.0:
; X86-NEXT:    pushl %ebp
; X86-NEXT:    pushl %ebx
; X86-NEXT:    pushl %edi
; X86-NEXT:    pushl %esi
; X86-NEXT:    subl $12, %esp
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ebp
; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl %eax, %ebx
; X86-NEXT:    shll %cl, %ebx
; X86-NEXT:    movzwl %bx, %edi
; X86-NEXT:    shrl %cl, %edi
; X86-NEXT:    cmpw %di, %ax
; X86-NEXT:    movl $65535, %eax # imm = 0xFFFF
; X86-NEXT:    cmovnel %eax, %ebx
; X86-NEXT:    movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movl %esi, %eax
; X86-NEXT:    movl %edx, %ecx
; X86-NEXT:    shll %cl, %eax
; X86-NEXT:    movzwl %ax, %edi
; X86-NEXT:    shrl %cl, %edi
; X86-NEXT:    cmpw %di, %si
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl $65535, %esi # imm = 0xFFFF
; X86-NEXT:    cmovnel %esi, %eax
; X86-NEXT:    movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movl %ebp, %eax
; X86-NEXT:    shll %cl, %eax
; X86-NEXT:    movzwl %ax, %edx
; X86-NEXT:    shrl %cl, %edx
; X86-NEXT:    cmpw %dx, %bp
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    cmovnel %esi, %eax
; X86-NEXT:    movl %eax, (%esp) # 4-byte Spill
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %ebp
; X86-NEXT:    shll %cl, %ebp
; X86-NEXT:    movzwl %bp, %edx
; X86-NEXT:    shrl %cl, %edx
; X86-NEXT:    cmpw %dx, %ax
; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    cmovnel %esi, %ebp
; X86-NEXT:    movl %edx, %ebx
; X86-NEXT:    shll %cl, %ebx
; X86-NEXT:    movzwl %bx, %esi
; X86-NEXT:    shrl %cl, %esi
; X86-NEXT:    cmpw %si, %dx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movl $65535, %esi # imm = 0xFFFF
; X86-NEXT:    cmovnel %esi, %ebx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl %edx, %edi
; X86-NEXT:    shll %cl, %edi
; X86-NEXT:    movzwl %di, %eax
; X86-NEXT:    shrl %cl, %eax
; X86-NEXT:    cmpw %ax, %dx
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
; X86-NEXT:    movl %esi, %edx
; X86-NEXT:    shll %cl, %edx
; X86-NEXT:    movzwl %dx, %eax
; X86-NEXT:    shrl %cl, %eax
; X86-NEXT:    cmpw %ax, %si
; X86-NEXT:    movl $65535, %eax # imm = 0xFFFF
; X86-NEXT:    cmovnel %eax, %edx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    shll %cl, %eax
; X86-NEXT:    movzwl %ax, %esi
; X86-NEXT:    shrl %cl, %esi
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    cmpw %si, %cx
; X86-NEXT:    movl $65535, %ecx # imm = 0xFFFF
; X86-NEXT:    cmovnel %ecx, %eax
; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movw %ax, 14(%ecx)
; X86-NEXT:    movw %dx, 12(%ecx)
; X86-NEXT:    movw %di, 10(%ecx)
; X86-NEXT:    movw %bx, 8(%ecx)
; X86-NEXT:    movw %bp, 6(%ecx)
; X86-NEXT:    movl (%esp), %eax # 4-byte Reload
; X86-NEXT:    movw %ax, 4(%ecx)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
; X86-NEXT:    movw %ax, 2(%ecx)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload
; X86-NEXT:    movw %ax, (%ecx)
; X86-NEXT:    movl %ecx, %eax
; X86-NEXT:    addl $12, %esp
; X86-NEXT:    popl %esi
; X86-NEXT:    popl %edi
; X86-NEXT:    popl %ebx
; X86-NEXT:    popl %ebp
; X86-NEXT:    retl $4
  %tmp = call <8 x i16> @llvm.ushl.sat.v8i16(<8 x i16> %x, <8 x i16> %y)
  ret <8 x i16> %tmp
}

define <16 x i8> @vec_v16i8(<16 x i8> %x, <16 x i8> %y) nounwind {
; X64-LABEL: vec_v16i8:
; X64:       # %bb.0:
; X64-NEXT:    psllw $5, %xmm1
; X64-NEXT:    pxor %xmm3, %xmm3
; X64-NEXT:    pxor %xmm4, %xmm4
; X64-NEXT:    pcmpgtb %xmm1, %xmm4
; X64-NEXT:    movdqa %xmm4, %xmm5
; X64-NEXT:    pandn %xmm0, %xmm5
; X64-NEXT:    movdqa %xmm0, %xmm2
; X64-NEXT:    psllw $4, %xmm2
; X64-NEXT:    pand %xmm4, %xmm2
; X64-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2
; X64-NEXT:    por %xmm5, %xmm2
; X64-NEXT:    paddb %xmm1, %xmm1
; X64-NEXT:    pxor %xmm5, %xmm5
; X64-NEXT:    pcmpgtb %xmm1, %xmm5
; X64-NEXT:    movdqa %xmm5, %xmm6
; X64-NEXT:    pandn %xmm2, %xmm6
; X64-NEXT:    psllw $2, %xmm2
; X64-NEXT:    pand %xmm5, %xmm2
; X64-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2
; X64-NEXT:    por %xmm6, %xmm2
; X64-NEXT:    paddb %xmm1, %xmm1
; X64-NEXT:    pcmpgtb %xmm1, %xmm3
; X64-NEXT:    movdqa %xmm3, %xmm1
; X64-NEXT:    pandn %xmm2, %xmm1
; X64-NEXT:    paddb %xmm2, %xmm2
; X64-NEXT:    pand %xmm3, %xmm2
; X64-NEXT:    por %xmm1, %xmm2
; X64-NEXT:    movdqa %xmm2, %xmm1
; X64-NEXT:    psrlw $4, %xmm1
; X64-NEXT:    pand %xmm4, %xmm1
; X64-NEXT:    pandn %xmm2, %xmm4
; X64-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
; X64-NEXT:    por %xmm4, %xmm1
; X64-NEXT:    movdqa %xmm5, %xmm4
; X64-NEXT:    pandn %xmm1, %xmm4
; X64-NEXT:    psrlw $2, %xmm1
; X64-NEXT:    pand %xmm5, %xmm1
; X64-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
; X64-NEXT:    por %xmm4, %xmm1
; X64-NEXT:    movdqa %xmm3, %xmm4
; X64-NEXT:    pandn %xmm1, %xmm4
; X64-NEXT:    psrlw $1, %xmm1
; X64-NEXT:    pand %xmm3, %xmm1
; X64-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
; X64-NEXT:    por %xmm4, %xmm1
; X64-NEXT:    pcmpeqb %xmm1, %xmm0
; X64-NEXT:    pcmpeqd %xmm1, %xmm1
; X64-NEXT:    pxor %xmm1, %xmm0
; X64-NEXT:    por %xmm2, %xmm0
; X64-NEXT:    retq
;
; X64-AVX2-LABEL: vec_v16i8:
; X64-AVX2:       # %bb.0:
; X64-AVX2-NEXT:    vpsllw $5, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpsllw $4, %xmm0, %xmm2
; X64-AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2, %xmm2
; X64-AVX2-NEXT:    vpblendvb %xmm1, %xmm2, %xmm0, %xmm2
; X64-AVX2-NEXT:    vpsllw $2, %xmm2, %xmm3
; X64-AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
; X64-AVX2-NEXT:    vpaddb %xmm1, %xmm1, %xmm4
; X64-AVX2-NEXT:    vpblendvb %xmm4, %xmm3, %xmm2, %xmm2
; X64-AVX2-NEXT:    vpaddb %xmm2, %xmm2, %xmm3
; X64-AVX2-NEXT:    vpaddb %xmm4, %xmm4, %xmm5
; X64-AVX2-NEXT:    vpblendvb %xmm5, %xmm3, %xmm2, %xmm2
; X64-AVX2-NEXT:    vpsrlw $4, %xmm2, %xmm3
; X64-AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
; X64-AVX2-NEXT:    vpblendvb %xmm1, %xmm3, %xmm2, %xmm1
; X64-AVX2-NEXT:    vpsrlw $2, %xmm1, %xmm3
; X64-AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
; X64-AVX2-NEXT:    vpblendvb %xmm4, %xmm3, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpsrlw $1, %xmm1, %xmm3
; X64-AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm3, %xmm3
; X64-AVX2-NEXT:    vpblendvb %xmm5, %xmm3, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpcmpeqb %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT:    vpcmpeqd %xmm1, %xmm1, %xmm1
; X64-AVX2-NEXT:    vpblendvb %xmm0, %xmm2, %xmm1, %xmm0
; X64-AVX2-NEXT:    retq
;
; X86-LABEL: vec_v16i8:
; X86:       # %bb.0:
; X86-NEXT:    pushl %ebp
; X86-NEXT:    pushl %ebx
; X86-NEXT:    pushl %edi
; X86-NEXT:    pushl %esi
; X86-NEXT:    subl $48, %esp
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    movb {{[0-9]+}}(%esp), %ch
; X86-NEXT:    movb {{[0-9]+}}(%esp), %ah
; X86-NEXT:    movb {{[0-9]+}}(%esp), %dh
; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ebx
; X86-NEXT:    movb %bl, %bh
; X86-NEXT:    shlb %cl, %bh
; X86-NEXT:    movzbl %bh, %edi
; X86-NEXT:    shrb %cl, %bh
; X86-NEXT:    cmpb %bh, %bl
; X86-NEXT:    movl $255, %esi
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movb %dh, %bl
; X86-NEXT:    movb %ah, %cl
; X86-NEXT:    shlb %cl, %bl
; X86-NEXT:    movzbl %bl, %edi
; X86-NEXT:    shrb %cl, %bl
; X86-NEXT:    cmpb %bl, %dh
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movb %ch, %ah
; X86-NEXT:    movb %dl, %cl
; X86-NEXT:    shlb %cl, %ah
; X86-NEXT:    movzbl %ah, %edi
; X86-NEXT:    shrb %cl, %ah
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %edx
; X86-NEXT:    cmpb %ah, %ch
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movb %dl, %ah
; X86-NEXT:    movl %eax, %ecx
; X86-NEXT:    shlb %cl, %ah
; X86-NEXT:    movzbl %ah, %edi
; X86-NEXT:    shrb %cl, %ah
; X86-NEXT:    cmpb %ah, %dl
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movl %edi, (%esp) # 4-byte Spill
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %ebp
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %ebp
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %edi
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %edi
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl %eax, %edx
; X86-NEXT:    shlb %cl, %dl
; X86-NEXT:    movzbl %dl, %ebx
; X86-NEXT:    shrb %cl, %dl
; X86-NEXT:    cmpb %dl, %al
; X86-NEXT:    cmovnel %esi, %ebx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movb %al, %ah
; X86-NEXT:    shlb %cl, %ah
; X86-NEXT:    movzbl %ah, %edx
; X86-NEXT:    shrb %cl, %ah
; X86-NEXT:    cmpb %ah, %al
; X86-NEXT:    cmovnel %esi, %edx
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movb %dl, 15(%eax)
; X86-NEXT:    movb %bl, 14(%eax)
; X86-NEXT:    movl %edi, %ecx
; X86-NEXT:    movb %cl, 13(%eax)
; X86-NEXT:    movl %ebp, %ecx
; X86-NEXT:    movb %cl, 12(%eax)
; X86-NEXT:    movl (%esp), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 11(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 10(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 9(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 8(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 7(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 6(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 5(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 4(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 3(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 2(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, 1(%eax)
; X86-NEXT:    movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload
; X86-NEXT:    movb %cl, (%eax)
; X86-NEXT:    addl $48, %esp
; X86-NEXT:    popl %esi
; X86-NEXT:    popl %edi
; X86-NEXT:    popl %ebx
; X86-NEXT:    popl %ebp
; X86-NEXT:    retl $4
  %tmp = call <16 x i8> @llvm.ushl.sat.v16i8(<16 x i8> %x, <16 x i8> %y)
  ret <16 x i8> %tmp
}