forked from OSchip/llvm-project
3310 lines
138 KiB
LLVM
3310 lines
138 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=CHECK,SSE,SSE41
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2ORLATER,AVX2
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX,AVX2ORLATER,AVX512,AVX512F
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=CHECK,AVX,AVX2ORLATER,AVX512,AVX512BW
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+xop | FileCheck %s --check-prefixes=CHECK,AVX,XOP
|
|
|
|
; fold (sdiv x, 1) -> x
|
|
define i32 @combine_sdiv_by_one(i32 %x) {
|
|
; CHECK-LABEL: combine_sdiv_by_one:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movl %edi, %eax
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv i32 %x, 1
|
|
ret i32 %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_one(<4 x i32> %x) {
|
|
; CHECK-LABEL: combine_vec_sdiv_by_one:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 1, i32 1, i32 1, i32 1>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; fold (sdiv x, -1) -> 0 - x
|
|
define i32 @combine_sdiv_by_negone(i32 %x) {
|
|
; CHECK-LABEL: combine_sdiv_by_negone:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movl %edi, %eax
|
|
; CHECK-NEXT: negl %eax
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv i32 %x, -1
|
|
ret i32 %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_negone(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_by_negone:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: pxor %xmm1, %xmm1
|
|
; SSE-NEXT: psubd %xmm0, %xmm1
|
|
; SSE-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_by_negone:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX-NEXT: vpsubd %xmm0, %xmm1, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 -1, i32 -1, i32 -1, i32 -1>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; fold (sdiv x, INT_MIN) -> select((icmp eq x, INT_MIN), 1, 0)
|
|
define i32 @combine_sdiv_by_minsigned(i32 %x) {
|
|
; CHECK-LABEL: combine_sdiv_by_minsigned:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xorl %eax, %eax
|
|
; CHECK-NEXT: cmpl $-2147483648, %edi # imm = 0x80000000
|
|
; CHECK-NEXT: sete %al
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv i32 %x, -2147483648
|
|
ret i32 %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_minsigned(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: pcmpeqd {{.*}}(%rip), %xmm0
|
|
; SSE-NEXT: psrld $31, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpcmpeqd {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm1 = [2147483648,2147483648,2147483648,2147483648]
|
|
; AVX2-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpbroadcastd {{.*#+}} xmm1 = [2147483648,2147483648,2147483648,2147483648]
|
|
; AVX512F-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpsrld $31, %xmm0, %xmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpcmpeqd {{.*}}(%rip){1to4}, %xmm0, %k1
|
|
; AVX512BW-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z}
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_minsigned:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpcomeqd {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpsrld $31, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; TODO fold (sdiv 0, x) -> 0
|
|
define i32 @combine_sdiv_zero(i32 %x) {
|
|
; CHECK-LABEL: combine_sdiv_zero:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xorl %eax, %eax
|
|
; CHECK-NEXT: xorl %edx, %edx
|
|
; CHECK-NEXT: idivl %edi
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv i32 0, %x
|
|
ret i32 %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_zero(<4 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_zero:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3]
|
|
; SSE2-NEXT: movd %xmm1, %ecx
|
|
; SSE2-NEXT: xorl %eax, %eax
|
|
; SSE2-NEXT: xorl %edx, %edx
|
|
; SSE2-NEXT: idivl %ecx
|
|
; SSE2-NEXT: movd %eax, %xmm1
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,0,1]
|
|
; SSE2-NEXT: movd %xmm2, %ecx
|
|
; SSE2-NEXT: xorl %eax, %eax
|
|
; SSE2-NEXT: xorl %edx, %edx
|
|
; SSE2-NEXT: idivl %ecx
|
|
; SSE2-NEXT: movd %eax, %xmm2
|
|
; SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
|
|
; SSE2-NEXT: movd %xmm0, %ecx
|
|
; SSE2-NEXT: xorl %eax, %eax
|
|
; SSE2-NEXT: xorl %edx, %edx
|
|
; SSE2-NEXT: idivl %ecx
|
|
; SSE2-NEXT: movd %eax, %xmm1
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
|
|
; SSE2-NEXT: movd %xmm0, %ecx
|
|
; SSE2-NEXT: xorl %eax, %eax
|
|
; SSE2-NEXT: xorl %edx, %edx
|
|
; SSE2-NEXT: idivl %ecx
|
|
; SSE2-NEXT: movd %eax, %xmm0
|
|
; SSE2-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
|
|
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_zero:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: pextrd $1, %xmm0, %ecx
|
|
; SSE41-NEXT: xorl %eax, %eax
|
|
; SSE41-NEXT: xorl %edx, %edx
|
|
; SSE41-NEXT: idivl %ecx
|
|
; SSE41-NEXT: movl %eax, %ecx
|
|
; SSE41-NEXT: movd %xmm0, %esi
|
|
; SSE41-NEXT: xorl %eax, %eax
|
|
; SSE41-NEXT: xorl %edx, %edx
|
|
; SSE41-NEXT: idivl %esi
|
|
; SSE41-NEXT: movd %eax, %xmm1
|
|
; SSE41-NEXT: pinsrd $1, %ecx, %xmm1
|
|
; SSE41-NEXT: pextrd $2, %xmm0, %ecx
|
|
; SSE41-NEXT: xorl %eax, %eax
|
|
; SSE41-NEXT: xorl %edx, %edx
|
|
; SSE41-NEXT: idivl %ecx
|
|
; SSE41-NEXT: pinsrd $2, %eax, %xmm1
|
|
; SSE41-NEXT: pextrd $3, %xmm0, %ecx
|
|
; SSE41-NEXT: xorl %eax, %eax
|
|
; SSE41-NEXT: xorl %edx, %edx
|
|
; SSE41-NEXT: idivl %ecx
|
|
; SSE41-NEXT: pinsrd $3, %eax, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_zero:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpextrd $1, %xmm0, %ecx
|
|
; AVX-NEXT: xorl %eax, %eax
|
|
; AVX-NEXT: xorl %edx, %edx
|
|
; AVX-NEXT: idivl %ecx
|
|
; AVX-NEXT: movl %eax, %ecx
|
|
; AVX-NEXT: vmovd %xmm0, %esi
|
|
; AVX-NEXT: xorl %eax, %eax
|
|
; AVX-NEXT: xorl %edx, %edx
|
|
; AVX-NEXT: idivl %esi
|
|
; AVX-NEXT: vmovd %eax, %xmm1
|
|
; AVX-NEXT: vpinsrd $1, %ecx, %xmm1, %xmm1
|
|
; AVX-NEXT: vpextrd $2, %xmm0, %ecx
|
|
; AVX-NEXT: xorl %eax, %eax
|
|
; AVX-NEXT: xorl %edx, %edx
|
|
; AVX-NEXT: idivl %ecx
|
|
; AVX-NEXT: vpinsrd $2, %eax, %xmm1, %xmm1
|
|
; AVX-NEXT: vpextrd $3, %xmm0, %ecx
|
|
; AVX-NEXT: xorl %eax, %eax
|
|
; AVX-NEXT: xorl %edx, %edx
|
|
; AVX-NEXT: idivl %ecx
|
|
; AVX-NEXT: vpinsrd $3, %eax, %xmm1, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = sdiv <4 x i32> zeroinitializer, %x
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; fold (sdiv x, x) -> 1
|
|
define i32 @combine_sdiv_dupe(i32 %x) {
|
|
; CHECK-LABEL: combine_sdiv_dupe:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movl $1, %eax
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv i32 %x, %x
|
|
ret i32 %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_dupe(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_dupe:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: movaps {{.*#+}} xmm0 = [1,1,1,1]
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_dupe:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vmovaps {{.*#+}} xmm0 = [1,1,1,1]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_dupe:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vbroadcastss {{.*#+}} xmm0 = [1,1,1,1]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_dupe:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vmovaps {{.*#+}} xmm0 = [1,1,1,1]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, %x
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; fold (sdiv x, y) -> (udiv x, y) iff x and y are positive
|
|
define <4 x i32> @combine_vec_sdiv_by_pos0(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_by_pos0:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE-NEXT: psrld $2, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_by_pos0:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX-NEXT: vpsrld $2, %xmm0, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255>
|
|
%2 = sdiv <4 x i32> %1, <i32 4, i32 4, i32 4, i32 4>
|
|
ret <4 x i32> %2
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pos1(<4 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pos1:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrld $4, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrld $3, %xmm2
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1],xmm1[1]
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrld $2, %xmm1
|
|
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3],xmm2[0,3]
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pos1:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrld $3, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: psrld $4, %xmm0
|
|
; SSE41-NEXT: psrld $2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pos1:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpsrld $4, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrld $2, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $3, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_by_pos1:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2ORLATER-NEXT: vpsrlvd {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pos1:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpshld {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255>
|
|
%2 = sdiv <4 x i32> %1, <i32 1, i32 4, i32 8, i32 16>
|
|
ret <4 x i32> %2
|
|
}
|
|
|
|
; fold (sdiv x, (1 << c)) -> x >>u c
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2a(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_by_pow2a:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE-NEXT: psrad $31, %xmm1
|
|
; SSE-NEXT: psrld $30, %xmm1
|
|
; SSE-NEXT: paddd %xmm0, %xmm1
|
|
; SSE-NEXT: psrad $2, %xmm1
|
|
; SSE-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_by_pow2a:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX-NEXT: vpsrld $30, %xmm1, %xmm1
|
|
; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0
|
|
; AVX-NEXT: vpsrad $2, %xmm0, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 4, i32 4, i32 4, i32 4>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2a_neg(<4 x i32> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_by_pow2a_neg:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE-NEXT: psrad $31, %xmm1
|
|
; SSE-NEXT: psrld $30, %xmm1
|
|
; SSE-NEXT: paddd %xmm0, %xmm1
|
|
; SSE-NEXT: psrad $2, %xmm1
|
|
; SSE-NEXT: pxor %xmm0, %xmm0
|
|
; SSE-NEXT: psubd %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_by_pow2a_neg:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX-NEXT: vpsrld $30, %xmm1, %xmm1
|
|
; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0
|
|
; AVX-NEXT: vpsrad $2, %xmm0, %xmm0
|
|
; AVX-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX-NEXT: vpsubd %xmm0, %xmm1, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 -4, i32 -4, i32 -4, i32 -4>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <16 x i8> @combine_vec_sdiv_by_pow2b_v16i8(<16 x i8> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: pxor %xmm1, %xmm1
|
|
; SSE2-NEXT: pxor %xmm2, %xmm2
|
|
; SSE2-NEXT: pcmpgtb %xmm0, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm1[8],xmm3[9],xmm1[9],xmm3[10],xmm1[10],xmm3[11],xmm1[11],xmm3[12],xmm1[12],xmm3[13],xmm1[13],xmm3[14],xmm1[14],xmm3[15],xmm1[15]
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [1,4,2,16,8,32,64,2]
|
|
; SSE2-NEXT: pmullw %xmm4, %xmm3
|
|
; SSE2-NEXT: psrlw $8, %xmm3
|
|
; SSE2-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]
|
|
; SSE2-NEXT: pmullw %xmm4, %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: packuswb %xmm3, %xmm2
|
|
; SSE2-NEXT: paddb %xmm0, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm2[8],xmm1[9],xmm2[9],xmm1[10],xmm2[10],xmm1[11],xmm2[11],xmm1[12],xmm2[12],xmm1[13],xmm2[13],xmm1[14],xmm2[14],xmm1[15],xmm2[15]
|
|
; SSE2-NEXT: psraw $8, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [256,64,128,16,32,8,4,128]
|
|
; SSE2-NEXT: pmullw %xmm3, %xmm1
|
|
; SSE2-NEXT: psrlw $8, %xmm1
|
|
; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: pmullw %xmm3, %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: packuswb %xmm1, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; SSE2-NEXT: pand %xmm1, %xmm2
|
|
; SSE2-NEXT: pandn %xmm0, %xmm1
|
|
; SSE2-NEXT: por %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: pxor %xmm0, %xmm0
|
|
; SSE41-NEXT: pxor %xmm3, %xmm3
|
|
; SSE41-NEXT: pcmpgtb %xmm1, %xmm3
|
|
; SSE41-NEXT: pmovzxbw {{.*#+}} xmm2 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
|
|
; SSE41-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm0[8],xmm3[9],xmm0[9],xmm3[10],xmm0[10],xmm3[11],xmm0[11],xmm3[12],xmm0[12],xmm3[13],xmm0[13],xmm3[14],xmm0[14],xmm3[15],xmm0[15]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm0 = [1,4,2,16,8,32,64,2]
|
|
; SSE41-NEXT: pmullw %xmm0, %xmm3
|
|
; SSE41-NEXT: psrlw $8, %xmm3
|
|
; SSE41-NEXT: pmullw %xmm0, %xmm2
|
|
; SSE41-NEXT: psrlw $8, %xmm2
|
|
; SSE41-NEXT: packuswb %xmm3, %xmm2
|
|
; SSE41-NEXT: paddb %xmm1, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE41-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm2[8],xmm0[9],xmm2[9],xmm0[10],xmm2[10],xmm0[11],xmm2[11],xmm0[12],xmm2[12],xmm0[13],xmm2[13],xmm0[14],xmm2[14],xmm0[15],xmm2[15]
|
|
; SSE41-NEXT: psraw $8, %xmm0
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [256,64,128,16,32,8,4,128]
|
|
; SSE41-NEXT: pmullw %xmm3, %xmm0
|
|
; SSE41-NEXT: psrlw $8, %xmm0
|
|
; SSE41-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; SSE41-NEXT: psraw $8, %xmm2
|
|
; SSE41-NEXT: pmullw %xmm3, %xmm2
|
|
; SSE41-NEXT: psrlw $8, %xmm2
|
|
; SSE41-NEXT: packuswb %xmm0, %xmm2
|
|
; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,4,2,16,8,32,64,2]
|
|
; AVX1-NEXT: vpmullw %xmm3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
|
|
; AVX1-NEXT: vpmullw %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
|
|
; AVX1-NEXT: vpaddb %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
|
|
; AVX1-NEXT: vpsraw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [256,64,128,16,32,8,4,128]
|
|
; AVX1-NEXT: vpmullw %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; AVX1-NEXT: vpsraw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpmullw %xmm3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; AVX1-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpmovzxbw {{.*#+}} 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,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
|
; AVX2-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpaddb %xmm1, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpmovsxbw %xmm1, %ymm1
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
|
; AVX2-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; AVX2-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: vzeroupper
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpmovzxbd {{.*#+}} zmm1 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero,xmm1[2],zero,zero,zero,xmm1[3],zero,zero,zero,xmm1[4],zero,zero,zero,xmm1[5],zero,zero,zero,xmm1[6],zero,zero,zero,xmm1[7],zero,zero,zero,xmm1[8],zero,zero,zero,xmm1[9],zero,zero,zero,xmm1[10],zero,zero,zero,xmm1[11],zero,zero,zero,xmm1[12],zero,zero,zero,xmm1[13],zero,zero,zero,xmm1[14],zero,zero,zero,xmm1[15],zero,zero,zero
|
|
; AVX512F-NEXT: vpsrlvd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: vpmovdb %zmm1, %xmm1
|
|
; AVX512F-NEXT: vpaddb %xmm1, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmovsxbd %xmm1, %zmm1
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: vpmovdb %zmm1, %xmm1
|
|
; AVX512F-NEXT: vmovdqa {{.*#+}} xmm2 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; AVX512F-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpmovzxbw {{.*#+}} 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,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
|
|
; AVX512BW-NEXT: vpsrlvw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpmovwb %ymm1, %xmm1
|
|
; AVX512BW-NEXT: vpaddb %xmm1, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpmovsxbw %xmm1, %ymm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpmovwb %ymm1, %xmm1
|
|
; AVX512BW-NEXT: movw $257, %ax # imm = 0x101
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqu8 %xmm0, %xmm1 {%k1}
|
|
; AVX512BW-NEXT: vmovdqa %xmm1, %xmm0
|
|
; AVX512BW-NEXT: vzeroupper
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v16i8:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vpshlb {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddb %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshab {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm2 = [0,255,255,255,255,255,255,255,0,255,255,255,255,255,255,255]
|
|
; XOP-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <16 x i8> %x, <i8 1, i8 4, i8 2, i8 16, i8 8, i8 32, i8 64, i8 2, i8 1, i8 4, i8 2, i8 16, i8 8, i8 32, i8 64, i8 2>
|
|
ret <16 x i8> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_by_pow2b_v8i16(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psraw $15, %xmm1
|
|
; SSE2-NEXT: pmulhuw {{.*}}(%rip), %xmm1
|
|
; SSE2-NEXT: paddw %xmm0, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,65535,0,65535,0,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $4, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,65535,65535,0,65535,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: pand %xmm1, %xmm3
|
|
; SSE2-NEXT: psraw $2, %xmm2
|
|
; SSE2-NEXT: pandn %xmm2, %xmm1
|
|
; SSE2-NEXT: por %xmm3, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,0,65535,0,0,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $1, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,65535,65535,65535,65535,65535,65535,65535]
|
|
; SSE2-NEXT: pand %xmm1, %xmm2
|
|
; SSE2-NEXT: pandn %xmm0, %xmm1
|
|
; SSE2-NEXT: por %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psraw $15, %xmm1
|
|
; SSE41-NEXT: pmulhuw {{.*}}(%rip), %xmm1
|
|
; SSE41-NEXT: paddw %xmm0, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psraw $1, %xmm2
|
|
; SSE41-NEXT: pmulhw {{.*}}(%rip), %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2],xmm1[3,4,5,6],xmm2[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsraw $15, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpmulhuw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsraw $1, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2],xmm1[3,4,5,6],xmm2[7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsraw $15, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpmulhuw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpsraw $1, %xmm1, %xmm2
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2],xmm1[3,4,5,6],xmm2[7]
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpsraw $15, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmulhuw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpaddw %xmm1, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmovsxwd %xmm1, %ymm1
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512F-NEXT: vpmovdw %zmm1, %ymm1
|
|
; AVX512F-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsrlvw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v8i16:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpsraw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshlw {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3,4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <16 x i16> @combine_vec_sdiv_by_pow2b_v16i16(<16 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psraw $15, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [1,4,2,16,8,32,64,2]
|
|
; SSE2-NEXT: pmulhuw %xmm8, %xmm0
|
|
; SSE2-NEXT: paddw %xmm3, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [65535,65535,65535,0,65535,0,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: pand %xmm4, %xmm2
|
|
; SSE2-NEXT: psraw $4, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm6
|
|
; SSE2-NEXT: pandn %xmm0, %xmm6
|
|
; SSE2-NEXT: por %xmm2, %xmm6
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm5 = [65535,0,65535,65535,0,65535,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm6, %xmm0
|
|
; SSE2-NEXT: pand %xmm5, %xmm0
|
|
; SSE2-NEXT: psraw $2, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm2
|
|
; SSE2-NEXT: pandn %xmm6, %xmm2
|
|
; SSE2-NEXT: por %xmm0, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm7 = [65535,65535,0,65535,0,0,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE2-NEXT: pand %xmm7, %xmm0
|
|
; SSE2-NEXT: psraw $1, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm7, %xmm6
|
|
; SSE2-NEXT: pandn %xmm2, %xmm6
|
|
; SSE2-NEXT: por %xmm0, %xmm6
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,65535,65535,65535,65535,65535,65535,65535]
|
|
; SSE2-NEXT: pand %xmm2, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE2-NEXT: pandn %xmm3, %xmm0
|
|
; SSE2-NEXT: por %xmm6, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psraw $15, %xmm3
|
|
; SSE2-NEXT: pmulhuw %xmm8, %xmm3
|
|
; SSE2-NEXT: paddw %xmm1, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm6
|
|
; SSE2-NEXT: pand %xmm4, %xmm6
|
|
; SSE2-NEXT: psraw $4, %xmm3
|
|
; SSE2-NEXT: pandn %xmm3, %xmm4
|
|
; SSE2-NEXT: por %xmm6, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm3
|
|
; SSE2-NEXT: pand %xmm5, %xmm3
|
|
; SSE2-NEXT: psraw $2, %xmm4
|
|
; SSE2-NEXT: pandn %xmm4, %xmm5
|
|
; SSE2-NEXT: por %xmm3, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm3
|
|
; SSE2-NEXT: pand %xmm7, %xmm3
|
|
; SSE2-NEXT: psraw $1, %xmm5
|
|
; SSE2-NEXT: pandn %xmm5, %xmm7
|
|
; SSE2-NEXT: por %xmm3, %xmm7
|
|
; SSE2-NEXT: pand %xmm2, %xmm7
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm7, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psraw $15, %xmm2
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,4,2,16,8,32,64,2]
|
|
; SSE41-NEXT: pmulhuw %xmm4, %xmm2
|
|
; SSE41-NEXT: paddw %xmm0, %xmm2
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm5 = <u,16384,32768,4096,8192,2048,1024,32768>
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE41-NEXT: pmulhw %xmm5, %xmm3
|
|
; SSE41-NEXT: psraw $1, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm3[0,1],xmm2[2],xmm3[3,4,5,6],xmm2[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm0[0],xmm2[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE41-NEXT: psraw $15, %xmm3
|
|
; SSE41-NEXT: pmulhuw %xmm4, %xmm3
|
|
; SSE41-NEXT: paddw %xmm1, %xmm3
|
|
; SSE41-NEXT: pmulhw %xmm3, %xmm5
|
|
; SSE41-NEXT: psraw $1, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm5[0,1],xmm3[2],xmm5[3,4,5,6],xmm3[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm1[0],xmm3[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm1
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
|
|
; AVX1-NEXT: vpsraw $15, %xmm1, %xmm2
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,4,2,16,8,32,64,2]
|
|
; AVX1-NEXT: vpmulhuw %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpaddw %xmm2, %xmm1, %xmm1
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = <u,16384,32768,4096,8192,2048,1024,32768>
|
|
; AVX1-NEXT: vpmulhw %xmm2, %xmm1, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm1[0],xmm4[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm4[0,1],xmm1[2],xmm4[3,4,5,6],xmm1[7]
|
|
; AVX1-NEXT: vpsraw $15, %xmm0, %xmm4
|
|
; AVX1-NEXT: vpmulhuw %xmm3, %xmm4, %xmm3
|
|
; AVX1-NEXT: vpaddw %xmm3, %xmm0, %xmm3
|
|
; AVX1-NEXT: vpmulhw %xmm2, %xmm3, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0],xmm2[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2],xmm2[3,4,5,6],xmm3[7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; AVX1-NEXT: vbroadcastf128 {{.*#+}} ymm2 = [0,65535,65535,65535,65535,65535,65535,65535,0,65535,65535,65535,65535,65535,65535,65535]
|
|
; AVX1-NEXT: # ymm2 = mem[0,1,0,1]
|
|
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
|
|
; AVX1-NEXT: vandnps %ymm0, %ymm2, %ymm0
|
|
; AVX1-NEXT: vorps %ymm0, %ymm1, %ymm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsraw $15, %ymm0, %ymm1
|
|
; AVX2-NEXT: vpmulhuw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpaddw %ymm1, %ymm0, %ymm1
|
|
; AVX2-NEXT: vpsraw $1, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm1 = ymm1[0,1],ymm2[2],ymm1[3,4,5,6],ymm2[7],ymm1[8,9],ymm2[10],ymm1[11,12,13,14],ymm2[15]
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7],ymm0[8],ymm1[9,10,11,12,13,14,15]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpsraw $15, %ymm0, %ymm1
|
|
; AVX512F-NEXT: vpmulhuw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512F-NEXT: vpaddw %ymm1, %ymm0, %ymm1
|
|
; AVX512F-NEXT: vpmovsxwd %ymm1, %zmm1
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: vpmovdw %zmm1, %ymm1
|
|
; AVX512F-NEXT: vpblendw {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7],ymm0[8],ymm1[9,10,11,12,13,14,15]
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraw $15, %ymm0, %ymm1
|
|
; AVX512BW-NEXT: vpsrlvw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpaddw %ymm1, %ymm0, %ymm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpblendw {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3,4,5,6,7],ymm0[8],ymm1[9,10,11,12,13,14,15]
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v16i16:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm1
|
|
; XOP-NEXT: vpsraw $15, %xmm1, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [65520,65522,65521,65524,65523,65525,65526,65521]
|
|
; XOP-NEXT: vpshlw %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpaddw %xmm2, %xmm1, %xmm1
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm2 = [0,65534,65535,65532,65533,65531,65530,65535]
|
|
; XOP-NEXT: vpshaw %xmm2, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsraw $15, %xmm0, %xmm4
|
|
; XOP-NEXT: vpshlw %xmm3, %xmm4, %xmm3
|
|
; XOP-NEXT: vpaddw %xmm3, %xmm0, %xmm3
|
|
; XOP-NEXT: vpshaw %xmm2, %xmm3, %xmm2
|
|
; XOP-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; XOP-NEXT: vbroadcastf128 {{.*#+}} ymm2 = [0,65535,65535,65535,65535,65535,65535,65535,0,65535,65535,65535,65535,65535,65535,65535]
|
|
; XOP-NEXT: # ymm2 = mem[0,1,0,1]
|
|
; XOP-NEXT: vpcmov %ymm2, %ymm0, %ymm1, %ymm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <16 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2>
|
|
ret <16 x i16> %1
|
|
}
|
|
|
|
define <32 x i16> @combine_vec_sdiv_by_pow2b_v32i16(<32 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm8
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psraw $15, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [1,4,2,16,8,32,64,2]
|
|
; SSE2-NEXT: pmulhuw %xmm9, %xmm0
|
|
; SSE2-NEXT: paddw %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm11 = [65535,65535,65535,0,65535,0,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE2-NEXT: pand %xmm11, %xmm4
|
|
; SSE2-NEXT: psraw $4, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm11, %xmm5
|
|
; SSE2-NEXT: pandn %xmm0, %xmm5
|
|
; SSE2-NEXT: por %xmm4, %xmm5
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm7 = [65535,0,65535,65535,0,65535,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm0
|
|
; SSE2-NEXT: pand %xmm7, %xmm0
|
|
; SSE2-NEXT: psraw $2, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm7, %xmm4
|
|
; SSE2-NEXT: pandn %xmm5, %xmm4
|
|
; SSE2-NEXT: por %xmm0, %xmm4
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [65535,65535,0,65535,0,0,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm0
|
|
; SSE2-NEXT: pand %xmm10, %xmm0
|
|
; SSE2-NEXT: psraw $1, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm10, %xmm5
|
|
; SSE2-NEXT: pandn %xmm4, %xmm5
|
|
; SSE2-NEXT: por %xmm0, %xmm5
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm12 = [0,65535,65535,65535,65535,65535,65535,65535]
|
|
; SSE2-NEXT: pand %xmm12, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm12, %xmm0
|
|
; SSE2-NEXT: pandn %xmm1, %xmm0
|
|
; SSE2-NEXT: por %xmm5, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm8, %xmm1
|
|
; SSE2-NEXT: psraw $15, %xmm1
|
|
; SSE2-NEXT: pmulhuw %xmm9, %xmm1
|
|
; SSE2-NEXT: paddw %xmm8, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE2-NEXT: pand %xmm11, %xmm5
|
|
; SSE2-NEXT: psraw $4, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm11, %xmm6
|
|
; SSE2-NEXT: pandn %xmm1, %xmm6
|
|
; SSE2-NEXT: por %xmm5, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm6, %xmm1
|
|
; SSE2-NEXT: pand %xmm7, %xmm1
|
|
; SSE2-NEXT: psraw $2, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm7, %xmm5
|
|
; SSE2-NEXT: pandn %xmm6, %xmm5
|
|
; SSE2-NEXT: por %xmm1, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm1
|
|
; SSE2-NEXT: pand %xmm10, %xmm1
|
|
; SSE2-NEXT: psraw $1, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm10, %xmm6
|
|
; SSE2-NEXT: pandn %xmm5, %xmm6
|
|
; SSE2-NEXT: por %xmm1, %xmm6
|
|
; SSE2-NEXT: pand %xmm12, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm12, %xmm1
|
|
; SSE2-NEXT: pandn %xmm8, %xmm1
|
|
; SSE2-NEXT: por %xmm6, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm5
|
|
; SSE2-NEXT: psraw $15, %xmm5
|
|
; SSE2-NEXT: pmulhuw %xmm9, %xmm5
|
|
; SSE2-NEXT: paddw %xmm2, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm6
|
|
; SSE2-NEXT: pand %xmm11, %xmm6
|
|
; SSE2-NEXT: psraw $4, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm11, %xmm4
|
|
; SSE2-NEXT: pandn %xmm5, %xmm4
|
|
; SSE2-NEXT: por %xmm6, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE2-NEXT: pand %xmm7, %xmm5
|
|
; SSE2-NEXT: psraw $2, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm7, %xmm6
|
|
; SSE2-NEXT: pandn %xmm4, %xmm6
|
|
; SSE2-NEXT: por %xmm5, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm6, %xmm4
|
|
; SSE2-NEXT: pand %xmm10, %xmm4
|
|
; SSE2-NEXT: psraw $1, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm10, %xmm5
|
|
; SSE2-NEXT: pandn %xmm6, %xmm5
|
|
; SSE2-NEXT: por %xmm4, %xmm5
|
|
; SSE2-NEXT: pand %xmm12, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm12, %xmm8
|
|
; SSE2-NEXT: pandn %xmm2, %xmm8
|
|
; SSE2-NEXT: por %xmm5, %xmm8
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm2
|
|
; SSE2-NEXT: psraw $15, %xmm2
|
|
; SSE2-NEXT: pmulhuw %xmm9, %xmm2
|
|
; SSE2-NEXT: paddw %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE2-NEXT: pand %xmm11, %xmm4
|
|
; SSE2-NEXT: psraw $4, %xmm2
|
|
; SSE2-NEXT: pandn %xmm2, %xmm11
|
|
; SSE2-NEXT: por %xmm4, %xmm11
|
|
; SSE2-NEXT: movdqa %xmm11, %xmm2
|
|
; SSE2-NEXT: pand %xmm7, %xmm2
|
|
; SSE2-NEXT: psraw $2, %xmm11
|
|
; SSE2-NEXT: pandn %xmm11, %xmm7
|
|
; SSE2-NEXT: por %xmm2, %xmm7
|
|
; SSE2-NEXT: movdqa %xmm7, %xmm2
|
|
; SSE2-NEXT: pand %xmm10, %xmm2
|
|
; SSE2-NEXT: psraw $1, %xmm7
|
|
; SSE2-NEXT: pandn %xmm7, %xmm10
|
|
; SSE2-NEXT: por %xmm2, %xmm10
|
|
; SSE2-NEXT: pand %xmm12, %xmm10
|
|
; SSE2-NEXT: pandn %xmm3, %xmm12
|
|
; SSE2-NEXT: por %xmm10, %xmm12
|
|
; SSE2-NEXT: movdqa %xmm8, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm12, %xmm3
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psraw $15, %xmm0
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm7 = [1,4,2,16,8,32,64,2]
|
|
; SSE41-NEXT: pmulhuw %xmm7, %xmm0
|
|
; SSE41-NEXT: paddw %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm6 = <u,16384,32768,4096,8192,2048,1024,32768>
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE41-NEXT: pmulhw %xmm6, %xmm5
|
|
; SSE41-NEXT: psraw $1, %xmm0
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm5[0,1],xmm0[2],xmm5[3,4,5,6],xmm0[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm1
|
|
; SSE41-NEXT: psraw $15, %xmm1
|
|
; SSE41-NEXT: pmulhuw %xmm7, %xmm1
|
|
; SSE41-NEXT: paddw %xmm4, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE41-NEXT: pmulhw %xmm6, %xmm5
|
|
; SSE41-NEXT: psraw $1, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm5[0,1],xmm1[2],xmm5[3,4,5,6],xmm1[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm4[0],xmm1[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE41-NEXT: psraw $15, %xmm4
|
|
; SSE41-NEXT: pmulhuw %xmm7, %xmm4
|
|
; SSE41-NEXT: paddw %xmm2, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE41-NEXT: pmulhw %xmm6, %xmm5
|
|
; SSE41-NEXT: psraw $1, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm5[0,1],xmm4[2],xmm5[3,4,5,6],xmm4[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm2[0],xmm4[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm5
|
|
; SSE41-NEXT: psraw $15, %xmm5
|
|
; SSE41-NEXT: pmulhuw %xmm7, %xmm5
|
|
; SSE41-NEXT: paddw %xmm3, %xmm5
|
|
; SSE41-NEXT: pmulhw %xmm5, %xmm6
|
|
; SSE41-NEXT: psraw $1, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm6[0,1],xmm5[2],xmm6[3,4,5,6],xmm5[7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm3[0],xmm5[1,2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm3
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
|
|
; AVX1-NEXT: vpsraw $15, %xmm2, %xmm3
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,4,2,16,8,32,64,2]
|
|
; AVX1-NEXT: vpmulhuw %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpaddw %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = <u,16384,32768,4096,8192,2048,1024,32768>
|
|
; AVX1-NEXT: vpmulhw %xmm3, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm2[0],xmm5[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm5[0,1],xmm2[2],xmm5[3,4,5,6],xmm2[7]
|
|
; AVX1-NEXT: vpsraw $15, %xmm0, %xmm5
|
|
; AVX1-NEXT: vpmulhuw %xmm4, %xmm5, %xmm5
|
|
; AVX1-NEXT: vpaddw %xmm5, %xmm0, %xmm5
|
|
; AVX1-NEXT: vpmulhw %xmm3, %xmm5, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm6 = xmm5[0],xmm6[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm5, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm6[0,1],xmm5[2],xmm6[3,4,5,6],xmm5[7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm5, %ymm2
|
|
; AVX1-NEXT: vbroadcastf128 {{.*#+}} ymm5 = [0,65535,65535,65535,65535,65535,65535,65535,0,65535,65535,65535,65535,65535,65535,65535]
|
|
; AVX1-NEXT: # ymm5 = mem[0,1,0,1]
|
|
; AVX1-NEXT: vandps %ymm5, %ymm2, %ymm2
|
|
; AVX1-NEXT: vandnps %ymm0, %ymm5, %ymm0
|
|
; AVX1-NEXT: vorps %ymm0, %ymm2, %ymm0
|
|
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
|
|
; AVX1-NEXT: vpsraw $15, %xmm2, %xmm6
|
|
; AVX1-NEXT: vpmulhuw %xmm4, %xmm6, %xmm6
|
|
; AVX1-NEXT: vpaddw %xmm6, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpmulhw %xmm3, %xmm2, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm6 = xmm2[0],xmm6[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm6[0,1],xmm2[2],xmm6[3,4,5,6],xmm2[7]
|
|
; AVX1-NEXT: vpsraw $15, %xmm1, %xmm6
|
|
; AVX1-NEXT: vpmulhuw %xmm4, %xmm6, %xmm4
|
|
; AVX1-NEXT: vpaddw %xmm4, %xmm1, %xmm4
|
|
; AVX1-NEXT: vpmulhw %xmm3, %xmm4, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0],xmm3[1,2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2],xmm3[3,4,5,6],xmm4[7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
|
|
; AVX1-NEXT: vandps %ymm5, %ymm2, %ymm2
|
|
; AVX1-NEXT: vandnps %ymm1, %ymm5, %ymm1
|
|
; AVX1-NEXT: vorps %ymm1, %ymm2, %ymm1
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsraw $15, %ymm0, %ymm2
|
|
; AVX2-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,4,2,16,8,32,64,2,1,4,2,16,8,32,64,2]
|
|
; AVX2-NEXT: # ymm3 = mem[0,1,0,1]
|
|
; AVX2-NEXT: vpmulhuw %ymm3, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpaddw %ymm2, %ymm0, %ymm2
|
|
; AVX2-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [0,16384,32768,4096,8192,2048,1024,32768,0,16384,32768,4096,8192,2048,1024,32768]
|
|
; AVX2-NEXT: # ymm4 = mem[0,1,0,1]
|
|
; AVX2-NEXT: vpmulhw %ymm4, %ymm2, %ymm5
|
|
; AVX2-NEXT: vpsraw $1, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm2 = ymm5[0,1],ymm2[2],ymm5[3,4,5,6],ymm2[7],ymm5[8,9],ymm2[10],ymm5[11,12,13,14],ymm2[15]
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3,4,5,6,7],ymm0[8],ymm2[9,10,11,12,13,14,15]
|
|
; AVX2-NEXT: vpsraw $15, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpmulhuw %ymm3, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpaddw %ymm2, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpmulhw %ymm4, %ymm2, %ymm3
|
|
; AVX2-NEXT: vpsraw $1, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm2 = ymm3[0,1],ymm2[2],ymm3[3,4,5,6],ymm2[7],ymm3[8,9],ymm2[10],ymm3[11,12,13,14],ymm2[15]
|
|
; AVX2-NEXT: vpblendw {{.*#+}} ymm1 = ymm1[0],ymm2[1,2,3,4,5,6,7],ymm1[8],ymm2[9,10,11,12,13,14,15]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpsraw $15, %ymm0, %ymm2
|
|
; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,4,2,16,8,32,64,2,1,4,2,16,8,32,64,2]
|
|
; AVX512F-NEXT: # ymm3 = mem[0,1,0,1]
|
|
; AVX512F-NEXT: vpmulhuw %ymm3, %ymm2, %ymm2
|
|
; AVX512F-NEXT: vpaddw %ymm2, %ymm0, %ymm2
|
|
; AVX512F-NEXT: vpmovsxwd %ymm2, %zmm2
|
|
; AVX512F-NEXT: vbroadcasti64x4 {{.*#+}} zmm4 = [0,2,1,4,3,5,6,1,0,2,1,4,3,5,6,1]
|
|
; AVX512F-NEXT: # zmm4 = mem[0,1,2,3,0,1,2,3]
|
|
; AVX512F-NEXT: vpsravd %zmm4, %zmm2, %zmm2
|
|
; AVX512F-NEXT: vpmovdw %zmm2, %ymm2
|
|
; AVX512F-NEXT: vpblendw {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3,4,5,6,7],ymm0[8],ymm2[9,10,11,12,13,14,15]
|
|
; AVX512F-NEXT: vpsraw $15, %ymm1, %ymm2
|
|
; AVX512F-NEXT: vpmulhuw %ymm3, %ymm2, %ymm2
|
|
; AVX512F-NEXT: vpaddw %ymm2, %ymm1, %ymm2
|
|
; AVX512F-NEXT: vpmovsxwd %ymm2, %zmm2
|
|
; AVX512F-NEXT: vpsravd %zmm4, %zmm2, %zmm2
|
|
; AVX512F-NEXT: vpmovdw %zmm2, %ymm2
|
|
; AVX512F-NEXT: vpblendw {{.*#+}} ymm1 = ymm1[0],ymm2[1,2,3,4,5,6,7],ymm1[8],ymm2[9,10,11,12,13,14,15]
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraw $15, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsrlvw {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: vpaddw %zmm1, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: movl $16843009, %eax # imm = 0x1010101
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqu16 %zmm0, %zmm1 {%k1}
|
|
; AVX512BW-NEXT: vmovdqa64 %zmm1, %zmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v32i16:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm2
|
|
; XOP-NEXT: vpsraw $15, %xmm2, %xmm3
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm4 = [65520,65522,65521,65524,65523,65525,65526,65521]
|
|
; XOP-NEXT: vpshlw %xmm4, %xmm3, %xmm3
|
|
; XOP-NEXT: vpaddw %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [0,65534,65535,65532,65533,65531,65530,65535]
|
|
; XOP-NEXT: vpshaw %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpsraw $15, %xmm0, %xmm5
|
|
; XOP-NEXT: vpshlw %xmm4, %xmm5, %xmm5
|
|
; XOP-NEXT: vpaddw %xmm5, %xmm0, %xmm5
|
|
; XOP-NEXT: vpshaw %xmm3, %xmm5, %xmm5
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm5, %ymm2
|
|
; XOP-NEXT: vbroadcastf128 {{.*#+}} ymm5 = [0,65535,65535,65535,65535,65535,65535,65535,0,65535,65535,65535,65535,65535,65535,65535]
|
|
; XOP-NEXT: # ymm5 = mem[0,1,0,1]
|
|
; XOP-NEXT: vpcmov %ymm5, %ymm0, %ymm2, %ymm0
|
|
; XOP-NEXT: vextractf128 $1, %ymm1, %xmm2
|
|
; XOP-NEXT: vpsraw $15, %xmm2, %xmm6
|
|
; XOP-NEXT: vpshlw %xmm4, %xmm6, %xmm6
|
|
; XOP-NEXT: vpaddw %xmm6, %xmm2, %xmm2
|
|
; XOP-NEXT: vpshaw %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpsraw $15, %xmm1, %xmm6
|
|
; XOP-NEXT: vpshlw %xmm4, %xmm6, %xmm4
|
|
; XOP-NEXT: vpaddw %xmm4, %xmm1, %xmm4
|
|
; XOP-NEXT: vpshaw %xmm3, %xmm4, %xmm3
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
|
|
; XOP-NEXT: vpcmov %ymm5, %ymm1, %ymm2, %ymm1
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <32 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2>
|
|
ret <32 x i16> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2b_v4i32(<4 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v4i32:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrad $31, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psrld $28, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psrld $29, %xmm3
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
|
|
; SSE2-NEXT: psrld $30, %xmm1
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm3[0,3]
|
|
; SSE2-NEXT: paddd %xmm0, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psrad $4, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psrad $3, %xmm3
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
|
|
; SSE2-NEXT: psrad $2, %xmm1
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm3[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3]
|
|
; SSE2-NEXT: movaps %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v4i32:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrad $31, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psrld $28, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE41-NEXT: psrld $30, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm1
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
|
|
; SSE41-NEXT: paddd %xmm0, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE41-NEXT: psrad $4, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE41-NEXT: psrad $2, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3,4,5,6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v4i32:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrld $28, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrld $30, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm3[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrad $4, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrad $2, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_by_pow2b_v4i32:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsrlvd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v4i32:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshld {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshad {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 1, i32 4, i32 8, i32 16>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <8 x i32> @combine_vec_sdiv_by_pow2b_v8i32(<8 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v8i32:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrad $31, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psrld $28, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE2-NEXT: psrld $29, %xmm4
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm4 = xmm4[1],xmm3[1]
|
|
; SSE2-NEXT: psrld $30, %xmm0
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3],xmm4[0,3]
|
|
; SSE2-NEXT: paddd %xmm2, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psrad $4, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE2-NEXT: psrad $3, %xmm4
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm4 = xmm4[1],xmm3[1]
|
|
; SSE2-NEXT: psrad $2, %xmm0
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm4[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm0 = xmm2[0],xmm0[1,2,3]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psrad $31, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: psrld $28, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE2-NEXT: psrld $29, %xmm4
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm4 = xmm4[1],xmm3[1]
|
|
; SSE2-NEXT: psrld $30, %xmm2
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,3],xmm4[0,3]
|
|
; SSE2-NEXT: paddd %xmm1, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: psrad $4, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE2-NEXT: psrad $3, %xmm4
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm4 = xmm4[1],xmm3[1]
|
|
; SSE2-NEXT: psrad $2, %xmm2
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,1],xmm4[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm2 = xmm1[0],xmm2[1,2,3]
|
|
; SSE2-NEXT: movaps %xmm2, %xmm1
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v8i32:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psrad $31, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE41-NEXT: psrld $28, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE41-NEXT: psrld $30, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm0
|
|
; SSE41-NEXT: pxor %xmm3, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm3[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm4[2,3],xmm0[4,5],xmm4[6,7]
|
|
; SSE41-NEXT: paddd %xmm2, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE41-NEXT: psrad $4, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE41-NEXT: psrad $2, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm0
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm5[2,3],xmm0[4,5],xmm5[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm2[0,1],xmm0[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psrad $31, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE41-NEXT: psrld $28, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm5
|
|
; SSE41-NEXT: psrld $30, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm5[2,3],xmm2[4,5],xmm5[6,7]
|
|
; SSE41-NEXT: paddd %xmm1, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE41-NEXT: psrad $4, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE41-NEXT: psrad $2, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm4[2,3],xmm2[4,5],xmm4[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm1[0,1],xmm2[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v8i32:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
|
|
; AVX1-NEXT: vpsrad $31, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrld $28, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpsrld $30, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpxor %xmm4, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsrad $4, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrad $2, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpsrld $28, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpsrld $30, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm5[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm2, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpsrad $4, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpsrad $2, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_by_pow2b_v8i32:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpsrad $31, %ymm0, %ymm1
|
|
; AVX2ORLATER-NEXT: vpsrlvd {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2ORLATER-NEXT: vpaddd %ymm1, %ymm0, %ymm1
|
|
; AVX2ORLATER-NEXT: vpsravd {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v8i32:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm1
|
|
; XOP-NEXT: vpsrad $31, %xmm1, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [4294967264,4294967266,4294967267,4294967268]
|
|
; XOP-NEXT: vpshld %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpaddd %xmm2, %xmm1, %xmm1
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm2 = [0,4294967294,4294967293,4294967292]
|
|
; XOP-NEXT: vpshad %xmm2, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsrad $31, %xmm0, %xmm4
|
|
; XOP-NEXT: vpshld %xmm3, %xmm4, %xmm3
|
|
; XOP-NEXT: vpaddd %xmm3, %xmm0, %xmm3
|
|
; XOP-NEXT: vpshad %xmm2, %xmm3, %xmm2
|
|
; XOP-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i32> %x, <i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16>
|
|
ret <8 x i32> %1
|
|
}
|
|
|
|
define <16 x i32> @combine_vec_sdiv_by_pow2b_v16i32(<16 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrad $31, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE2-NEXT: psrld $28, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm6
|
|
; SSE2-NEXT: psrld $29, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrld $30, %xmm0
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3],xmm6[0,3]
|
|
; SSE2-NEXT: paddd %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE2-NEXT: psrad $4, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm6
|
|
; SSE2-NEXT: psrad $3, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrad $2, %xmm0
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm6[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm1
|
|
; SSE2-NEXT: psrad $31, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE2-NEXT: psrld $28, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm6
|
|
; SSE2-NEXT: psrld $29, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrld $30, %xmm1
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm6[0,3]
|
|
; SSE2-NEXT: paddd %xmm4, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE2-NEXT: psrad $4, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm6
|
|
; SSE2-NEXT: psrad $3, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrad $2, %xmm1
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm6[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm1 = xmm4[0],xmm1[1,2,3]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE2-NEXT: psrad $31, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE2-NEXT: psrld $28, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm6
|
|
; SSE2-NEXT: psrld $29, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrld $30, %xmm4
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm4 = zero,zero,zero,zero,zero,zero,zero,zero,xmm4[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,3],xmm6[0,3]
|
|
; SSE2-NEXT: paddd %xmm2, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE2-NEXT: psrad $4, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm6
|
|
; SSE2-NEXT: psrad $3, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm5[1]
|
|
; SSE2-NEXT: psrad $2, %xmm4
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm4 = xmm4[0,1],xmm6[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm4 = xmm2[0],xmm4[1,2,3]
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm5
|
|
; SSE2-NEXT: psrad $31, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm2
|
|
; SSE2-NEXT: psrld $28, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm6
|
|
; SSE2-NEXT: psrld $29, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm2[1]
|
|
; SSE2-NEXT: psrld $30, %xmm5
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm5 = zero,zero,zero,zero,zero,zero,zero,zero,xmm5[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm5 = xmm5[0,3],xmm6[0,3]
|
|
; SSE2-NEXT: paddd %xmm3, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm2
|
|
; SSE2-NEXT: psrad $4, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm6
|
|
; SSE2-NEXT: psrad $3, %xmm6
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm2[1]
|
|
; SSE2-NEXT: psrad $2, %xmm5
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm5 = xmm5[0,1],xmm6[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm5 = xmm3[0],xmm5[1,2,3]
|
|
; SSE2-NEXT: movaps %xmm4, %xmm2
|
|
; SSE2-NEXT: movaps %xmm5, %xmm3
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrad $31, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE41-NEXT: psrld $28, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm7
|
|
; SSE41-NEXT: psrld $30, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm0
|
|
; SSE41-NEXT: pxor %xmm6, %xmm6
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm6[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm7[2,3],xmm0[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: paddd %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm5
|
|
; SSE41-NEXT: psrad $4, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm7
|
|
; SSE41-NEXT: psrad $2, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm0
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm7[2,3],xmm0[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm1
|
|
; SSE41-NEXT: psrad $31, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE41-NEXT: psrld $28, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm7
|
|
; SSE41-NEXT: psrld $30, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm6[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm7[2,3],xmm1[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: paddd %xmm4, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE41-NEXT: psrad $4, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm7
|
|
; SSE41-NEXT: psrad $2, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm7[2,3],xmm1[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm4[0,1],xmm1[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm4
|
|
; SSE41-NEXT: psrad $31, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE41-NEXT: psrld $28, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm7
|
|
; SSE41-NEXT: psrld $30, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm6[0,1,2,3],xmm4[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1],xmm7[2,3],xmm4[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: paddd %xmm2, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm5
|
|
; SSE41-NEXT: psrad $4, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm7
|
|
; SSE41-NEXT: psrad $2, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1],xmm7[2,3],xmm4[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm2[0,1],xmm4[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm5
|
|
; SSE41-NEXT: psrad $31, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm2
|
|
; SSE41-NEXT: psrld $28, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm7
|
|
; SSE41-NEXT: psrld $30, %xmm7
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm7 = xmm7[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1],xmm7[2,3],xmm5[4,5],xmm7[6,7]
|
|
; SSE41-NEXT: paddd %xmm3, %xmm5
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm2
|
|
; SSE41-NEXT: psrad $4, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm6
|
|
; SSE41-NEXT: psrad $2, %xmm6
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm6 = xmm6[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1],xmm6[2,3],xmm5[4,5],xmm6[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm3[0,1],xmm5[2,3,4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm5, %xmm3
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
|
|
; AVX1-NEXT: vpsrad $31, %xmm3, %xmm2
|
|
; AVX1-NEXT: vpsrld $28, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpsrld $30, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm2[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1],xmm4[2,3],xmm5[4,5],xmm4[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsrad $4, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsrad $2, %xmm3, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm3, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2,3],xmm3[4,5],xmm4[6,7]
|
|
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm4
|
|
; AVX1-NEXT: vpsrld $28, %xmm4, %xmm5
|
|
; AVX1-NEXT: vpsrld $30, %xmm4, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1],xmm5[2,3],xmm4[4,5],xmm5[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm4, %xmm0, %xmm4
|
|
; AVX1-NEXT: vpsrad $4, %xmm4, %xmm5
|
|
; AVX1-NEXT: vpsrad $2, %xmm4, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm4, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm6[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1],xmm5[2,3],xmm4[4,5],xmm5[6,7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm4, %ymm3
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm3[1,2,3],ymm0[4],ymm3[5,6,7]
|
|
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
|
|
; AVX1-NEXT: vpsrad $31, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsrld $28, %xmm4, %xmm5
|
|
; AVX1-NEXT: vpsrld $30, %xmm4, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1],xmm5[2,3],xmm4[4,5],xmm5[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsrad $4, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsrad $2, %xmm3, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm3, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2,3],xmm3[4,5],xmm4[6,7]
|
|
; AVX1-NEXT: vpsrad $31, %xmm1, %xmm4
|
|
; AVX1-NEXT: vpsrld $28, %xmm4, %xmm5
|
|
; AVX1-NEXT: vpsrld $30, %xmm4, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm5[2,3],xmm2[4,5],xmm5[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrad $4, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpsrad $2, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm4[2,3],xmm2[4,5],xmm4[6,7]
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm2, %ymm2
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm1 = ymm1[0],ymm2[1,2,3],ymm1[4],ymm2[5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrad $31, %ymm0, %ymm2
|
|
; AVX2-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [32,30,29,28,32,30,29,28]
|
|
; AVX2-NEXT: # ymm3 = mem[0,1,0,1]
|
|
; AVX2-NEXT: vpsrlvd %ymm3, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpaddd %ymm2, %ymm0, %ymm2
|
|
; AVX2-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [0,2,3,4,0,2,3,4]
|
|
; AVX2-NEXT: # ymm4 = mem[0,1,0,1]
|
|
; AVX2-NEXT: vpsravd %ymm4, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3],ymm0[4],ymm2[5,6,7]
|
|
; AVX2-NEXT: vpsrad $31, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpsrlvd %ymm3, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpaddd %ymm2, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpsravd %ymm4, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],ymm2[1,2,3],ymm1[4],ymm2[5,6,7]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpsrad $31, %zmm0, %zmm1
|
|
; AVX512F-NEXT: vpsrlvd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: vpaddd %zmm1, %zmm0, %zmm1
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: movw $4369, %ax # imm = 0x1111
|
|
; AVX512F-NEXT: kmovw %eax, %k1
|
|
; AVX512F-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1}
|
|
; AVX512F-NEXT: vmovdqa64 %zmm1, %zmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsrad $31, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsrlvd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: vpaddd %zmm1, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsravd {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: movw $4369, %ax # imm = 0x1111
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1}
|
|
; AVX512BW-NEXT: vmovdqa64 %zmm1, %zmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v16i32:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm2
|
|
; XOP-NEXT: vpsrad $31, %xmm2, %xmm3
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm4 = [4294967264,4294967266,4294967267,4294967268]
|
|
; XOP-NEXT: vpshld %xmm4, %xmm3, %xmm3
|
|
; XOP-NEXT: vpaddd %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [0,4294967294,4294967293,4294967292]
|
|
; XOP-NEXT: vpshad %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpsrad $31, %xmm0, %xmm5
|
|
; XOP-NEXT: vpshld %xmm4, %xmm5, %xmm5
|
|
; XOP-NEXT: vpaddd %xmm5, %xmm0, %xmm5
|
|
; XOP-NEXT: vpshad %xmm3, %xmm5, %xmm5
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm5, %ymm2
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3],ymm0[4],ymm2[5,6,7]
|
|
; XOP-NEXT: vextractf128 $1, %ymm1, %xmm2
|
|
; XOP-NEXT: vpsrad $31, %xmm2, %xmm5
|
|
; XOP-NEXT: vpshld %xmm4, %xmm5, %xmm5
|
|
; XOP-NEXT: vpaddd %xmm5, %xmm2, %xmm2
|
|
; XOP-NEXT: vpshad %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vpsrad $31, %xmm1, %xmm5
|
|
; XOP-NEXT: vpshld %xmm4, %xmm5, %xmm4
|
|
; XOP-NEXT: vpaddd %xmm4, %xmm1, %xmm4
|
|
; XOP-NEXT: vpshad %xmm3, %xmm4, %xmm3
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm1 = ymm1[0],ymm2[1,2,3],ymm1[4],ymm2[5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <16 x i32> %x, <i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16>
|
|
ret <16 x i32> %1
|
|
}
|
|
|
|
define <2 x i64> @combine_vec_sdiv_by_pow2b_v2i64(<2 x i64> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrad $31, %xmm1
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE2-NEXT: psrlq $62, %xmm1
|
|
; SSE2-NEXT: pxor %xmm2, %xmm2
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm2 = xmm2[1],xmm1[1]
|
|
; SSE2-NEXT: paddq %xmm0, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE2-NEXT: psrlq $2, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1]
|
|
; SSE2-NEXT: movapd {{.*#+}} xmm2 = [9223372036854775808,2305843009213693952]
|
|
; SSE2-NEXT: xorpd %xmm2, %xmm1
|
|
; SSE2-NEXT: psubq %xmm2, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm0[0],xmm1[1]
|
|
; SSE2-NEXT: movapd %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrad $31, %xmm1
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE41-NEXT: psrlq $62, %xmm1
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm0, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm1
|
|
; SSE41-NEXT: psrlq $2, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [9223372036854775808,2305843009213693952]
|
|
; SSE41-NEXT: pxor %xmm2, %xmm1
|
|
; SSE41-NEXT: psubq %xmm2, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpcmpgtq %xmm0, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrlq $62, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrlq $2, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [9223372036854775808,2305843009213693952]
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsubq %xmm2, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpcmpgtq %xmm0, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpsrlvq {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpaddq %xmm1, %xmm0, %xmm1
|
|
; AVX2-NEXT: movl $2, %eax
|
|
; AVX2-NEXT: vmovq %rax, %xmm2
|
|
; AVX2-NEXT: vpslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1,2,3,4,5,6,7]
|
|
; AVX2-NEXT: vpsrlvq %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [9223372036854775808,2305843009213693952]
|
|
; AVX2-NEXT: vpxor %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpsubq %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 def $zmm0
|
|
; AVX512F-NEXT: movl $2, %eax
|
|
; AVX512F-NEXT: vmovq %rax, %xmm1
|
|
; AVX512F-NEXT: vpslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7]
|
|
; AVX512F-NEXT: vpsraq $63, %zmm0, %zmm2
|
|
; AVX512F-NEXT: vpsrlvq {{.*}}(%rip), %xmm2, %xmm2
|
|
; AVX512F-NEXT: vpaddq %xmm2, %xmm0, %xmm2
|
|
; AVX512F-NEXT: vpsravq %zmm1, %zmm2, %zmm1
|
|
; AVX512F-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraq $63, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsrlvq {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpaddq %xmm1, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: movl $2, %eax
|
|
; AVX512BW-NEXT: vmovq %rax, %xmm2
|
|
; AVX512BW-NEXT: vpslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1,2,3,4,5,6,7]
|
|
; AVX512BW-NEXT: vpsravq %xmm2, %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v2i64:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpshaq {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpshlq {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddq %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: movq $-2, %rax
|
|
; XOP-NEXT: vmovq %rax, %xmm2
|
|
; XOP-NEXT: vpslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1,2,3,4,5,6,7]
|
|
; XOP-NEXT: vpshaq %xmm2, %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <2 x i64> %x, <i64 1, i64 4>
|
|
ret <2 x i64> %1
|
|
}
|
|
|
|
define <4 x i64> @combine_vec_sdiv_by_pow2b_v4i64(<4 x i64> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psrad $31, %xmm1
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psrlq $61, %xmm3
|
|
; SSE2-NEXT: psrlq $60, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm3[0],xmm1[1]
|
|
; SSE2-NEXT: paddq %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psrlq $3, %xmm2
|
|
; SSE2-NEXT: psrlq $4, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm2[0],xmm1[1]
|
|
; SSE2-NEXT: movapd {{.*#+}} xmm2 = [1152921504606846976,576460752303423488]
|
|
; SSE2-NEXT: xorpd %xmm2, %xmm1
|
|
; SSE2-NEXT: psubq %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrad $31, %xmm2
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
|
|
; SSE2-NEXT: psrlq $62, %xmm2
|
|
; SSE2-NEXT: pxor %xmm3, %xmm3
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
|
|
; SSE2-NEXT: paddq %xmm0, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm2
|
|
; SSE2-NEXT: psrlq $2, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm3[0],xmm2[1]
|
|
; SSE2-NEXT: movapd {{.*#+}} xmm3 = [9223372036854775808,2305843009213693952]
|
|
; SSE2-NEXT: xorpd %xmm3, %xmm2
|
|
; SSE2-NEXT: psubq %xmm3, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
|
|
; SSE2-NEXT: movapd %xmm2, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psrad $31, %xmm1
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE41-NEXT: psrlq $60, %xmm3
|
|
; SSE41-NEXT: psrlq $61, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psrlq $4, %xmm2
|
|
; SSE41-NEXT: psrlq $3, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [1152921504606846976,576460752303423488]
|
|
; SSE41-NEXT: pxor %xmm2, %xmm1
|
|
; SSE41-NEXT: psubq %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psrad $31, %xmm2
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
|
|
; SSE41-NEXT: psrlq $62, %xmm2
|
|
; SSE41-NEXT: pxor %xmm3, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm0, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm2
|
|
; SSE41-NEXT: psrlq $2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [9223372036854775808,2305843009213693952]
|
|
; SSE41-NEXT: pxor %xmm3, %xmm2
|
|
; SSE41-NEXT: psubq %xmm3, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpsrlq $60, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsrlq $61, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsrlq $4, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpsrlq $3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1152921504606846976,576460752303423488]
|
|
; AVX1-NEXT: vpxor %xmm3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsubq %xmm3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpcmpgtq %xmm0, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpsrlq $62, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm2, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpsrlq $2, %xmm2, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm3[4,5,6,7]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [9223372036854775808,2305843009213693952]
|
|
; AVX1-NEXT: vpxor %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsubq %xmm3, %xmm2, %xmm2
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsrlvq {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpaddq %ymm1, %ymm0, %ymm1
|
|
; AVX2-NEXT: vpsrlvq {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [9223372036854775808,2305843009213693952,1152921504606846976,576460752303423488]
|
|
; AVX2-NEXT: vpxor %ymm2, %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsubq %ymm2, %ymm1, %ymm1
|
|
; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
|
|
; AVX512F-NEXT: vmovdqa {{.*#+}} ymm1 = [0,2,3,4]
|
|
; AVX512F-NEXT: vpsraq $63, %zmm0, %zmm2
|
|
; AVX512F-NEXT: vpsrlvq {{.*}}(%rip), %ymm2, %ymm2
|
|
; AVX512F-NEXT: vpaddq %ymm2, %ymm0, %ymm2
|
|
; AVX512F-NEXT: vpsravq %zmm1, %zmm2, %zmm1
|
|
; AVX512F-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraq $63, %ymm0, %ymm1
|
|
; AVX512BW-NEXT: vpsrlvq {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpaddq %ymm1, %ymm0, %ymm1
|
|
; AVX512BW-NEXT: vpsravq {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v4i64:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm1 = [18446744073709551553,18446744073709551553]
|
|
; XOP-NEXT: vpshaq %xmm1, %xmm0, %xmm2
|
|
; XOP-NEXT: vpshlq {{.*}}(%rip), %xmm2, %xmm2
|
|
; XOP-NEXT: vpaddq %xmm2, %xmm0, %xmm2
|
|
; XOP-NEXT: movq $-2, %rax
|
|
; XOP-NEXT: vmovq %rax, %xmm3
|
|
; XOP-NEXT: vpslldq {{.*#+}} xmm3 = zero,zero,zero,zero,zero,zero,zero,zero,xmm3[0,1,2,3,4,5,6,7]
|
|
; XOP-NEXT: vpshaq %xmm3, %xmm2, %xmm2
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm3
|
|
; XOP-NEXT: vpshaq %xmm1, %xmm3, %xmm1
|
|
; XOP-NEXT: vpshlq {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddq %xmm1, %xmm3, %xmm1
|
|
; XOP-NEXT: vpshaq {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <4 x i64> %x, <i64 1, i64 4, i64 8, i64 16>
|
|
ret <4 x i64> %1
|
|
}
|
|
|
|
define <8 x i64> @combine_vec_sdiv_by_pow2b_v8i64(<8 x i64> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm4
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psrad $31, %xmm1
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE2-NEXT: psrlq $61, %xmm5
|
|
; SSE2-NEXT: psrlq $60, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
|
|
; SSE2-NEXT: paddq %xmm3, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: psrlq $3, %xmm3
|
|
; SSE2-NEXT: psrlq $4, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm3[0],xmm1[1]
|
|
; SSE2-NEXT: movapd {{.*#+}} xmm5 = [1152921504606846976,576460752303423488]
|
|
; SSE2-NEXT: xorpd %xmm5, %xmm1
|
|
; SSE2-NEXT: psubq %xmm5, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm4, %xmm3
|
|
; SSE2-NEXT: psrad $31, %xmm3
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm6
|
|
; SSE2-NEXT: psrlq $61, %xmm6
|
|
; SSE2-NEXT: psrlq $60, %xmm3
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm3 = xmm6[0],xmm3[1]
|
|
; SSE2-NEXT: paddq %xmm4, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm3, %xmm4
|
|
; SSE2-NEXT: psrlq $3, %xmm4
|
|
; SSE2-NEXT: psrlq $4, %xmm3
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm3 = xmm4[0],xmm3[1]
|
|
; SSE2-NEXT: xorpd %xmm5, %xmm3
|
|
; SSE2-NEXT: psubq %xmm5, %xmm3
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE2-NEXT: psrad $31, %xmm4
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
|
|
; SSE2-NEXT: psrlq $62, %xmm4
|
|
; SSE2-NEXT: pxor %xmm6, %xmm6
|
|
; SSE2-NEXT: pxor %xmm5, %xmm5
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm5 = xmm5[1],xmm4[1]
|
|
; SSE2-NEXT: paddq %xmm0, %xmm5
|
|
; SSE2-NEXT: movdqa %xmm5, %xmm4
|
|
; SSE2-NEXT: psrlq $2, %xmm4
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm4 = xmm5[0],xmm4[1]
|
|
; SSE2-NEXT: movapd {{.*#+}} xmm7 = [9223372036854775808,2305843009213693952]
|
|
; SSE2-NEXT: xorpd %xmm7, %xmm4
|
|
; SSE2-NEXT: psubq %xmm7, %xmm4
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm4 = xmm0[0],xmm4[1]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE2-NEXT: psrad $31, %xmm0
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
|
|
; SSE2-NEXT: psrlq $62, %xmm0
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm6 = xmm6[1],xmm0[1]
|
|
; SSE2-NEXT: paddq %xmm2, %xmm6
|
|
; SSE2-NEXT: movdqa %xmm6, %xmm5
|
|
; SSE2-NEXT: psrlq $2, %xmm5
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm6[0],xmm5[1]
|
|
; SSE2-NEXT: xorpd %xmm7, %xmm5
|
|
; SSE2-NEXT: psubq %xmm7, %xmm5
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm2[0],xmm5[1]
|
|
; SSE2-NEXT: movapd %xmm4, %xmm0
|
|
; SSE2-NEXT: movapd %xmm5, %xmm2
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE41-NEXT: psrad $31, %xmm1
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm5
|
|
; SSE41-NEXT: psrlq $60, %xmm5
|
|
; SSE41-NEXT: psrlq $61, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm3, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE41-NEXT: psrlq $4, %xmm3
|
|
; SSE41-NEXT: psrlq $3, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [1152921504606846976,576460752303423488]
|
|
; SSE41-NEXT: pxor %xmm5, %xmm1
|
|
; SSE41-NEXT: psubq %xmm5, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm3
|
|
; SSE41-NEXT: psrad $31, %xmm3
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm6
|
|
; SSE41-NEXT: psrlq $60, %xmm6
|
|
; SSE41-NEXT: psrlq $61, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm6[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm4, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm4
|
|
; SSE41-NEXT: psrlq $4, %xmm4
|
|
; SSE41-NEXT: psrlq $3, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm4[4,5,6,7]
|
|
; SSE41-NEXT: pxor %xmm5, %xmm3
|
|
; SSE41-NEXT: psubq %xmm5, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE41-NEXT: psrad $31, %xmm4
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
|
|
; SSE41-NEXT: psrlq $62, %xmm4
|
|
; SSE41-NEXT: pxor %xmm5, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm0, %xmm4
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm6
|
|
; SSE41-NEXT: psrlq $2, %xmm6
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm6 = xmm4[0,1,2,3],xmm6[4,5,6,7]
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [9223372036854775808,2305843009213693952]
|
|
; SSE41-NEXT: pxor %xmm4, %xmm6
|
|
; SSE41-NEXT: psubq %xmm4, %xmm6
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm6[4,5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm6
|
|
; SSE41-NEXT: psrad $31, %xmm6
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
|
|
; SSE41-NEXT: psrlq $62, %xmm6
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm6 = xmm5[0,1,2,3],xmm6[4,5,6,7]
|
|
; SSE41-NEXT: paddq %xmm2, %xmm6
|
|
; SSE41-NEXT: movdqa %xmm6, %xmm5
|
|
; SSE41-NEXT: psrlq $2, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm6[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: pxor %xmm4, %xmm5
|
|
; SSE41-NEXT: psubq %xmm4, %xmm5
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm5[4,5,6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpsrlq $60, %xmm4, %xmm5
|
|
; AVX1-NEXT: vpsrlq $61, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsrlq $4, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsrlq $3, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1152921504606846976,576460752303423488]
|
|
; AVX1-NEXT: vpxor %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsubq %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpcmpgtq %xmm0, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpsrlq $62, %xmm5, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm2[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm5, %xmm0, %xmm5
|
|
; AVX1-NEXT: vpsrlq $2, %xmm5, %xmm6
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm6[4,5,6,7]
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [9223372036854775808,2305843009213693952]
|
|
; AVX1-NEXT: vpxor %xmm6, %xmm5, %xmm5
|
|
; AVX1-NEXT: vpsubq %xmm6, %xmm5, %xmm5
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm5, %ymm3
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm3[2,3,4,5,6,7]
|
|
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
|
|
; AVX1-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm5
|
|
; AVX1-NEXT: vpsrlq $60, %xmm5, %xmm7
|
|
; AVX1-NEXT: vpsrlq $61, %xmm5, %xmm5
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm7[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm5, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsrlq $4, %xmm3, %xmm5
|
|
; AVX1-NEXT: vpsrlq $3, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm5[4,5,6,7]
|
|
; AVX1-NEXT: vpxor %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsubq %xmm4, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpcmpgtq %xmm1, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpsrlq $62, %xmm4, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpaddq %xmm2, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrlq $2, %xmm2, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm4[4,5,6,7]
|
|
; AVX1-NEXT: vpxor %xmm6, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsubq %xmm6, %xmm2, %xmm2
|
|
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm2, %ymm2
|
|
; AVX1-NEXT: vblendps {{.*#+}} ymm1 = ymm1[0,1],ymm2[2,3,4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} ymm4 = [64,62,61,60]
|
|
; AVX2-NEXT: vpsrlvq %ymm4, %ymm3, %ymm3
|
|
; AVX2-NEXT: vpaddq %ymm3, %ymm0, %ymm3
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} ymm5 = [0,2,3,4]
|
|
; AVX2-NEXT: vpsrlvq %ymm5, %ymm3, %ymm3
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} ymm6 = [9223372036854775808,2305843009213693952,1152921504606846976,576460752303423488]
|
|
; AVX2-NEXT: vpxor %ymm6, %ymm3, %ymm3
|
|
; AVX2-NEXT: vpsubq %ymm6, %ymm3, %ymm3
|
|
; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm3[2,3,4,5,6,7]
|
|
; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpsrlvq %ymm4, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpaddq %ymm2, %ymm1, %ymm2
|
|
; AVX2-NEXT: vpsrlvq %ymm5, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpxor %ymm6, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpsubq %ymm6, %ymm2, %ymm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0,1],ymm2[2,3,4,5,6,7]
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpsraq $63, %zmm0, %zmm1
|
|
; AVX512F-NEXT: vpsrlvq {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: vpaddq %zmm1, %zmm0, %zmm1
|
|
; AVX512F-NEXT: vpsravq {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512F-NEXT: movb $17, %al
|
|
; AVX512F-NEXT: kmovw %eax, %k1
|
|
; AVX512F-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1}
|
|
; AVX512F-NEXT: vmovdqa64 %zmm1, %zmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpsraq $63, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsrlvq {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: vpaddq %zmm1, %zmm0, %zmm1
|
|
; AVX512BW-NEXT: vpsravq {{.*}}(%rip), %zmm1, %zmm1
|
|
; AVX512BW-NEXT: movb $17, %al
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1}
|
|
; AVX512BW-NEXT: vmovdqa64 %zmm1, %zmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_v8i64:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vextractf128 $1, %ymm0, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [18446744073709551553,18446744073709551553]
|
|
; XOP-NEXT: vpshaq %xmm3, %xmm2, %xmm4
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm8 = [18446744073709551555,18446744073709551556]
|
|
; XOP-NEXT: vpshlq %xmm8, %xmm4, %xmm4
|
|
; XOP-NEXT: vpaddq %xmm4, %xmm2, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm4 = [18446744073709551613,18446744073709551612]
|
|
; XOP-NEXT: vpshaq %xmm4, %xmm2, %xmm2
|
|
; XOP-NEXT: vpshaq %xmm3, %xmm0, %xmm6
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm7 = [18446744073709551552,18446744073709551554]
|
|
; XOP-NEXT: vpshlq %xmm7, %xmm6, %xmm6
|
|
; XOP-NEXT: vpaddq %xmm6, %xmm0, %xmm6
|
|
; XOP-NEXT: movq $-2, %rax
|
|
; XOP-NEXT: vmovq %rax, %xmm5
|
|
; XOP-NEXT: vpslldq {{.*#+}} xmm5 = zero,zero,zero,zero,zero,zero,zero,zero,xmm5[0,1,2,3,4,5,6,7]
|
|
; XOP-NEXT: vpshaq %xmm5, %xmm6, %xmm6
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm6, %ymm2
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm2[2,3,4,5,6,7]
|
|
; XOP-NEXT: vextractf128 $1, %ymm1, %xmm2
|
|
; XOP-NEXT: vpshaq %xmm3, %xmm2, %xmm6
|
|
; XOP-NEXT: vpshlq %xmm8, %xmm6, %xmm6
|
|
; XOP-NEXT: vpaddq %xmm6, %xmm2, %xmm2
|
|
; XOP-NEXT: vpshaq %xmm4, %xmm2, %xmm2
|
|
; XOP-NEXT: vpshaq %xmm3, %xmm1, %xmm3
|
|
; XOP-NEXT: vpshlq %xmm7, %xmm3, %xmm3
|
|
; XOP-NEXT: vpaddq %xmm3, %xmm1, %xmm3
|
|
; XOP-NEXT: vpshaq %xmm5, %xmm3, %xmm3
|
|
; XOP-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
|
|
; XOP-NEXT: vblendps {{.*#+}} ymm1 = ymm1[0,1],ymm2[2,3,4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i64> %x, <i64 1, i64 4, i64 8, i64 16, i64 1, i64 4, i64 8, i64 16>
|
|
ret <8 x i64> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2b_PosAndNeg(<4 x i32> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psrad $31, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrld $28, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psrld $29, %xmm3
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
|
|
; SSE2-NEXT: psrld $30, %xmm0
|
|
; SSE2-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,3],xmm3[0,3]
|
|
; SSE2-NEXT: paddd %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrad $4, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psrad $3, %xmm3
|
|
; SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
|
|
; SSE2-NEXT: psrad $2, %xmm0
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm3[0,3]
|
|
; SSE2-NEXT: movss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
|
|
; SSE2-NEXT: pxor %xmm1, %xmm1
|
|
; SSE2-NEXT: psubd %xmm0, %xmm1
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2,2,3]
|
|
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
|
|
; SSE2-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrad $31, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psrld $28, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE41-NEXT: psrld $30, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; SSE41-NEXT: psrld $29, %xmm0
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm2[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm3[2,3],xmm0[4,5],xmm3[6,7]
|
|
; SSE41-NEXT: paddd %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE41-NEXT: psrad $4, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm4
|
|
; SSE41-NEXT: psrad $2, %xmm4
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
|
|
; SSE41-NEXT: psrad $3, %xmm0
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm4[2,3],xmm0[4,5],xmm4[6,7]
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7]
|
|
; SSE41-NEXT: psubd %xmm0, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm2[2,3],xmm0[4,5],xmm2[6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrld $28, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrld $30, %xmm1, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpsrld $29, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm3[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrad $4, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsrad $2, %xmm1, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm4[0,1,2,3],xmm2[4,5,6,7]
|
|
; AVX1-NEXT: vpsrad $3, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
|
|
; AVX1-NEXT: vpsubd %xmm0, %xmm3, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsrlvd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
|
|
; AVX2ORLATER-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsubd %xmm0, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshld {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshad {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7]
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsubd %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5],xmm1[6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 1, i32 -4, i32 8, i32 -16>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2b_undef1(<4 x i32> %x) {
|
|
; CHECK-LABEL: combine_vec_sdiv_by_pow2b_undef1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 undef, i32 -4, i32 undef, i32 -16>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2b_undef2(<4 x i32> %x) {
|
|
; CHECK-LABEL: combine_vec_sdiv_by_pow2b_undef2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 undef, i32 4, i32 undef, i32 16>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
define <4 x i32> @combine_vec_sdiv_by_pow2b_undef3(<4 x i32> %x) {
|
|
; CHECK-LABEL: combine_vec_sdiv_by_pow2b_undef3:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: retq
|
|
%1 = sdiv <4 x i32> %x, <i32 undef, i32 -4, i32 undef, i32 16>
|
|
ret <4 x i32> %1
|
|
}
|
|
|
|
; PR37119
|
|
define <16 x i8> @non_splat_minus_one_divisor_0(<16 x i8> %A) {
|
|
; SSE2-LABEL: non_splat_minus_one_divisor_0:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; SSE2-NEXT: pxor %xmm2, %xmm2
|
|
; SSE2-NEXT: psubb %xmm0, %xmm2
|
|
; SSE2-NEXT: pand %xmm1, %xmm0
|
|
; SSE2-NEXT: pandn %xmm2, %xmm1
|
|
; SSE2-NEXT: por %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: non_splat_minus_one_divisor_0:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: psubb %xmm0, %xmm2
|
|
; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; SSE41-NEXT: pblendvb %xmm0, %xmm1, %xmm2
|
|
; SSE41-NEXT: movdqa %xmm2, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: non_splat_minus_one_divisor_0:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: non_splat_minus_one_divisor_0:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX2-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: non_splat_minus_one_divisor_0:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX512F-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: non_splat_minus_one_divisor_0:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512BW-NEXT: movw $443, %ax # imm = 0x1BB
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vpsubb %xmm0, %xmm1, %xmm0 {%k1}
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: non_splat_minus_one_divisor_0:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; XOP-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0
|
|
; XOP-NEXT: retq
|
|
%div = sdiv <16 x i8> %A, <i8 -1, i8 -1, i8 1, i8 -1, i8 -1, i8 -1, i8 1, i8 -1, i8 -1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
|
|
ret <16 x i8> %div
|
|
}
|
|
|
|
define <16 x i8> @non_splat_minus_one_divisor_1(<16 x i8> %A) {
|
|
; SSE2-LABEL: non_splat_minus_one_divisor_1:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: pxor %xmm0, %xmm0
|
|
; SSE2-NEXT: pxor %xmm2, %xmm2
|
|
; SSE2-NEXT: pcmpgtb %xmm1, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm0[8],xmm3[9],xmm0[9],xmm3[10],xmm0[10],xmm3[11],xmm0[11],xmm3[12],xmm0[12],xmm3[13],xmm0[13],xmm3[14],xmm0[14],xmm3[15],xmm0[15]
|
|
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
|
|
; SSE2-NEXT: psrlw $8, %xmm3
|
|
; SSE2-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]
|
|
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: packuswb %xmm3, %xmm2
|
|
; SSE2-NEXT: paddb %xmm1, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
|
|
; SSE2-NEXT: psraw $8, %xmm3
|
|
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
|
|
; SSE2-NEXT: psrlw $8, %xmm3
|
|
; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: packuswb %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; SSE2-NEXT: pand %xmm3, %xmm2
|
|
; SSE2-NEXT: pandn %xmm1, %xmm3
|
|
; SSE2-NEXT: por %xmm2, %xmm3
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; SSE2-NEXT: psubb %xmm3, %xmm0
|
|
; SSE2-NEXT: pand %xmm1, %xmm0
|
|
; SSE2-NEXT: pandn %xmm3, %xmm1
|
|
; SSE2-NEXT: por %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: non_splat_minus_one_divisor_1:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: pxor %xmm3, %xmm3
|
|
; SSE41-NEXT: pcmpgtb %xmm0, %xmm3
|
|
; SSE41-NEXT: pmovzxbw {{.*#+}} xmm4 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
|
|
; SSE41-NEXT: movdqa %xmm4, %xmm0
|
|
; SSE41-NEXT: psllw $1, %xmm0
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm4[0,1],xmm0[2],xmm4[3,4,5],xmm0[6],xmm4[7]
|
|
; SSE41-NEXT: psrlw $8, %xmm0
|
|
; SSE41-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
|
|
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm3
|
|
; SSE41-NEXT: psrlw $8, %xmm3
|
|
; SSE41-NEXT: packuswb %xmm3, %xmm0
|
|
; SSE41-NEXT: paddb %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE41-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]
|
|
; SSE41-NEXT: psraw $8, %xmm3
|
|
; SSE41-NEXT: movdqa %xmm3, %xmm4
|
|
; SSE41-NEXT: psllw $7, %xmm4
|
|
; SSE41-NEXT: psllw $8, %xmm3
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2],xmm3[3,4,5],xmm4[6],xmm3[7]
|
|
; SSE41-NEXT: psrlw $8, %xmm3
|
|
; SSE41-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
|
|
; SSE41-NEXT: psraw $8, %xmm0
|
|
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: psrlw $8, %xmm0
|
|
; SSE41-NEXT: packuswb %xmm0, %xmm3
|
|
; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; SSE41-NEXT: pblendvb %xmm0, %xmm3, %xmm1
|
|
; SSE41-NEXT: psubb %xmm1, %xmm2
|
|
; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: non_splat_minus_one_divisor_1:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm3 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
|
|
; AVX1-NEXT: vpsllw $1, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2],xmm3[3,4,5],xmm4[6],xmm3[7]
|
|
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
|
|
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpackuswb %xmm2, %xmm3, %xmm2
|
|
; AVX1-NEXT: vpaddb %xmm2, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm3 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; AVX1-NEXT: vpsraw $8, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpsllw $7, %xmm3, %xmm4
|
|
; AVX1-NEXT: vpsllw $8, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2],xmm3[3,4,5],xmm4[6],xmm3[7]
|
|
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
|
|
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
|
|
; AVX1-NEXT: vpsraw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpackuswb %xmm2, %xmm3, %xmm2
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX1-NEXT: vpblendvb %xmm3, %xmm2, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; AVX1-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: non_splat_minus_one_divisor_1:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; AVX2-NEXT: vpmovzxbw {{.*#+}} ymm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero,xmm2[8],zero,xmm2[9],zero,xmm2[10],zero,xmm2[11],zero,xmm2[12],zero,xmm2[13],zero,xmm2[14],zero,xmm2[15],zero
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
|
|
; AVX2-NEXT: vpsrlw $8, %ymm2, %ymm2
|
|
; AVX2-NEXT: vextracti128 $1, %ymm2, %xmm3
|
|
; AVX2-NEXT: vpackuswb %xmm3, %xmm2, %xmm2
|
|
; AVX2-NEXT: vpaddb %xmm2, %xmm0, %xmm2
|
|
; AVX2-NEXT: vpmovsxbw %xmm2, %ymm2
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
|
|
; AVX2-NEXT: vpsrlw $8, %ymm2, %ymm2
|
|
; AVX2-NEXT: vextracti128 $1, %ymm2, %xmm3
|
|
; AVX2-NEXT: vpackuswb %xmm3, %xmm2, %xmm2
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX2-NEXT: vpblendvb %xmm3, %xmm2, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; AVX2-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: vzeroupper
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: non_splat_minus_one_divisor_1:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; AVX512F-NEXT: vpmovzxbd {{.*#+}} zmm2 = xmm2[0],zero,zero,zero,xmm2[1],zero,zero,zero,xmm2[2],zero,zero,zero,xmm2[3],zero,zero,zero,xmm2[4],zero,zero,zero,xmm2[5],zero,zero,zero,xmm2[6],zero,zero,zero,xmm2[7],zero,zero,zero,xmm2[8],zero,zero,zero,xmm2[9],zero,zero,zero,xmm2[10],zero,zero,zero,xmm2[11],zero,zero,zero,xmm2[12],zero,zero,zero,xmm2[13],zero,zero,zero,xmm2[14],zero,zero,zero,xmm2[15],zero,zero,zero
|
|
; AVX512F-NEXT: vpsrlvd {{.*}}(%rip), %zmm2, %zmm2
|
|
; AVX512F-NEXT: vpmovdb %zmm2, %xmm2
|
|
; AVX512F-NEXT: vpaddb %xmm2, %xmm0, %xmm2
|
|
; AVX512F-NEXT: vpmovsxbd %xmm2, %zmm2
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %zmm2, %zmm2
|
|
; AVX512F-NEXT: vpmovdb %zmm2, %xmm2
|
|
; AVX512F-NEXT: vmovdqa {{.*#+}} xmm3 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; AVX512F-NEXT: vpblendvb %xmm3, %xmm2, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; AVX512F-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: non_splat_minus_one_divisor_1:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; AVX512BW-NEXT: vpmovzxbw {{.*#+}} ymm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero,xmm2[8],zero,xmm2[9],zero,xmm2[10],zero,xmm2[11],zero,xmm2[12],zero,xmm2[13],zero,xmm2[14],zero,xmm2[15],zero
|
|
; AVX512BW-NEXT: vpsrlvw {{.*}}(%rip), %ymm2, %ymm2
|
|
; AVX512BW-NEXT: vpmovwb %ymm2, %xmm2
|
|
; AVX512BW-NEXT: vpaddb %xmm2, %xmm0, %xmm2
|
|
; AVX512BW-NEXT: vpmovsxbw %xmm2, %ymm2
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %ymm2, %ymm2
|
|
; AVX512BW-NEXT: vpmovwb %ymm2, %xmm2
|
|
; AVX512BW-NEXT: movw $443, %ax # imm = 0x1BB
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqu8 %xmm0, %xmm2 {%k1}
|
|
; AVX512BW-NEXT: vpsubb %xmm2, %xmm1, %xmm0
|
|
; AVX512BW-NEXT: movw $24132, %ax # imm = 0x5E44
|
|
; AVX512BW-NEXT: kmovd %eax, %k1
|
|
; AVX512BW-NEXT: vmovdqu8 %xmm2, %xmm0 {%k1}
|
|
; AVX512BW-NEXT: vzeroupper
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: non_splat_minus_one_divisor_1:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpcmpgtb %xmm0, %xmm1, %xmm2
|
|
; XOP-NEXT: vpshlb {{.*}}(%rip), %xmm2, %xmm2
|
|
; XOP-NEXT: vpaddb %xmm2, %xmm0, %xmm2
|
|
; XOP-NEXT: vpshab {{.*}}(%rip), %xmm2, %xmm2
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm3 = [0,0,255,0,0,0,255,0,0,255,255,255,255,255,255,255]
|
|
; XOP-NEXT: vpblendvb %xmm3, %xmm2, %xmm0, %xmm0
|
|
; XOP-NEXT: vpsubb %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,0,255,255,255,0,255,255,0,0,0,0,255,0,255]
|
|
; XOP-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%div = sdiv <16 x i8> %A, <i8 -1, i8 -1, i8 2, i8 -1, i8 -1, i8 -1, i8 2, i8 -1, i8 -1, i8 2, i8 2, i8 2, i8 2, i8 -128, i8 2, i8 -128>
|
|
ret <16 x i8> %div
|
|
}
|
|
|
|
define <4 x i32> @non_splat_minus_one_divisor_2(<4 x i32> %A) {
|
|
; SSE2-LABEL: non_splat_minus_one_divisor_2:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psrld $31, %xmm2
|
|
; SSE2-NEXT: xorpd %xmm1, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm1[0],xmm2[1]
|
|
; SSE2-NEXT: paddd %xmm0, %xmm2
|
|
; SSE2-NEXT: psrad $1, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
|
|
; SSE2-NEXT: psubd %xmm2, %xmm1
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm2[1,2]
|
|
; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2,3,1]
|
|
; SSE2-NEXT: movaps %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: non_splat_minus_one_divisor_2:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psrld $31, %xmm1
|
|
; SSE41-NEXT: pxor %xmm2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: paddd %xmm0, %xmm1
|
|
; SSE41-NEXT: psrad $1, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: psubd %xmm1, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3,4,5],xmm2[6,7]
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: non_splat_minus_one_divisor_2:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrld $31, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsrad $1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpsubd %xmm0, %xmm2, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: non_splat_minus_one_divisor_2:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsrlvd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
|
|
; AVX2ORLATER-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsubd %xmm0, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm1[0],xmm0[1,2],xmm1[3]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: non_splat_minus_one_divisor_2:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpsrad $31, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshld {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpaddd %xmm1, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshad {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsubd %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7]
|
|
; XOP-NEXT: retq
|
|
%div = sdiv <4 x i32> %A, <i32 -1, i32 1, i32 2, i32 -2>
|
|
ret <4 x i32> %div
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform(<8 x i16> %x) {
|
|
; SSE-LABEL: combine_vec_sdiv_nonuniform:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE-NEXT: psrlw $15, %xmm1
|
|
; SSE-NEXT: paddw %xmm0, %xmm1
|
|
; SSE-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX-LABEL: combine_vec_sdiv_nonuniform:
|
|
; AVX: # %bb.0:
|
|
; AVX-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 3, i16 3, i16 3, i16 3, i16 22, i16 22, i16 22, i16 22>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform2(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: psraw $2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psraw $1, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm1[0],xmm2[1]
|
|
; SSE2-NEXT: psrlw $15, %xmm0
|
|
; SSE2-NEXT: paddw %xmm2, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: psraw $1, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psraw $2, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; SSE41-NEXT: psrlw $15, %xmm0
|
|
; SSE41-NEXT: paddw %xmm2, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpsraw $1, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsraw $2, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2-NEXT: vpsraw $1, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpsraw $2, %xmm0, %xmm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpsraw $1, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpsraw $2, %xmm0, %xmm2
|
|
; AVX512F-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX512F-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform2:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 24, i16 24, i16 24, i16 24, i16 25, i16 25, i16 25, i16 25>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform3(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [45591,45591,45591,45591,32833,32833,32833,32833]
|
|
; SSE2-NEXT: pmulhw %xmm0, %xmm1
|
|
; SSE2-NEXT: paddw %xmm0, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: psraw $4, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
|
|
; SSE2-NEXT: psrlw $15, %xmm1
|
|
; SSE2-NEXT: paddw %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [45591,45591,45591,45591,32833,32833,32833,32833]
|
|
; SSE41-NEXT: pmulhw %xmm0, %xmm1
|
|
; SSE41-NEXT: paddw %xmm0, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: psraw $8, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psraw $4, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: psrlw $15, %xmm1
|
|
; SSE41-NEXT: paddw %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX512F-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX512F-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform3:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; XOP-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 23, i16 23, i16 23, i16 23, i16 511, i16 511, i16 511, i16 511>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform4(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [19945,19945,19945,19945,32639,32639,32639,32639]
|
|
; SSE2-NEXT: pmulhw %xmm0, %xmm1
|
|
; SSE2-NEXT: psubw %xmm0, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: psraw $4, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm2 = xmm0[0],xmm2[1]
|
|
; SSE2-NEXT: psrlw $15, %xmm1
|
|
; SSE2-NEXT: paddw %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [19945,19945,19945,19945,32639,32639,32639,32639]
|
|
; SSE41-NEXT: pmulhw %xmm0, %xmm1
|
|
; SSE41-NEXT: psubw %xmm0, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: psraw $8, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psraw $4, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: psrlw $15, %xmm1
|
|
; SSE41-NEXT: paddw %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsubw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX1-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1,2,3],xmm1[4,5,6,7]
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpsubw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX2-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpsubw %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: vpsraw $8, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpsraw $4, %xmm0, %xmm2
|
|
; AVX512F-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3]
|
|
; AVX512F-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsubw %xmm0, %xmm1, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform4:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpsubw %xmm0, %xmm1, %xmm0
|
|
; XOP-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 -23, i16 -23, i16 -23, i16 -23, i16 -510, i16 -510, i16 -510, i16 -510>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform5(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,65535,0,0,0,1,1]
|
|
; SSE2-NEXT: pmullw %xmm0, %xmm1
|
|
; SSE2-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: paddw %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,65535,65535,65535,65535,65535,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: pand %xmm1, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm3
|
|
; SSE2-NEXT: psraw $8, %xmm3
|
|
; SSE2-NEXT: pandn %xmm3, %xmm1
|
|
; SSE2-NEXT: por %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,0,65535,65535,65535,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $4, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,65535,65535,65535,65535,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: pand %xmm1, %xmm3
|
|
; SSE2-NEXT: psraw $2, %xmm2
|
|
; SSE2-NEXT: pandn %xmm2, %xmm1
|
|
; SSE2-NEXT: por %xmm3, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,65535,65535,65535,0,0,65535]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $1, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: psrlw $15, %xmm0
|
|
; SSE2-NEXT: paddw %xmm2, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,65535,0,0,0,1,1]
|
|
; SSE41-NEXT: pmullw %xmm0, %xmm1
|
|
; SSE41-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: paddw %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = <256,16384,4096,u,u,32768,512,256>
|
|
; SSE41-NEXT: pmulhw %xmm0, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[3,4],xmm1[5,6,7]
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-NEXT: psraw $1, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm1[0,1,2,3,4],xmm2[5],xmm1[6,7]
|
|
; SSE41-NEXT: psrlw $15, %xmm0
|
|
; SSE41-NEXT: paddw %xmm2, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[3,4],xmm1[5,6,7]
|
|
; AVX1-NEXT: vpsraw $1, %xmm0, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3,4],xmm2[5],xmm1[6,7]
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[3,4],xmm1[5,6,7]
|
|
; AVX2-NEXT: vpsraw $1, %xmm0, %xmm2
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3,4],xmm2[5],xmm1[6,7]
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmovsxwd %xmm0, %ymm0
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %ymm0, %ymm0
|
|
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
|
|
; AVX512F-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform5:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpmacsww %xmm1, {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 -510, i16 -24, i16 -23, i16 3, i16 22, i16 25, i16 255, i16 511>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform6(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,65535,65535,65535,1,1,1,0]
|
|
; SSE2-NEXT: pmullw %xmm0, %xmm1
|
|
; SSE2-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: paddw %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,65535,65535,0,0,0,65535,65535]
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: pand %xmm1, %xmm2
|
|
; SSE2-NEXT: pandn %xmm0, %xmm1
|
|
; SSE2-NEXT: por %xmm2, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,65535,65535,65535,65535,0,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $4, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,65535,65535,65535,65535,0,65535,65535]
|
|
; SSE2-NEXT: movdqa %xmm2, %xmm3
|
|
; SSE2-NEXT: pand %xmm1, %xmm3
|
|
; SSE2-NEXT: psraw $2, %xmm2
|
|
; SSE2-NEXT: pandn %xmm2, %xmm1
|
|
; SSE2-NEXT: por %xmm3, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,65535,65535,65535,65535,0,65535,0]
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: pand %xmm2, %xmm3
|
|
; SSE2-NEXT: psraw $1, %xmm1
|
|
; SSE2-NEXT: pandn %xmm1, %xmm2
|
|
; SSE2-NEXT: por %xmm3, %xmm2
|
|
; SSE2-NEXT: psrlw $15, %xmm0
|
|
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: paddw %xmm2, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [65535,65535,65535,65535,1,1,1,0]
|
|
; SSE41-NEXT: pmullw %xmm0, %xmm1
|
|
; SSE41-NEXT: pmulhw {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: paddw %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = <4,256,256,u,u,512,256,8>
|
|
; SSE41-NEXT: pmulhw %xmm0, %xmm2
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2],xmm0[3,4],xmm2[5,6,7]
|
|
; SSE41-NEXT: psrlw $15, %xmm0
|
|
; SSE41-NEXT: pxor %xmm1, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1,2],xmm1[3,4],xmm0[5,6,7]
|
|
; SSE41-NEXT: paddw %xmm2, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[3,4],xmm1[5,6,7]
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2],xmm2[3,4],xmm0[5,6,7]
|
|
; AVX1-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[3,4],xmm1[5,6,7]
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX2-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2],xmm2[3,4],xmm0[5,6,7]
|
|
; AVX2-NEXT: vpaddw %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX512F-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm2[3,4],xmm1[5,6,7]
|
|
; AVX512F-NEXT: vpmovsxwd %xmm0, %ymm0
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %ymm0, %ymm0
|
|
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
|
|
; AVX512F-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX512BW-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm2[3,4],xmm1[5,6,7]
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform6:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpmulhw {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpmacsww %xmm1, {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpsrlw $15, %xmm0, %xmm1
|
|
; XOP-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2],xmm2[3,4],xmm1[5,6,7]
|
|
; XOP-NEXT: vpshaw {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddw %xmm1, %xmm0, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 -32768, i16 -512, i16 -511, i16 -1, i16 1, i16 255, i16 512, i16 32767>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <8 x i16> @combine_vec_sdiv_nonuniform7(<8 x i16> %x) {
|
|
; SSE2-LABEL: combine_vec_sdiv_nonuniform7:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: pxor %xmm1, %xmm1
|
|
; SSE2-NEXT: psubw %xmm0, %xmm1
|
|
; SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: combine_vec_sdiv_nonuniform7:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: pxor %xmm1, %xmm1
|
|
; SSE41-NEXT: psubw %xmm0, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: combine_vec_sdiv_nonuniform7:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsubw %xmm0, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2ORLATER-LABEL: combine_vec_sdiv_nonuniform7:
|
|
; AVX2ORLATER: # %bb.0:
|
|
; AVX2ORLATER-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpsubw %xmm0, %xmm1, %xmm1
|
|
; AVX2ORLATER-NEXT: vpblendd {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
|
|
; AVX2ORLATER-NEXT: retq
|
|
;
|
|
; XOP-LABEL: combine_vec_sdiv_nonuniform7:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpxor %xmm1, %xmm1, %xmm1
|
|
; XOP-NEXT: vpsubw %xmm0, %xmm1, %xmm1
|
|
; XOP-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <8 x i16> %x, <i16 -1, i16 -1, i16 -1, i16 -1, i16 1, i16 1, i16 1, i16 1>
|
|
ret <8 x i16> %1
|
|
}
|
|
|
|
define <16 x i8> @pr38658(<16 x i8> %x) {
|
|
; SSE2-LABEL: pr38658:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: movdqa {{.*#+}} xmm0 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,147]
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
|
|
; SSE2-NEXT: psraw $8, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: pmullw %xmm0, %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: movdqa %xmm1, %xmm3
|
|
; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3],xmm3[4],xmm1[4],xmm3[5],xmm1[5],xmm3[6],xmm1[6],xmm3[7],xmm1[7]
|
|
; SSE2-NEXT: psraw $8, %xmm3
|
|
; SSE2-NEXT: pxor %xmm0, %xmm0
|
|
; SSE2-NEXT: pmullw %xmm3, %xmm0
|
|
; SSE2-NEXT: psrlw $8, %xmm0
|
|
; SSE2-NEXT: packuswb %xmm2, %xmm0
|
|
; SSE2-NEXT: paddb %xmm1, %xmm0
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
|
|
; SSE2-NEXT: psraw $8, %xmm1
|
|
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm1
|
|
; SSE2-NEXT: psrlw $8, %xmm1
|
|
; SSE2-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE2-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]
|
|
; SSE2-NEXT: psraw $8, %xmm2
|
|
; SSE2-NEXT: psllw $8, %xmm2
|
|
; SSE2-NEXT: psrlw $8, %xmm2
|
|
; SSE2-NEXT: packuswb %xmm1, %xmm2
|
|
; SSE2-NEXT: psrlw $7, %xmm0
|
|
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE2-NEXT: paddb %xmm2, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE41-LABEL: pr38658:
|
|
; SSE41: # %bb.0:
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
|
|
; SSE41-NEXT: pmovsxbw %xmm0, %xmm2
|
|
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
|
|
; SSE41-NEXT: psrlw $8, %xmm2
|
|
; SSE41-NEXT: pxor %xmm0, %xmm0
|
|
; SSE41-NEXT: packuswb %xmm2, %xmm0
|
|
; SSE41-NEXT: paddb %xmm1, %xmm0
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm1
|
|
; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
|
|
; SSE41-NEXT: psraw $8, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm1, %xmm2
|
|
; SSE41-NEXT: psllw $6, %xmm2
|
|
; SSE41-NEXT: psllw $8, %xmm1
|
|
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,5,6],xmm2[7]
|
|
; SSE41-NEXT: psrlw $8, %xmm1
|
|
; SSE41-NEXT: movdqa %xmm0, %xmm2
|
|
; SSE41-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]
|
|
; SSE41-NEXT: psraw $8, %xmm2
|
|
; SSE41-NEXT: psllw $8, %xmm2
|
|
; SSE41-NEXT: psrlw $8, %xmm2
|
|
; SSE41-NEXT: packuswb %xmm1, %xmm2
|
|
; SSE41-NEXT: psrlw $7, %xmm0
|
|
; SSE41-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: pand {{.*}}(%rip), %xmm0
|
|
; SSE41-NEXT: paddb %xmm2, %xmm0
|
|
; SSE41-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: pr38658:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
|
|
; AVX1-NEXT: vpmovsxbw %xmm1, %xmm1
|
|
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
|
|
; AVX1-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
|
|
; AVX1-NEXT: vpsraw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpsllw $6, %xmm1, %xmm2
|
|
; AVX1-NEXT: vpsllw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,5,6],xmm2[7]
|
|
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
|
|
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm2 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
|
|
; AVX1-NEXT: vpsraw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsllw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
|
|
; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
|
|
; AVX1-NEXT: vpsrlw $7, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX1-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: pr38658:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpmovsxbw %xmm0, %ymm1
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
|
; AVX2-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: vpmovsxbw %xmm0, %ymm1
|
|
; AVX2-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX2-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
|
|
; AVX2-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpsrlw $7, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; AVX2-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX2-NEXT: vzeroupper
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512F-LABEL: pr38658:
|
|
; AVX512F: # %bb.0:
|
|
; AVX512F-NEXT: vpmovsxbw %xmm0, %ymm1
|
|
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX512F-NEXT: vextracti128 $1, %ymm1, %xmm2
|
|
; AVX512F-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX512F-NEXT: vpsrlw $7, %xmm0, %xmm1
|
|
; AVX512F-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512F-NEXT: vpmovsxbd %xmm0, %zmm0
|
|
; AVX512F-NEXT: vpsravd {{.*}}(%rip), %zmm0, %zmm0
|
|
; AVX512F-NEXT: vpmovdb %zmm0, %xmm0
|
|
; AVX512F-NEXT: vpaddb %xmm1, %xmm0, %xmm0
|
|
; AVX512F-NEXT: vzeroupper
|
|
; AVX512F-NEXT: retq
|
|
;
|
|
; AVX512BW-LABEL: pr38658:
|
|
; AVX512BW: # %bb.0:
|
|
; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm1
|
|
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpsrlw $8, %ymm1, %ymm1
|
|
; AVX512BW-NEXT: vpmovwb %ymm1, %xmm1
|
|
; AVX512BW-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; AVX512BW-NEXT: vpsrlw $7, %xmm0, %xmm1
|
|
; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
|
|
; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm0
|
|
; AVX512BW-NEXT: vpsravw {{.*}}(%rip), %ymm0, %ymm0
|
|
; AVX512BW-NEXT: vpmovwb %ymm0, %xmm0
|
|
; AVX512BW-NEXT: vpaddb %xmm1, %xmm0, %xmm0
|
|
; AVX512BW-NEXT: vzeroupper
|
|
; AVX512BW-NEXT: retq
|
|
;
|
|
; XOP-LABEL: pr38658:
|
|
; XOP: # %bb.0:
|
|
; XOP-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
|
|
; XOP-NEXT: vpmovsxbw %xmm1, %xmm1
|
|
; XOP-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
|
|
; XOP-NEXT: vpxor %xmm2, %xmm2, %xmm2
|
|
; XOP-NEXT: vpperm {{.*#+}} xmm1 = xmm2[1,3,5,7,9,11,13,15],xmm1[1,3,5,7,9,11,13,15]
|
|
; XOP-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; XOP-NEXT: vpshab {{.*}}(%rip), %xmm0, %xmm1
|
|
; XOP-NEXT: vpshlb {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
|
|
; XOP-NEXT: vpaddb %xmm0, %xmm1, %xmm0
|
|
; XOP-NEXT: retq
|
|
%1 = sdiv <16 x i8> %x, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 7>
|
|
ret <16 x i8> %1
|
|
}
|
|
|
|
define i1 @bool_sdiv(i1 %x, i1 %y) {
|
|
; CHECK-LABEL: bool_sdiv:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movl %edi, %eax
|
|
; CHECK-NEXT: # kill: def $al killed $al killed $eax
|
|
; CHECK-NEXT: retq
|
|
%r = sdiv i1 %x, %y
|
|
ret i1 %r
|
|
}
|
|
|
|
define <4 x i1> @boolvec_sdiv(<4 x i1> %x, <4 x i1> %y) {
|
|
; CHECK-LABEL: boolvec_sdiv:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: retq
|
|
%r = sdiv <4 x i1> %x, %y
|
|
ret <4 x i1> %r
|
|
}
|