[llvm] 4f95f90 - [X86] Remove unused check-prefixes from vector funnel shift tests
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Tue Nov 10 07:45:57 PST 2020
Author: Simon Pilgrim
Date: 2020-11-10T15:45:37Z
New Revision: 4f95f90b9445d56ba889a6f427cb20b24c2bb66a
URL: https://github.com/llvm/llvm-project/commit/4f95f90b9445d56ba889a6f427cb20b24c2bb66a
DIFF: https://github.com/llvm/llvm-project/commit/4f95f90b9445d56ba889a6f427cb20b24c2bb66a.diff
LOG: [X86] Remove unused check-prefixes from vector funnel shift tests
Added:
Modified:
llvm/test/CodeGen/X86/vector-fshl-128.ll
llvm/test/CodeGen/X86/vector-fshl-256.ll
llvm/test/CodeGen/X86/vector-fshl-512.ll
llvm/test/CodeGen/X86/vector-fshl-rot-128.ll
llvm/test/CodeGen/X86/vector-fshl-rot-256.ll
llvm/test/CodeGen/X86/vector-fshl-rot-sub128.ll
llvm/test/CodeGen/X86/vector-fshr-128.ll
llvm/test/CodeGen/X86/vector-fshr-256.ll
llvm/test/CodeGen/X86/vector-fshr-512.ll
llvm/test/CodeGen/X86/vector-fshr-rot-128.ll
llvm/test/CodeGen/X86/vector-fshr-rot-256.ll
llvm/test/CodeGen/X86/vector-fshr-rot-sub128.ll
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/X86/vector-fshl-128.ll b/llvm/test/CodeGen/X86/vector-fshl-128.ll
index a98dfa8f11c1..d760ebdd1aa1 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-128.ll
@@ -3,17 +3,17 @@
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i64> @llvm.fshl.v2i64(<2 x i64>, <2 x i64>, <2 x i64>)
declare <4 x i32> @llvm.fshl.v4i32(<4 x i32>, <4 x i32>, <4 x i32>)
@@ -177,25 +177,25 @@ define <2 x i64> @var_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %amt)
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: pandn %xmm3, %xmm4
-; X32-SSE-NEXT: psrlq $1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllq %xmm2, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: pandn %xmm3, %xmm4
+; X86-SSE2-NEXT: psrlq $1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllq %xmm2, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %amt)
ret <2 x i64> %res
}
@@ -385,40 +385,40 @@ define <4 x i32> @var_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %amt)
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [31,31,31,31]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm5
-; X32-SSE-NEXT: pandn %xmm4, %xmm5
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: psrld $1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm6
-; X32-SSE-NEXT: psrld %xmm3, %xmm6
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm7 = xmm5[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-NEXT: psrld %xmm7, %xmm3
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm6[0]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm5[2,3,2,3]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm6 = xmm5[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm7
-; X32-SSE-NEXT: psrld %xmm6, %xmm7
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: psrld %xmm5, %xmm1
-; X32-SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm7[1]
-; X32-SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,3],xmm1[0,3]
-; X32-SSE-NEXT: pand %xmm4, %xmm2
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [31,31,31,31]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm5
+; X86-SSE2-NEXT: pandn %xmm4, %xmm5
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: psrld $1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm6
+; X86-SSE2-NEXT: psrld %xmm3, %xmm6
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm7 = xmm5[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-SSE2-NEXT: psrld %xmm7, %xmm3
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm6[0]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm5[2,3,2,3]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm6 = xmm5[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm7
+; X86-SSE2-NEXT: psrld %xmm6, %xmm7
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: psrld %xmm5, %xmm1
+; X86-SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm7[1]
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,3],xmm1[0,3]
+; X86-SSE2-NEXT: pand %xmm4, %xmm2
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %amt)
ret <4 x i32> %res
}
@@ -673,64 +673,64 @@ define <8 x i16> @var_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %amt)
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pcmpeqd %xmm4, %xmm4
-; X32-SSE-NEXT: pxor %xmm2, %xmm4
-; X32-SSE-NEXT: psllw $12, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm3
-; X32-SSE-NEXT: psraw $15, %xmm3
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $1, %xmm5
-; X32-SSE-NEXT: psrlw $9, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pandn %xmm5, %xmm3
-; X32-SSE-NEXT: por %xmm1, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm1
-; X32-SSE-NEXT: psraw $15, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psrlw $4, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: por %xmm5, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm1
-; X32-SSE-NEXT: psraw $15, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psrlw $2, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: por %xmm5, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: psraw $15, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm5
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psrlw $1, %xmm3
-; X32-SSE-NEXT: pand %xmm4, %xmm3
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm1
-; X32-SSE-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [1065353216,1065353216,1065353216,1065353216]
-; X32-SSE-NEXT: paddd %xmm4, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm6 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd %xmm4, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm6[0]
-; X32-SSE-NEXT: pmullw %xmm0, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: por %xmm3, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pcmpeqd %xmm4, %xmm4
+; X86-SSE2-NEXT: pxor %xmm2, %xmm4
+; X86-SSE2-NEXT: psllw $12, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm3
+; X86-SSE2-NEXT: psraw $15, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $1, %xmm5
+; X86-SSE2-NEXT: psrlw $9, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pandn %xmm5, %xmm3
+; X86-SSE2-NEXT: por %xmm1, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm1
+; X86-SSE2-NEXT: psraw $15, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psrlw $4, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: por %xmm5, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm1
+; X86-SSE2-NEXT: psraw $15, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psrlw $2, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: por %xmm5, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: psraw $15, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm5
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psrlw $1, %xmm3
+; X86-SSE2-NEXT: pand %xmm4, %xmm3
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm1
+; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [1065353216,1065353216,1065353216,1065353216]
+; X86-SSE2-NEXT: paddd %xmm4, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd %xmm4, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm6[0]
+; X86-SSE2-NEXT: pmullw %xmm0, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: por %xmm3, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %amt)
ret <8 x i16> %res
}
@@ -1007,70 +1007,70 @@ define <16 x i8> @var_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %amt)
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm5 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127]
-; X32-SSE-NEXT: pand %xmm5, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm6
-; X32-SSE-NEXT: pandn %xmm4, %xmm6
-; X32-SSE-NEXT: psllw $5, %xmm6
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: pcmpgtb %xmm6, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm7
-; X32-SSE-NEXT: pandn %xmm1, %xmm7
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm7, %xmm1
-; X32-SSE-NEXT: paddb %xmm6, %xmm6
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: pcmpgtb %xmm6, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm7
-; X32-SSE-NEXT: pandn %xmm1, %xmm7
-; X32-SSE-NEXT: psrlw $2, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm7, %xmm1
-; X32-SSE-NEXT: paddb %xmm6, %xmm6
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: pcmpgtb %xmm6, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm6
-; X32-SSE-NEXT: pand %xmm5, %xmm3
-; X32-SSE-NEXT: pandn %xmm1, %xmm6
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: por %xmm6, %xmm3
-; X32-SSE-NEXT: pand %xmm4, %xmm2
-; X32-SSE-NEXT: psllw $5, %xmm2
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm4
-; X32-SSE-NEXT: pandn %xmm0, %xmm4
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand %xmm1, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm4, %xmm0
-; X32-SSE-NEXT: paddb %xmm2, %xmm2
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm4
-; X32-SSE-NEXT: pandn %xmm0, %xmm4
-; X32-SSE-NEXT: psllw $2, %xmm0
-; X32-SSE-NEXT: pand %xmm1, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm4, %xmm0
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: paddb %xmm2, %xmm2
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: pandn %xmm0, %xmm2
-; X32-SSE-NEXT: por %xmm3, %xmm2
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: pand %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm5 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127]
+; X86-SSE2-NEXT: pand %xmm5, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm6
+; X86-SSE2-NEXT: pandn %xmm4, %xmm6
+; X86-SSE2-NEXT: psllw $5, %xmm6
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: pcmpgtb %xmm6, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm7
+; X86-SSE2-NEXT: pandn %xmm1, %xmm7
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm7, %xmm1
+; X86-SSE2-NEXT: paddb %xmm6, %xmm6
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: pcmpgtb %xmm6, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm7
+; X86-SSE2-NEXT: pandn %xmm1, %xmm7
+; X86-SSE2-NEXT: psrlw $2, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm7, %xmm1
+; X86-SSE2-NEXT: paddb %xmm6, %xmm6
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: pcmpgtb %xmm6, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm6
+; X86-SSE2-NEXT: pand %xmm5, %xmm3
+; X86-SSE2-NEXT: pandn %xmm1, %xmm6
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: por %xmm6, %xmm3
+; X86-SSE2-NEXT: pand %xmm4, %xmm2
+; X86-SSE2-NEXT: psllw $5, %xmm2
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm4
+; X86-SSE2-NEXT: pandn %xmm0, %xmm4
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand %xmm1, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm4, %xmm0
+; X86-SSE2-NEXT: paddb %xmm2, %xmm2
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm4
+; X86-SSE2-NEXT: pandn %xmm0, %xmm4
+; X86-SSE2-NEXT: psllw $2, %xmm0
+; X86-SSE2-NEXT: pand %xmm1, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm4, %xmm0
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: paddb %xmm2, %xmm2
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: pandn %xmm0, %xmm2
+; X86-SSE2-NEXT: por %xmm3, %xmm2
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: pand %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %amt)
ret <16 x i8> %res
}
@@ -1231,26 +1231,26 @@ define <2 x i64> @splatvar_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,1,0,1]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: pandn %xmm3, %xmm4
-; X32-SSE-NEXT: psrlq $1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllq %xmm2, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,1,0,1]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: pandn %xmm3, %xmm4
+; X86-SSE2-NEXT: psrlq $1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllq %xmm2, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i64> %amt, <2 x i64> undef, <2 x i32> zeroinitializer
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %splat)
ret <2 x i64> %res
@@ -1447,32 +1447,32 @@ define <4 x i32> @splatvar_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: psrld $1, %xmm1
-; X32-SSE-NEXT: pandn {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrld %xmm4, %xmm5
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm6
-; X32-SSE-NEXT: psrld %xmm4, %xmm6
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm6 = xmm6[0],xmm5[0]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrld %xmm4, %xmm5
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: psrld %xmm3, %xmm1
-; X32-SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm5[1]
-; X32-SSE-NEXT: shufps {{.*#+}} xmm6 = xmm6[0,3],xmm1[0,3]
-; X32-SSE-NEXT: movd %xmm2, %eax
-; X32-SSE-NEXT: andl $31, %eax
-; X32-SSE-NEXT: movd %eax, %xmm1
-; X32-SSE-NEXT: pslld %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm6, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: psrld $1, %xmm1
+; X86-SSE2-NEXT: pandn {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrld %xmm4, %xmm5
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm6
+; X86-SSE2-NEXT: psrld %xmm4, %xmm6
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm6 = xmm6[0],xmm5[0]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm3[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrld %xmm4, %xmm5
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: psrld %xmm3, %xmm1
+; X86-SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm5[1]
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm6 = xmm6[0,3],xmm1[0,3]
+; X86-SSE2-NEXT: movd %xmm2, %eax
+; X86-SSE2-NEXT: andl $31, %eax
+; X86-SSE2-NEXT: movd %eax, %xmm1
+; X86-SSE2-NEXT: pslld %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm6, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <4 x i32> %amt, <4 x i32> undef, <4 x i32> zeroinitializer
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %splat)
ret <4 x i32> %res
@@ -1719,50 +1719,50 @@ define <8 x i16> @splatvar_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pcmpeqd %xmm3, %xmm3
-; X32-SSE-NEXT: pxor %xmm2, %xmm3
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,0,0]
-; X32-SSE-NEXT: psllw $12, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm3
-; X32-SSE-NEXT: psraw $15, %xmm3
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $1, %xmm5
-; X32-SSE-NEXT: psrlw $9, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pandn %xmm5, %xmm3
-; X32-SSE-NEXT: por %xmm1, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm1
-; X32-SSE-NEXT: psraw $15, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psrlw $4, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: por %xmm5, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm1
-; X32-SSE-NEXT: psraw $15, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psrlw $2, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: por %xmm5, %xmm3
-; X32-SSE-NEXT: paddw %xmm4, %xmm4
-; X32-SSE-NEXT: psraw $15, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm1
-; X32-SSE-NEXT: pandn %xmm3, %xmm1
-; X32-SSE-NEXT: psrlw $1, %xmm3
-; X32-SSE-NEXT: pand %xmm4, %xmm3
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psllw %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pcmpeqd %xmm3, %xmm3
+; X86-SSE2-NEXT: pxor %xmm2, %xmm3
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,0,0]
+; X86-SSE2-NEXT: psllw $12, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm3
+; X86-SSE2-NEXT: psraw $15, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $1, %xmm5
+; X86-SSE2-NEXT: psrlw $9, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pandn %xmm5, %xmm3
+; X86-SSE2-NEXT: por %xmm1, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm1
+; X86-SSE2-NEXT: psraw $15, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psrlw $4, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: por %xmm5, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm1
+; X86-SSE2-NEXT: psraw $15, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psrlw $2, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: por %xmm5, %xmm3
+; X86-SSE2-NEXT: paddw %xmm4, %xmm4
+; X86-SSE2-NEXT: psraw $15, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm1
+; X86-SSE2-NEXT: pandn %xmm3, %xmm1
+; X86-SSE2-NEXT: psrlw $1, %xmm3
+; X86-SSE2-NEXT: pand %xmm4, %xmm3
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psllw %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <8 x i16> %amt, <8 x i16> undef, <8 x i32> zeroinitializer
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %splat)
ret <8 x i16> %res
@@ -2065,57 +2065,57 @@ define <16 x i8> @splatvar_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm3[0,0,0,0]
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127]
-; X32-SSE-NEXT: pand %xmm4, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
-; X32-SSE-NEXT: pandn %xmm3, %xmm5
-; X32-SSE-NEXT: psllw $5, %xmm5
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm5, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm1, %xmm7
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand %xmm6, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm7, %xmm1
-; X32-SSE-NEXT: paddb %xmm5, %xmm5
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm5, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm1, %xmm7
-; X32-SSE-NEXT: psrlw $2, %xmm1
-; X32-SSE-NEXT: pand %xmm6, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm7, %xmm1
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: paddb %xmm5, %xmm5
-; X32-SSE-NEXT: pcmpgtb %xmm5, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm5
-; X32-SSE-NEXT: pand %xmm4, %xmm6
-; X32-SSE-NEXT: pandn %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pand %xmm1, %xmm6
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psllw %xmm2, %xmm0
-; X32-SSE-NEXT: pcmpeqd %xmm1, %xmm1
-; X32-SSE-NEXT: psllw %xmm2, %xmm1
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm0, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: por %xmm6, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm3[0,0,0,0]
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127]
+; X86-SSE2-NEXT: pand %xmm4, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
+; X86-SSE2-NEXT: pandn %xmm3, %xmm5
+; X86-SSE2-NEXT: psllw $5, %xmm5
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm5, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm1, %xmm7
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand %xmm6, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm7, %xmm1
+; X86-SSE2-NEXT: paddb %xmm5, %xmm5
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm5, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm1, %xmm7
+; X86-SSE2-NEXT: psrlw $2, %xmm1
+; X86-SSE2-NEXT: pand %xmm6, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm7, %xmm1
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: paddb %xmm5, %xmm5
+; X86-SSE2-NEXT: pcmpgtb %xmm5, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm5
+; X86-SSE2-NEXT: pand %xmm4, %xmm6
+; X86-SSE2-NEXT: pandn %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pand %xmm1, %xmm6
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psllw %xmm2, %xmm0
+; X86-SSE2-NEXT: pcmpeqd %xmm1, %xmm1
+; X86-SSE2-NEXT: psllw %xmm2, %xmm1
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm0, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: por %xmm6, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <16 x i8> %amt, <16 x i8> undef, <16 x i32> zeroinitializer
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %splat)
ret <16 x i8> %res
@@ -2342,40 +2342,40 @@ define void @sink_splatvar(i32* %p, i32 %shift_amt) {
; XOPAVX2-NEXT: # %bb.2: # %end
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: sink_splatvar:
-; X32-SSE: # %bb.0: # %entry
-; X32-SSE-NEXT: pushl %esi
-; X32-SSE-NEXT: .cfi_def_cfa_offset 8
-; X32-SSE-NEXT: .cfi_offset %esi, -8
-; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-SSE-NEXT: xorl %ecx, %ecx
-; X32-SSE-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-SSE-NEXT: movd %xmm0, %edx
-; X32-SSE-NEXT: andl $31, %edx
-; X32-SSE-NEXT: movl $32, %esi
-; X32-SSE-NEXT: subl %edx, %esi
-; X32-SSE-NEXT: movd %esi, %xmm0
-; X32-SSE-NEXT: movd %edx, %xmm1
-; X32-SSE-NEXT: xorl %edx, %edx
-; X32-SSE-NEXT: .p2align 4, 0x90
-; X32-SSE-NEXT: .LBB8_1: # %loop
-; X32-SSE-NEXT: # =>This Inner Loop Header: Depth=1
-; X32-SSE-NEXT: movdqu (%eax,%ecx,4), %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: psrld %xmm0, %xmm3
-; X32-SSE-NEXT: pslld %xmm1, %xmm2
-; X32-SSE-NEXT: por %xmm3, %xmm2
-; X32-SSE-NEXT: movdqu %xmm2, (%eax,%ecx,4)
-; X32-SSE-NEXT: addl $4, %ecx
-; X32-SSE-NEXT: adcl $0, %edx
-; X32-SSE-NEXT: movl %ecx, %esi
-; X32-SSE-NEXT: xorl $256, %esi # imm = 0x100
-; X32-SSE-NEXT: orl %edx, %esi
-; X32-SSE-NEXT: jne .LBB8_1
-; X32-SSE-NEXT: # %bb.2: # %end
-; X32-SSE-NEXT: popl %esi
-; X32-SSE-NEXT: .cfi_def_cfa_offset 4
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: sink_splatvar:
+; X86-SSE2: # %bb.0: # %entry
+; X86-SSE2-NEXT: pushl %esi
+; X86-SSE2-NEXT: .cfi_def_cfa_offset 8
+; X86-SSE2-NEXT: .cfi_offset %esi, -8
+; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-SSE2-NEXT: xorl %ecx, %ecx
+; X86-SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-SSE2-NEXT: movd %xmm0, %edx
+; X86-SSE2-NEXT: andl $31, %edx
+; X86-SSE2-NEXT: movl $32, %esi
+; X86-SSE2-NEXT: subl %edx, %esi
+; X86-SSE2-NEXT: movd %esi, %xmm0
+; X86-SSE2-NEXT: movd %edx, %xmm1
+; X86-SSE2-NEXT: xorl %edx, %edx
+; X86-SSE2-NEXT: .p2align 4, 0x90
+; X86-SSE2-NEXT: .LBB8_1: # %loop
+; X86-SSE2-NEXT: # =>This Inner Loop Header: Depth=1
+; X86-SSE2-NEXT: movdqu (%eax,%ecx,4), %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: psrld %xmm0, %xmm3
+; X86-SSE2-NEXT: pslld %xmm1, %xmm2
+; X86-SSE2-NEXT: por %xmm3, %xmm2
+; X86-SSE2-NEXT: movdqu %xmm2, (%eax,%ecx,4)
+; X86-SSE2-NEXT: addl $4, %ecx
+; X86-SSE2-NEXT: adcl $0, %edx
+; X86-SSE2-NEXT: movl %ecx, %esi
+; X86-SSE2-NEXT: xorl $256, %esi # imm = 0x100
+; X86-SSE2-NEXT: orl %edx, %esi
+; X86-SSE2-NEXT: jne .LBB8_1
+; X86-SSE2-NEXT: # %bb.2: # %end
+; X86-SSE2-NEXT: popl %esi
+; X86-SSE2-NEXT: .cfi_def_cfa_offset 4
+; X86-SSE2-NEXT: retl
entry:
%ins = insertelement <4 x i32> undef, i32 %shift_amt, i32 0
%splat = shufflevector <4 x i32> %ins, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2499,26 +2499,26 @@ define <2 x i64> @constant_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y) nounwind {
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = <4,u,14,u>
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: pandn %xmm2, %xmm4
-; X32-SSE-NEXT: psrlq $1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pand %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: psllq %xmm3, %xmm2
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm3, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = <4,u,14,u>
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: pandn %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlq $1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pand %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: psllq %xmm3, %xmm2
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm3, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> <i64 4, i64 14>)
ret <2 x i64> %res
}
@@ -2637,28 +2637,28 @@ define <4 x i32> @constant_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y) nounwind {
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: psrld $25, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-NEXT: psrld $26, %xmm3
-; X32-SSE-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: psrld $27, %xmm2
-; X32-SSE-NEXT: psrld $28, %xmm1
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; X32-SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm3[0,3]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [16,32,64,128]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm3, %xmm2
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: psrld $25, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-SSE2-NEXT: psrld $26, %xmm3
+; X86-SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: psrld $27, %xmm2
+; X86-SSE2-NEXT: psrld $28, %xmm1
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm3[0,3]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [16,32,64,128]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm3, %xmm2
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
ret <4 x i32> %res
}
@@ -2741,13 +2741,13 @@ define <8 x i16> @constant_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y) nounwind {
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pmulhuw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pmulhuw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
ret <8 x i16> %res
}
@@ -2935,30 +2935,30 @@ define <16 x i8> @constant_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y) nounwind {
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: pand %xmm3, %xmm0
-; X32-SSE-NEXT: packuswb %xmm2, %xmm0
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: psrlw $8, %xmm3
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psrlw $8, %xmm1
-; X32-SSE-NEXT: packuswb %xmm3, %xmm1
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: pand %xmm3, %xmm0
+; X86-SSE2-NEXT: packuswb %xmm2, %xmm0
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: psrlw $8, %xmm3
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psrlw $8, %xmm1
+; X86-SSE2-NEXT: packuswb %xmm3, %xmm1
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1>)
ret <16 x i8> %res
}
@@ -3029,14 +3029,14 @@ define <2 x i64> @splatconstant_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlq $50, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
-; X32-SSE-NEXT: psllq $14, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlq $50, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
+; X86-SSE2-NEXT: psllq $14, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> <i64 14, i64 14>)
ret <2 x i64> %res
}
@@ -3103,12 +3103,12 @@ define <4 x i32> @splatconstant_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrld $28, %xmm1
-; X32-SSE-NEXT: pslld $4, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrld $28, %xmm1
+; X86-SSE2-NEXT: pslld $4, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> <i32 4, i32 4, i32 4, i32 4>)
ret <4 x i32> %res
}
@@ -3175,12 +3175,12 @@ define <8 x i16> @splatconstant_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlw $9, %xmm1
-; X32-SSE-NEXT: psllw $7, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlw $9, %xmm1
+; X86-SSE2-NEXT: psllw $7, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
ret <8 x i16> %res
}
@@ -3259,14 +3259,14 @@ define <16 x i8> @splatconstant_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>)
ret <16 x i8> %res
}
diff --git a/llvm/test/CodeGen/X86/vector-fshl-256.ll b/llvm/test/CodeGen/X86/vector-fshl-256.ll
index 3c621b6aeac4..3ca4333cbc0b 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-256.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-256.ll
@@ -1,14 +1,14 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOPAVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOPAVX2
declare <4 x i64> @llvm.fshl.v4i64(<4 x i64>, <4 x i64>, <4 x i64>)
declare <8 x i32> @llvm.fshl.v8i32(<8 x i32>, <8 x i32>, <8 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshl-512.ll b/llvm/test/CodeGen/X86/vector-fshl-512.ll
index 62660f3186ff..4e3fcca69514 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-512.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-512.ll
@@ -1,10 +1,10 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
declare <8 x i64> @llvm.fshl.v8i64(<8 x i64>, <8 x i64>, <8 x i64>)
declare <16 x i32> @llvm.fshl.v16i32(<16 x i32>, <16 x i32>, <16 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshl-rot-128.ll b/llvm/test/CodeGen/X86/vector-fshl-rot-128.ll
index 231b314fa9f4..96c3759956cd 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-rot-128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-rot-128.ll
@@ -11,7 +11,7 @@
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i64> @llvm.fshl.v2i64(<2 x i64>, <2 x i64>, <2 x i64>)
declare <4 x i32> @llvm.fshl.v4i32(<4 x i32>, <4 x i32>, <4 x i32>)
@@ -128,26 +128,26 @@ define <2 x i64> @var_funnnel_v2i64(<2 x i64> %x, <2 x i64> %amt) nounwind {
; XOP-NEXT: vprotq %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm1, %xmm3
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psllq %xmm1, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psllq %xmm1, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm1, %xmm3
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psllq %xmm1, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psllq %xmm1, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> %amt)
ret <2 x i64> %res
}
@@ -250,24 +250,24 @@ define <4 x i32> @var_funnnel_v4i32(<4 x i32> %x, <4 x i32> %amt) nounwind {
; XOP-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> %amt)
ret <4 x i32> %res
}
@@ -412,31 +412,31 @@ define <8 x i16> @var_funnnel_v8i16(<8 x i16> %x, <8 x i16> %amt) nounwind {
; XOP-NEXT: vprotw %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: punpckhwd {{.*#+}} xmm2 = xmm2[4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [1065353216,1065353216,1065353216,1065353216]
-; X32-SSE-NEXT: paddd %xmm3, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm2
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X32-SSE-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3]
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: paddd %xmm3, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pmulhuw %xmm1, %xmm2
-; X32-SSE-NEXT: pmullw %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm2 = xmm2[4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [1065353216,1065353216,1065353216,1065353216]
+; X86-SSE2-NEXT: paddd %xmm3, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm2
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3]
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: paddd %xmm3, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pmulhuw %xmm1, %xmm2
+; X86-SSE2-NEXT: pmullw %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> %amt)
ret <8 x i16> %res
}
@@ -618,48 +618,48 @@ define <16 x i8> @var_funnnel_v16i8(<16 x i8> %x, <16 x i8> %amt) nounwind {
; XOP-NEXT: vprotb %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: psllw $5, %xmm1
-; X32-SSE-NEXT: pxor %xmm0, %xmm0
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: pcmpgtb %xmm1, %xmm3
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: psrlw $4, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: movdqa %xmm2, %xmm5
-; X32-SSE-NEXT: psllw $4, %xmm5
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm5
-; X32-SSE-NEXT: por %xmm4, %xmm5
-; X32-SSE-NEXT: pand %xmm3, %xmm5
-; X32-SSE-NEXT: pandn %xmm2, %xmm3
-; X32-SSE-NEXT: por %xmm5, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm2
-; X32-SSE-NEXT: psrlw $6, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: psllw $2, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: por %xmm2, %xmm4
-; X32-SSE-NEXT: paddb %xmm1, %xmm1
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: pcmpgtb %xmm1, %xmm2
-; X32-SSE-NEXT: pand %xmm2, %xmm4
-; X32-SSE-NEXT: pandn %xmm3, %xmm2
-; X32-SSE-NEXT: por %xmm4, %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: paddb %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: psrlw $7, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: por %xmm3, %xmm4
-; X32-SSE-NEXT: paddb %xmm1, %xmm1
-; X32-SSE-NEXT: pcmpgtb %xmm1, %xmm0
-; X32-SSE-NEXT: pand %xmm0, %xmm4
-; X32-SSE-NEXT: pandn %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm4, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: psllw $5, %xmm1
+; X86-SSE2-NEXT: pxor %xmm0, %xmm0
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: pcmpgtb %xmm1, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlw $4, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm5
+; X86-SSE2-NEXT: psllw $4, %xmm5
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm5
+; X86-SSE2-NEXT: por %xmm4, %xmm5
+; X86-SSE2-NEXT: pand %xmm3, %xmm5
+; X86-SSE2-NEXT: pandn %xmm2, %xmm3
+; X86-SSE2-NEXT: por %xmm5, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm2
+; X86-SSE2-NEXT: psrlw $6, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: psllw $2, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: por %xmm2, %xmm4
+; X86-SSE2-NEXT: paddb %xmm1, %xmm1
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: pcmpgtb %xmm1, %xmm2
+; X86-SSE2-NEXT: pand %xmm2, %xmm4
+; X86-SSE2-NEXT: pandn %xmm3, %xmm2
+; X86-SSE2-NEXT: por %xmm4, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: paddb %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlw $7, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: por %xmm3, %xmm4
+; X86-SSE2-NEXT: paddb %xmm1, %xmm1
+; X86-SSE2-NEXT: pcmpgtb %xmm1, %xmm0
+; X86-SSE2-NEXT: pand %xmm0, %xmm4
+; X86-SSE2-NEXT: pandn %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm4, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> %amt)
ret <16 x i8> %res
}
@@ -736,27 +736,27 @@ define <2 x i64> @splatvar_funnnel_v2i64(<2 x i64> %x, <2 x i64> %amt) nounwind
; XOPAVX2-NEXT: vprotq %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,1]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm1, %xmm3
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psllq %xmm1, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psllq %xmm1, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,1]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm1, %xmm3
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psllq %xmm1, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psllq %xmm1, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i64> %amt, <2 x i64> undef, <2 x i32> zeroinitializer
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> %splat)
ret <2 x i64> %res
@@ -857,19 +857,19 @@ define <4 x i32> @splatvar_funnnel_v4i32(<4 x i32> %x, <4 x i32> %amt) nounwind
; XOPAVX2-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movd %xmm1, %eax
-; X32-SSE-NEXT: andl $31, %eax
-; X32-SSE-NEXT: movd %eax, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pslld %xmm1, %xmm2
-; X32-SSE-NEXT: movl $32, %ecx
-; X32-SSE-NEXT: subl %eax, %ecx
-; X32-SSE-NEXT: movd %ecx, %xmm1
-; X32-SSE-NEXT: psrld %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movd %xmm1, %eax
+; X86-SSE2-NEXT: andl $31, %eax
+; X86-SSE2-NEXT: movd %eax, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pslld %xmm1, %xmm2
+; X86-SSE2-NEXT: movl $32, %ecx
+; X86-SSE2-NEXT: subl %eax, %ecx
+; X86-SSE2-NEXT: movd %ecx, %xmm1
+; X86-SSE2-NEXT: psrld %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <4 x i32> %amt, <4 x i32> undef, <4 x i32> zeroinitializer
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> %splat)
ret <4 x i32> %res
@@ -941,20 +941,20 @@ define <8 x i16> @splatvar_funnnel_v8i16(<8 x i16> %x, <8 x i16> %amt) nounwind
; XOPAVX2-NEXT: vprotw %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [65535,0,0,0]
-; X32-SSE-NEXT: pand %xmm1, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllw %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [16,16,16,16,16,16,16,16]
-; X32-SSE-NEXT: psubw %xmm1, %xmm2
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [65535,0,0,0]
+; X86-SSE2-NEXT: pand %xmm1, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllw %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [16,16,16,16,16,16,16,16]
+; X86-SSE2-NEXT: psubw %xmm1, %xmm2
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <8 x i16> %amt, <8 x i16> undef, <8 x i32> zeroinitializer
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> %splat)
ret <8 x i16> %res
@@ -1136,35 +1136,35 @@ define <16 x i8> @splatvar_funnnel_v16i8(<16 x i8> %x, <16 x i8> %amt) nounwind
; XOPAVX2-NEXT: vprotb %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8]
-; X32-SSE-NEXT: psubb %xmm1, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm3 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm3[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm3 = xmm3[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllw %xmm3, %xmm1
-; X32-SSE-NEXT: pcmpeqd %xmm4, %xmm4
-; X32-SSE-NEXT: pcmpeqd %xmm5, %xmm5
-; X32-SSE-NEXT: psllw %xmm3, %xmm5
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm5 = xmm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm2, %xmm0
-; X32-SSE-NEXT: psrlw %xmm2, %xmm4
-; X32-SSE-NEXT: psrlw $8, %xmm4
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm4[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm0, %xmm2
-; X32-SSE-NEXT: por %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8]
+; X86-SSE2-NEXT: psubb %xmm1, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm3 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm3[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm3 = xmm3[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllw %xmm3, %xmm1
+; X86-SSE2-NEXT: pcmpeqd %xmm4, %xmm4
+; X86-SSE2-NEXT: pcmpeqd %xmm5, %xmm5
+; X86-SSE2-NEXT: psllw %xmm3, %xmm5
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm5 = xmm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm0
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlw $8, %xmm4
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm4[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm0, %xmm2
+; X86-SSE2-NEXT: por %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <16 x i8> %amt, <16 x i8> undef, <16 x i32> zeroinitializer
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> %splat)
ret <16 x i8> %res
@@ -1254,27 +1254,27 @@ define <2 x i64> @constant_funnnel_v2i64(<2 x i64> %x) nounwind {
; XOP-NEXT: vprotq {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = <4,u,14,u>
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm2, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psllq %xmm2, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psllq %xmm2, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = <4,u,14,u>
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm2, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psllq %xmm2, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psllq %xmm2, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> <i64 4, i64 14>)
ret <2 x i64> %res
}
@@ -1364,21 +1364,21 @@ define <4 x i32> @constant_funnnel_v4i32(<4 x i32> %x) nounwind {
; XOP-NEXT: vprotd {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [16,32,64,128]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [16,32,64,128]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
ret <4 x i32> %res
}
@@ -1440,14 +1440,14 @@ define <8 x i16> @constant_funnnel_v8i16(<8 x i16> %x) nounwind {
; XOP-NEXT: vprotw {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [1,2,4,8,16,32,64,128]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pmulhuw %xmm1, %xmm2
-; X32-SSE-NEXT: pmullw %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [1,2,4,8,16,32,64,128]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pmulhuw %xmm1, %xmm2
+; X86-SSE2-NEXT: pmullw %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
ret <8 x i16> %res
}
@@ -1583,29 +1583,29 @@ define <16 x i8> @constant_funnnel_v16i8(<16 x i8> %x) nounwind {
; XOP-NEXT: vprotb {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: psrlw $8, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: psrlw $8, %xmm3
-; X32-SSE-NEXT: packuswb %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: pand %xmm2, %xmm0
-; X32-SSE-NEXT: packuswb %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: psrlw $8, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: psrlw $8, %xmm3
+; X86-SSE2-NEXT: packuswb %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: pand %xmm2, %xmm0
+; X86-SSE2-NEXT: packuswb %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1>)
ret <16 x i8> %res
}
@@ -1661,15 +1661,15 @@ define <2 x i64> @splatconstant_funnnel_v2i64(<2 x i64> %x) nounwind {
; XOP-NEXT: vprotq $14, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlq $50, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
-; X32-SSE-NEXT: psllq $14, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlq $50, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
+; X86-SSE2-NEXT: psllq $14, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> <i64 14, i64 14>)
ret <2 x i64> %res
}
@@ -1721,13 +1721,13 @@ define <4 x i32> @splatconstant_funnnel_v4i32(<4 x i32> %x) nounwind {
; XOP-NEXT: vprotd $4, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrld $28, %xmm1
-; X32-SSE-NEXT: pslld $4, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrld $28, %xmm1
+; X86-SSE2-NEXT: pslld $4, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> <i32 4, i32 4, i32 4, i32 4>)
ret <4 x i32> %res
}
@@ -1760,13 +1760,13 @@ define <8 x i16> @splatconstant_funnnel_v8i16(<8 x i16> %x) nounwind {
; XOP-NEXT: vprotw $7, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlw $9, %xmm1
-; X32-SSE-NEXT: psllw $7, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlw $9, %xmm1
+; X86-SSE2-NEXT: psllw $7, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
ret <8 x i16> %res
}
@@ -1828,15 +1828,15 @@ define <16 x i8> @splatconstant_funnnel_v16i8(<16 x i8> %x) nounwind {
; XOP-NEXT: vprotb $4, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>)
ret <16 x i8> %res
}
diff --git a/llvm/test/CodeGen/X86/vector-fshl-rot-256.ll b/llvm/test/CodeGen/X86/vector-fshl-rot-256.ll
index 05bd0e01ec9b..f463b059a052 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-rot-256.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-rot-256.ll
@@ -1,12 +1,12 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOPAVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOPAVX2
declare <4 x i64> @llvm.fshl.v4i64(<4 x i64>, <4 x i64>, <4 x i64>)
declare <8 x i32> @llvm.fshl.v8i32(<8 x i32>, <8 x i32>, <8 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshl-rot-sub128.ll b/llvm/test/CodeGen/X86/vector-fshl-rot-sub128.ll
index d21fd8f53104..a5950ff60ead 100644
--- a/llvm/test/CodeGen/X86/vector-fshl-rot-sub128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshl-rot-sub128.ll
@@ -1,17 +1,17 @@
; 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=SSE,SSE2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE41
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i32> @llvm.fshl.v2i32(<2 x i32>, <2 x i32>, <2 x i32>)
@@ -117,24 +117,24 @@ define <2 x i32> @var_funnnel_v2i32(<2 x i32> %x, <2 x i32> %amt) nounwind {
; XOP-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshl.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> %amt)
ret <2 x i32> %res
}
@@ -256,25 +256,25 @@ define <2 x i32> @splatvar_funnnel_v2i32(<2 x i32> %x, <2 x i32> %amt) nounwind
; XOPAVX2-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i32> %amt, <2 x i32> undef, <2 x i32> zeroinitializer
%res = call <2 x i32> @llvm.fshl.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> %splat)
ret <2 x i32> %res
@@ -369,21 +369,21 @@ define <2 x i32> @constant_funnnel_v2i32(<2 x i32> %x) nounwind {
; XOP-NEXT: vprotd {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [16,32,1,1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [16,32,1,1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshl.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> <i32 4, i32 5>)
ret <2 x i32> %res
}
@@ -461,16 +461,16 @@ define <2 x i32> @splatconstant_funnnel_v2i32(<2 x i32> %x) nounwind {
; XOP-NEXT: vprotd $4, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: psrld $28, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: pslld $4, %xmm1
-; X32-SSE-NEXT: por %xmm2, %xmm1
-; X32-SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3]
-; X32-SSE-NEXT: movaps %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: psrld $28, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: pslld $4, %xmm1
+; X86-SSE2-NEXT: por %xmm2, %xmm1
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3]
+; X86-SSE2-NEXT: movaps %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshl.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> <i32 4, i32 4>)
ret <2 x i32> %res
}
diff --git a/llvm/test/CodeGen/X86/vector-fshr-128.ll b/llvm/test/CodeGen/X86/vector-fshr-128.ll
index 20033b4dbc9b..7fc12f3094c3 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-128.ll
@@ -3,17 +3,17 @@
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i64> @llvm.fshr.v2i64(<2 x i64>, <2 x i64>, <2 x i64>)
declare <4 x i32> @llvm.fshr.v4i32(<4 x i32>, <4 x i32>, <4 x i32>)
@@ -178,25 +178,25 @@ define <2 x i64> @var_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %amt)
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: pand %xmm3, %xmm4
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pandn %xmm3, %xmm2
-; X32-SSE-NEXT: psllq $1, %xmm0
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllq %xmm2, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: pand %xmm3, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pandn %xmm3, %xmm2
+; X86-SSE2-NEXT: psllq $1, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllq %xmm2, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %amt)
ret <2 x i64> %res
}
@@ -387,40 +387,40 @@ define <4 x i32> @var_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %amt)
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [31,31,31,31]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm5
-; X32-SSE-NEXT: pand %xmm4, %xmm5
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm6
-; X32-SSE-NEXT: psrld %xmm3, %xmm6
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm7 = xmm5[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-NEXT: psrld %xmm7, %xmm3
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm6[0]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm5[2,3,2,3]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm6 = xmm5[2,3,3,3,4,5,6,7]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm7
-; X32-SSE-NEXT: psrld %xmm6, %xmm7
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,1,1,1,4,5,6,7]
-; X32-SSE-NEXT: psrld %xmm5, %xmm1
-; X32-SSE-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm7[1]
-; X32-SSE-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,3],xmm1[0,3]
-; X32-SSE-NEXT: pandn %xmm4, %xmm2
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pslld $1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [31,31,31,31]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm5
+; X86-SSE2-NEXT: pand %xmm4, %xmm5
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm5[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm6
+; X86-SSE2-NEXT: psrld %xmm3, %xmm6
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm7 = xmm5[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-SSE2-NEXT: psrld %xmm7, %xmm3
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm6[0]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm5[2,3,2,3]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm6 = xmm5[2,3,3,3,4,5,6,7]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm7
+; X86-SSE2-NEXT: psrld %xmm6, %xmm7
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,1,1,1,4,5,6,7]
+; X86-SSE2-NEXT: psrld %xmm5, %xmm1
+; X86-SSE2-NEXT: punpckhqdq {{.*#+}} xmm1 = xmm1[1],xmm7[1]
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm3 = xmm3[0,3],xmm1[0,3]
+; X86-SSE2-NEXT: pandn %xmm4, %xmm2
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pslld $1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %amt)
ret <4 x i32> %res
}
@@ -674,62 +674,62 @@ define <8 x i16> @var_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %amt)
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: psllw $12, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: psraw $15, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm5
-; X32-SSE-NEXT: pandn %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $8, %xmm1
-; X32-SSE-NEXT: pand %xmm4, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: paddw %xmm3, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: psraw $15, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm5
-; X32-SSE-NEXT: pandn %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand %xmm4, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: paddw %xmm3, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: psraw $15, %xmm4
-; X32-SSE-NEXT: movdqa %xmm4, %xmm5
-; X32-SSE-NEXT: pandn %xmm1, %xmm5
-; X32-SSE-NEXT: psrlw $2, %xmm1
-; X32-SSE-NEXT: pand %xmm4, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: paddw %xmm3, %xmm3
-; X32-SSE-NEXT: psraw $15, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: pandn %xmm1, %xmm4
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pandn {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: punpckhwd {{.*#+}} xmm3 = xmm3[4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pslld $23, %xmm3
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm5 = [1065353216,1065353216,1065353216,1065353216]
-; X32-SSE-NEXT: paddd %xmm5, %xmm3
-; X32-SSE-NEXT: cvttps2dq %xmm3, %xmm3
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
-; X32-SSE-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd %xmm5, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm2
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
-; X32-SSE-NEXT: psllw $1, %xmm0
-; X32-SSE-NEXT: pmullw %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm4, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: psllw $12, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: psraw $15, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm5
+; X86-SSE2-NEXT: pandn %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $8, %xmm1
+; X86-SSE2-NEXT: pand %xmm4, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: paddw %xmm3, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: psraw $15, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm5
+; X86-SSE2-NEXT: pandn %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand %xmm4, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: paddw %xmm3, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: psraw $15, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm4, %xmm5
+; X86-SSE2-NEXT: pandn %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlw $2, %xmm1
+; X86-SSE2-NEXT: pand %xmm4, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: paddw %xmm3, %xmm3
+; X86-SSE2-NEXT: psraw $15, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: pandn %xmm1, %xmm4
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pandn {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm3 = xmm3[4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pslld $23, %xmm3
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm5 = [1065353216,1065353216,1065353216,1065353216]
+; X86-SSE2-NEXT: paddd %xmm5, %xmm3
+; X86-SSE2-NEXT: cvttps2dq %xmm3, %xmm3
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
+; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd %xmm5, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm2
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
+; X86-SSE2-NEXT: psllw $1, %xmm0
+; X86-SSE2-NEXT: pmullw %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm4, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %amt)
ret <8 x i16> %res
}
@@ -996,68 +996,68 @@ define <16 x i8> @var_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %amt)
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm5 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: pandn %xmm5, %xmm4
-; X32-SSE-NEXT: psllw $5, %xmm4
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm6
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm0, %xmm7
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand %xmm6, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm7, %xmm0
-; X32-SSE-NEXT: paddb %xmm4, %xmm4
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm0, %xmm7
-; X32-SSE-NEXT: psllw $2, %xmm0
-; X32-SSE-NEXT: pand %xmm6, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm7, %xmm0
-; X32-SSE-NEXT: paddb %xmm4, %xmm4
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm4
-; X32-SSE-NEXT: pandn %xmm0, %xmm4
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: pand %xmm6, %xmm0
-; X32-SSE-NEXT: pand %xmm5, %xmm2
-; X32-SSE-NEXT: psllw $5, %xmm2
-; X32-SSE-NEXT: pxor %xmm5, %xmm5
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm5
-; X32-SSE-NEXT: movdqa %xmm5, %xmm6
-; X32-SSE-NEXT: pandn %xmm1, %xmm6
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand %xmm5, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm6, %xmm1
-; X32-SSE-NEXT: paddb %xmm2, %xmm2
-; X32-SSE-NEXT: pxor %xmm5, %xmm5
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm5
-; X32-SSE-NEXT: movdqa %xmm5, %xmm6
-; X32-SSE-NEXT: pandn %xmm1, %xmm6
-; X32-SSE-NEXT: psrlw $2, %xmm1
-; X32-SSE-NEXT: pand %xmm5, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm6, %xmm1
-; X32-SSE-NEXT: paddb %xmm2, %xmm2
-; X32-SSE-NEXT: pcmpgtb %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm2
-; X32-SSE-NEXT: pandn %xmm1, %xmm2
-; X32-SSE-NEXT: psrlw $1, %xmm1
-; X32-SSE-NEXT: pand %xmm3, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: por %xmm2, %xmm1
-; X32-SSE-NEXT: por %xmm4, %xmm1
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm5 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: pandn %xmm5, %xmm4
+; X86-SSE2-NEXT: psllw $5, %xmm4
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm6
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm0, %xmm7
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand %xmm6, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm7, %xmm0
+; X86-SSE2-NEXT: paddb %xmm4, %xmm4
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm0, %xmm7
+; X86-SSE2-NEXT: psllw $2, %xmm0
+; X86-SSE2-NEXT: pand %xmm6, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm7, %xmm0
+; X86-SSE2-NEXT: paddb %xmm4, %xmm4
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm4
+; X86-SSE2-NEXT: pandn %xmm0, %xmm4
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: pand %xmm6, %xmm0
+; X86-SSE2-NEXT: pand %xmm5, %xmm2
+; X86-SSE2-NEXT: psllw $5, %xmm2
+; X86-SSE2-NEXT: pxor %xmm5, %xmm5
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm5
+; X86-SSE2-NEXT: movdqa %xmm5, %xmm6
+; X86-SSE2-NEXT: pandn %xmm1, %xmm6
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand %xmm5, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm6, %xmm1
+; X86-SSE2-NEXT: paddb %xmm2, %xmm2
+; X86-SSE2-NEXT: pxor %xmm5, %xmm5
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm5
+; X86-SSE2-NEXT: movdqa %xmm5, %xmm6
+; X86-SSE2-NEXT: pandn %xmm1, %xmm6
+; X86-SSE2-NEXT: psrlw $2, %xmm1
+; X86-SSE2-NEXT: pand %xmm5, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm6, %xmm1
+; X86-SSE2-NEXT: paddb %xmm2, %xmm2
+; X86-SSE2-NEXT: pcmpgtb %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm2
+; X86-SSE2-NEXT: pandn %xmm1, %xmm2
+; X86-SSE2-NEXT: psrlw $1, %xmm1
+; X86-SSE2-NEXT: pand %xmm3, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: por %xmm2, %xmm1
+; X86-SSE2-NEXT: por %xmm4, %xmm1
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %amt)
ret <16 x i8> %res
}
@@ -1217,26 +1217,26 @@ define <2 x i64> @splatvar_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,1,0,1]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: pand %xmm3, %xmm4
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pandn %xmm3, %xmm2
-; X32-SSE-NEXT: psllq $1, %xmm0
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllq %xmm2, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,1,0,1]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: pand %xmm3, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pandn %xmm3, %xmm2
+; X86-SSE2-NEXT: psllq $1, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllq %xmm2, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm3[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i64> %amt, <2 x i64> undef, <2 x i32> zeroinitializer
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> %splat)
ret <2 x i64> %res
@@ -1408,27 +1408,27 @@ define <4 x i32> @splatvar_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: pslld $1, %xmm0
-; X32-SSE-NEXT: pandn {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: pslld $23, %xmm3
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: cvttps2dq %xmm3, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm3, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm4, %xmm3
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
-; X32-SSE-NEXT: movd %xmm2, %eax
-; X32-SSE-NEXT: andl $31, %eax
-; X32-SSE-NEXT: movd %eax, %xmm2
-; X32-SSE-NEXT: psrld %xmm2, %xmm1
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: pslld $1, %xmm0
+; X86-SSE2-NEXT: pandn {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: pslld $23, %xmm3
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: cvttps2dq %xmm3, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm3, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm4, %xmm3
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
+; X86-SSE2-NEXT: movd %xmm2, %eax
+; X86-SSE2-NEXT: andl $31, %eax
+; X86-SSE2-NEXT: movd %eax, %xmm2
+; X86-SSE2-NEXT: psrld %xmm2, %xmm1
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <4 x i32> %amt, <4 x i32> undef, <4 x i32> zeroinitializer
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> %splat)
ret <4 x i32> %res
@@ -1642,37 +1642,37 @@ define <8 x i16> @splatvar_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm2[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm4 = [15,15,15,15,15,15,15,15]
-; X32-SSE-NEXT: pandn %xmm4, %xmm3
-; X32-SSE-NEXT: movdqa %xmm3, %xmm5
-; X32-SSE-NEXT: punpckhwd {{.*#+}} xmm5 = xmm5[4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pslld $23, %xmm5
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm6 = [1065353216,1065353216,1065353216,1065353216]
-; X32-SSE-NEXT: paddd %xmm6, %xmm5
-; X32-SSE-NEXT: cvttps2dq %xmm5, %xmm5
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm5 = xmm5[0,2,2,3]
-; X32-SSE-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3]
-; X32-SSE-NEXT: pslld $23, %xmm3
-; X32-SSE-NEXT: paddd %xmm6, %xmm3
-; X32-SSE-NEXT: cvttps2dq %xmm3, %xmm3
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm5[0]
-; X32-SSE-NEXT: psllw $1, %xmm0
-; X32-SSE-NEXT: pmullw %xmm3, %xmm0
-; X32-SSE-NEXT: pand %xmm4, %xmm2
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm2, %xmm1
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm2[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [15,15,15,15,15,15,15,15]
+; X86-SSE2-NEXT: pandn %xmm4, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm5
+; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm5 = xmm5[4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pslld $23, %xmm5
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm6 = [1065353216,1065353216,1065353216,1065353216]
+; X86-SSE2-NEXT: paddd %xmm6, %xmm5
+; X86-SSE2-NEXT: cvttps2dq %xmm5, %xmm5
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm5 = xmm5[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm5[0,2,2,3]
+; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3]
+; X86-SSE2-NEXT: pslld $23, %xmm3
+; X86-SSE2-NEXT: paddd %xmm6, %xmm3
+; X86-SSE2-NEXT: cvttps2dq %xmm3, %xmm3
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm5[0]
+; X86-SSE2-NEXT: psllw $1, %xmm0
+; X86-SSE2-NEXT: pmullw %xmm3, %xmm0
+; X86-SSE2-NEXT: pand %xmm4, %xmm2
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0,1]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm1
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <8 x i16> %amt, <8 x i16> undef, <8 x i32> zeroinitializer
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> %splat)
ret <8 x i16> %res
@@ -1959,54 +1959,54 @@ define <16 x i8> @splatvar_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,0,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
-; X32-SSE-NEXT: pandn %xmm3, %xmm4
-; X32-SSE-NEXT: psllw $5, %xmm4
-; X32-SSE-NEXT: pxor %xmm5, %xmm5
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm6
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm0, %xmm7
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand %xmm6, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm7, %xmm0
-; X32-SSE-NEXT: paddb %xmm4, %xmm4
-; X32-SSE-NEXT: pxor %xmm6, %xmm6
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm6
-; X32-SSE-NEXT: movdqa %xmm6, %xmm7
-; X32-SSE-NEXT: pandn %xmm0, %xmm7
-; X32-SSE-NEXT: psllw $2, %xmm0
-; X32-SSE-NEXT: pand %xmm6, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm7, %xmm0
-; X32-SSE-NEXT: paddb %xmm4, %xmm4
-; X32-SSE-NEXT: pcmpgtb %xmm4, %xmm5
-; X32-SSE-NEXT: movdqa %xmm5, %xmm4
-; X32-SSE-NEXT: pandn %xmm0, %xmm4
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: pand %xmm5, %xmm0
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm2, %xmm1
-; X32-SSE-NEXT: pcmpeqd %xmm3, %xmm3
-; X32-SSE-NEXT: psrlw %xmm2, %xmm3
-; X32-SSE-NEXT: psrlw $8, %xmm3
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm3[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm1, %xmm2
-; X32-SSE-NEXT: por %xmm4, %xmm2
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,0,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
+; X86-SSE2-NEXT: pandn %xmm3, %xmm4
+; X86-SSE2-NEXT: psllw $5, %xmm4
+; X86-SSE2-NEXT: pxor %xmm5, %xmm5
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm6
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm0, %xmm7
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand %xmm6, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm7, %xmm0
+; X86-SSE2-NEXT: paddb %xmm4, %xmm4
+; X86-SSE2-NEXT: pxor %xmm6, %xmm6
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm6
+; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
+; X86-SSE2-NEXT: pandn %xmm0, %xmm7
+; X86-SSE2-NEXT: psllw $2, %xmm0
+; X86-SSE2-NEXT: pand %xmm6, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm7, %xmm0
+; X86-SSE2-NEXT: paddb %xmm4, %xmm4
+; X86-SSE2-NEXT: pcmpgtb %xmm4, %xmm5
+; X86-SSE2-NEXT: movdqa %xmm5, %xmm4
+; X86-SSE2-NEXT: pandn %xmm0, %xmm4
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: pand %xmm5, %xmm0
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm1
+; X86-SSE2-NEXT: pcmpeqd %xmm3, %xmm3
+; X86-SSE2-NEXT: psrlw %xmm2, %xmm3
+; X86-SSE2-NEXT: psrlw $8, %xmm3
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm3[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm1, %xmm2
+; X86-SSE2-NEXT: por %xmm4, %xmm2
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <16 x i8> %amt, <16 x i8> undef, <16 x i32> zeroinitializer
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> %splat)
ret <16 x i8> %res
@@ -2116,26 +2116,26 @@ define <2 x i64> @constant_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y) nounwind {
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = <4,u,14,u>
-; X32-SSE-NEXT: movdqa %xmm3, %xmm4
-; X32-SSE-NEXT: pand %xmm2, %xmm4
-; X32-SSE-NEXT: movdqa %xmm1, %xmm5
-; X32-SSE-NEXT: psrlq %xmm4, %xmm5
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
-; X32-SSE-NEXT: psrlq %xmm4, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
-; X32-SSE-NEXT: pandn %xmm2, %xmm3
-; X32-SSE-NEXT: psllq $1, %xmm0
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: psllq %xmm3, %xmm2
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm3, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = <4,u,14,u>
+; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
+; X86-SSE2-NEXT: pand %xmm2, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm5
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[2,3,2,3]
+; X86-SSE2-NEXT: psrlq %xmm4, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm5[0],xmm1[1]
+; X86-SSE2-NEXT: pandn %xmm2, %xmm3
+; X86-SSE2-NEXT: psllq $1, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: psllq %xmm3, %xmm2
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm3, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm2[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> <i64 4, i64 14>)
ret <2 x i64> %res
}
@@ -2255,28 +2255,28 @@ define <4 x i32> @constant_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y) nounwind {
; XOPAVX2-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: psrld $7, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-NEXT: psrld $6, %xmm3
-; X32-SSE-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: psrld $5, %xmm2
-; X32-SSE-NEXT: psrld $4, %xmm1
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
-; X32-SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm3[0,3]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [268435456,134217728,67108864,33554432]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm3, %xmm2
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: psrld $7, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-SSE2-NEXT: psrld $6, %xmm3
+; X86-SSE2-NEXT: punpckhqdq {{.*#+}} xmm3 = xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: psrld $5, %xmm2
+; X86-SSE2-NEXT: psrld $4, %xmm1
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,3],xmm3[0,3]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [268435456,134217728,67108864,33554432]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm3, %xmm2
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
ret <4 x i32> %res
}
@@ -2378,18 +2378,18 @@ define <8 x i16> @constant_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y) nounwind {
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [0,65535,65535,65535,65535,65535,65535,65535]
-; X32-SSE-NEXT: movdqa %xmm2, %xmm3
-; X32-SSE-NEXT: pandn %xmm1, %xmm3
-; X32-SSE-NEXT: pmulhuw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: pand %xmm1, %xmm2
-; X32-SSE-NEXT: psllw $1, %xmm0
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,65535,65535,65535,65535,65535,65535,65535]
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
+; X86-SSE2-NEXT: pandn %xmm1, %xmm3
+; X86-SSE2-NEXT: pmulhuw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: pand %xmm1, %xmm2
+; X86-SSE2-NEXT: psllw $1, %xmm0
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
ret <8 x i16> %res
}
@@ -2566,29 +2566,29 @@ define <16 x i8> @constant_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y) nounwind {
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm3
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: psrlw $8, %xmm3
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psrlw $8, %xmm1
-; X32-SSE-NEXT: packuswb %xmm3, %xmm1
-; X32-SSE-NEXT: paddb %xmm0, %xmm0
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
-; X32-SSE-NEXT: pand %xmm3, %xmm2
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: pand %xmm3, %xmm0
-; X32-SSE-NEXT: packuswb %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: psrlw $8, %xmm3
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psrlw $8, %xmm1
+; X86-SSE2-NEXT: packuswb %xmm3, %xmm1
+; X86-SSE2-NEXT: paddb %xmm0, %xmm0
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
+; X86-SSE2-NEXT: pand %xmm3, %xmm2
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: pand %xmm3, %xmm0
+; X86-SSE2-NEXT: packuswb %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1>)
ret <16 x i8> %res
}
@@ -2659,14 +2659,14 @@ define <2 x i64> @splatconstant_funnnel_v2i64(<2 x i64> %x, <2 x i64> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlq $14, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
-; X32-SSE-NEXT: psllq $50, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlq $14, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
+; X86-SSE2-NEXT: psllq $50, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %y, <2 x i64> <i64 14, i64 14>)
ret <2 x i64> %res
}
@@ -2733,12 +2733,12 @@ define <4 x i32> @splatconstant_funnnel_v4i32(<4 x i32> %x, <4 x i32> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrld $4, %xmm1
-; X32-SSE-NEXT: pslld $28, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrld $4, %xmm1
+; X86-SSE2-NEXT: pslld $28, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %y, <4 x i32> <i32 4, i32 4, i32 4, i32 4>)
ret <4 x i32> %res
}
@@ -2805,12 +2805,12 @@ define <8 x i16> @splatconstant_funnnel_v8i16(<8 x i16> %x, <8 x i16> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlw $7, %xmm1
-; X32-SSE-NEXT: psllw $9, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlw $7, %xmm1
+; X86-SSE2-NEXT: psllw $9, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %y, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
ret <8 x i16> %res
}
@@ -2889,14 +2889,14 @@ define <16 x i8> @splatconstant_funnnel_v16i8(<16 x i8> %x, <16 x i8> %y) nounwi
; XOP-NEXT: vpor %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %y, <16 x i8> <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>)
ret <16 x i8> %res
}
diff --git a/llvm/test/CodeGen/X86/vector-fshr-256.ll b/llvm/test/CodeGen/X86/vector-fshr-256.ll
index e6f3097e7496..942a318ba238 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-256.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-256.ll
@@ -1,14 +1,14 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOPAVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOPAVX2
declare <4 x i64> @llvm.fshr.v4i64(<4 x i64>, <4 x i64>, <4 x i64>)
declare <8 x i32> @llvm.fshr.v8i32(<8 x i32>, <8 x i32>, <8 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshr-512.ll b/llvm/test/CodeGen/X86/vector-fshr-512.ll
index 0c001a8a8c0c..2e19d36d6d86 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-512.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-512.ll
@@ -1,10 +1,10 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512,AVX512VBMI2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLVBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
declare <8 x i64> @llvm.fshr.v8i64(<8 x i64>, <8 x i64>, <8 x i64>)
declare <16 x i32> @llvm.fshr.v16i32(<16 x i32>, <16 x i32>, <16 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshr-rot-128.ll b/llvm/test/CodeGen/X86/vector-fshr-rot-128.ll
index ecf27e5884da..64ccaff6d94c 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-rot-128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-rot-128.ll
@@ -11,7 +11,7 @@
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i64> @llvm.fshr.v2i64(<2 x i64>, <2 x i64>, <2 x i64>)
declare <4 x i32> @llvm.fshr.v4i32(<4 x i32>, <4 x i32>, <4 x i32>)
@@ -130,26 +130,26 @@ define <2 x i64> @var_funnnel_v2i64(<2 x i64> %x, <2 x i64> %amt) nounwind {
; XOP-NEXT: vprotq %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm1, %xmm3
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psrlq %xmm1, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psrlq %xmm1, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm1, %xmm3
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psrlq %xmm1, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm1, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> %amt)
ret <2 x i64> %res
}
@@ -262,26 +262,26 @@ define <4 x i32> @var_funnnel_v4i32(<4 x i32> %x, <4 x i32> %amt) nounwind {
; XOP-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubd %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubd %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> %amt)
ret <4 x i32> %res
}
@@ -444,33 +444,33 @@ define <8 x i16> @var_funnnel_v8i16(<8 x i16> %x, <8 x i16> %amt) nounwind {
; XOP-NEXT: vprotw %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubw %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm1
-; X32-SSE-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pslld $23, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [1065353216,1065353216,1065353216,1065353216]
-; X32-SSE-NEXT: paddd %xmm3, %xmm1
-; X32-SSE-NEXT: cvttps2dq %xmm1, %xmm1
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd %xmm3, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm2
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
-; X32-SSE-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X32-SSE-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: pmulhuw %xmm2, %xmm1
-; X32-SSE-NEXT: pmullw %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubw %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm1
+; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pslld $23, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [1065353216,1065353216,1065353216,1065353216]
+; X86-SSE2-NEXT: paddd %xmm3, %xmm1
+; X86-SSE2-NEXT: cvttps2dq %xmm1, %xmm1
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3]
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd %xmm3, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm2
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
+; X86-SSE2-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,6,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
+; X86-SSE2-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: pmulhuw %xmm2, %xmm1
+; X86-SSE2-NEXT: pmullw %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> %amt)
ret <8 x i16> %res
}
@@ -656,50 +656,50 @@ define <16 x i8> @var_funnnel_v16i8(<16 x i8> %x, <16 x i8> %amt) nounwind {
; XOP-NEXT: vprotb %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pxor %xmm0, %xmm0
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubb %xmm1, %xmm3
-; X32-SSE-NEXT: psllw $5, %xmm3
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: pcmpgtb %xmm3, %xmm1
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: psrlw $4, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: movdqa %xmm2, %xmm5
-; X32-SSE-NEXT: psllw $4, %xmm5
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm5
-; X32-SSE-NEXT: por %xmm4, %xmm5
-; X32-SSE-NEXT: pand %xmm1, %xmm5
-; X32-SSE-NEXT: pandn %xmm2, %xmm1
-; X32-SSE-NEXT: por %xmm5, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm2
-; X32-SSE-NEXT: psrlw $6, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa %xmm1, %xmm4
-; X32-SSE-NEXT: psllw $2, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: por %xmm2, %xmm4
-; X32-SSE-NEXT: paddb %xmm3, %xmm3
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: pcmpgtb %xmm3, %xmm2
-; X32-SSE-NEXT: pand %xmm2, %xmm4
-; X32-SSE-NEXT: pandn %xmm1, %xmm2
-; X32-SSE-NEXT: por %xmm4, %xmm2
-; X32-SSE-NEXT: movdqa %xmm2, %xmm1
-; X32-SSE-NEXT: paddb %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm2, %xmm4
-; X32-SSE-NEXT: psrlw $7, %xmm4
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm4
-; X32-SSE-NEXT: por %xmm1, %xmm4
-; X32-SSE-NEXT: paddb %xmm3, %xmm3
-; X32-SSE-NEXT: pcmpgtb %xmm3, %xmm0
-; X32-SSE-NEXT: pand %xmm0, %xmm4
-; X32-SSE-NEXT: pandn %xmm2, %xmm0
-; X32-SSE-NEXT: por %xmm4, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pxor %xmm0, %xmm0
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubb %xmm1, %xmm3
+; X86-SSE2-NEXT: psllw $5, %xmm3
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: pcmpgtb %xmm3, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlw $4, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm5
+; X86-SSE2-NEXT: psllw $4, %xmm5
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm5
+; X86-SSE2-NEXT: por %xmm4, %xmm5
+; X86-SSE2-NEXT: pand %xmm1, %xmm5
+; X86-SSE2-NEXT: pandn %xmm2, %xmm1
+; X86-SSE2-NEXT: por %xmm5, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
+; X86-SSE2-NEXT: psrlw $6, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm4
+; X86-SSE2-NEXT: psllw $2, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: por %xmm2, %xmm4
+; X86-SSE2-NEXT: paddb %xmm3, %xmm3
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: pcmpgtb %xmm3, %xmm2
+; X86-SSE2-NEXT: pand %xmm2, %xmm4
+; X86-SSE2-NEXT: pandn %xmm1, %xmm2
+; X86-SSE2-NEXT: por %xmm4, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm1
+; X86-SSE2-NEXT: paddb %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
+; X86-SSE2-NEXT: psrlw $7, %xmm4
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm4
+; X86-SSE2-NEXT: por %xmm1, %xmm4
+; X86-SSE2-NEXT: paddb %xmm3, %xmm3
+; X86-SSE2-NEXT: pcmpgtb %xmm3, %xmm0
+; X86-SSE2-NEXT: pand %xmm0, %xmm4
+; X86-SSE2-NEXT: pandn %xmm2, %xmm0
+; X86-SSE2-NEXT: por %xmm4, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> %amt)
ret <16 x i8> %res
}
@@ -780,27 +780,27 @@ define <2 x i64> @splatvar_funnnel_v2i64(<2 x i64> %x, <2 x i64> %amt) nounwind
; XOPAVX2-NEXT: vprotq %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,1]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm1, %xmm3
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psrlq %xmm1, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psrlq %xmm1, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,1]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [63,0,63,0]
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm1, %xmm3
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psrlq %xmm1, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm1, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i64> %amt, <2 x i64> undef, <2 x i32> zeroinitializer
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> %splat)
ret <2 x i64> %res
@@ -912,20 +912,20 @@ define <4 x i32> @splatvar_funnnel_v4i32(<4 x i32> %x, <4 x i32> %amt) nounwind
; XOPAVX2-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movd %xmm1, %eax
-; X32-SSE-NEXT: negl %eax
-; X32-SSE-NEXT: andl $31, %eax
-; X32-SSE-NEXT: movd %eax, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pslld %xmm1, %xmm2
-; X32-SSE-NEXT: movl $32, %ecx
-; X32-SSE-NEXT: subl %eax, %ecx
-; X32-SSE-NEXT: movd %ecx, %xmm1
-; X32-SSE-NEXT: psrld %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movd %xmm1, %eax
+; X86-SSE2-NEXT: negl %eax
+; X86-SSE2-NEXT: andl $31, %eax
+; X86-SSE2-NEXT: movd %eax, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pslld %xmm1, %xmm2
+; X86-SSE2-NEXT: movl $32, %ecx
+; X86-SSE2-NEXT: subl %eax, %ecx
+; X86-SSE2-NEXT: movd %ecx, %xmm1
+; X86-SSE2-NEXT: psrld %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <4 x i32> %amt, <4 x i32> undef, <4 x i32> zeroinitializer
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> %splat)
ret <4 x i32> %res
@@ -1009,22 +1009,22 @@ define <8 x i16> @splatvar_funnnel_v8i16(<8 x i16> %x, <8 x i16> %amt) nounwind
; XOPAVX2-NEXT: vprotw %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubw %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,0,0]
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-NEXT: psllw %xmm1, %xmm3
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [16,16,16,16,16,16,16,16]
-; X32-SSE-NEXT: psubw %xmm2, %xmm1
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm1 = xmm1[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubw %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [65535,0,0,0]
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-SSE2-NEXT: psllw %xmm1, %xmm3
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [16,16,16,16,16,16,16,16]
+; X86-SSE2-NEXT: psubw %xmm2, %xmm1
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm1 = xmm1[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <8 x i16> %amt, <8 x i16> undef, <8 x i32> zeroinitializer
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> %splat)
ret <8 x i16> %res
@@ -1215,36 +1215,36 @@ define <16 x i8> @splatvar_funnnel_v16i8(<16 x i8> %x, <16 x i8> %amt) nounwind
; XOPAVX2-NEXT: vprotb %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubb %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm3 = [8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8]
-; X32-SSE-NEXT: psubb %xmm2, %xmm3
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllw %xmm2, %xmm1
-; X32-SSE-NEXT: pcmpeqd %xmm4, %xmm4
-; X32-SSE-NEXT: pcmpeqd %xmm5, %xmm5
-; X32-SSE-NEXT: psllw %xmm2, %xmm5
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm5 = xmm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm5[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: pslldq {{.*#+}} xmm3 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm3[0]
-; X32-SSE-NEXT: psrldq {{.*#+}} xmm3 = xmm3[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
-; X32-SSE-NEXT: psrlw %xmm3, %xmm0
-; X32-SSE-NEXT: psrlw %xmm3, %xmm4
-; X32-SSE-NEXT: psrlw $8, %xmm4
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pshuflw {{.*#+}} xmm2 = xmm4[0,0,0,0,4,5,6,7]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
-; X32-SSE-NEXT: pand %xmm0, %xmm2
-; X32-SSE-NEXT: por %xmm2, %xmm1
-; X32-SSE-NEXT: movdqa %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubb %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8]
+; X86-SSE2-NEXT: psubb %xmm2, %xmm3
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm2 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm2[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm2 = xmm2[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllw %xmm2, %xmm1
+; X86-SSE2-NEXT: pcmpeqd %xmm4, %xmm4
+; X86-SSE2-NEXT: pcmpeqd %xmm5, %xmm5
+; X86-SSE2-NEXT: psllw %xmm2, %xmm5
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm5 = xmm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm5[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: pslldq {{.*#+}} xmm3 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm3[0]
+; X86-SSE2-NEXT: psrldq {{.*#+}} xmm3 = xmm3[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE2-NEXT: psrlw %xmm3, %xmm0
+; X86-SSE2-NEXT: psrlw %xmm3, %xmm4
+; X86-SSE2-NEXT: psrlw $8, %xmm4
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pshuflw {{.*#+}} xmm2 = xmm4[0,0,0,0,4,5,6,7]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
+; X86-SSE2-NEXT: pand %xmm0, %xmm2
+; X86-SSE2-NEXT: por %xmm2, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <16 x i8> %amt, <16 x i8> undef, <16 x i32> zeroinitializer
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> %splat)
ret <16 x i8> %res
@@ -1334,27 +1334,27 @@ define <2 x i64> @constant_funnnel_v2i64(<2 x i64> %x) nounwind {
; XOP-NEXT: vprotq {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [63,0,63,0]
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = <4,u,14,u>
-; X32-SSE-NEXT: pxor %xmm3, %xmm3
-; X32-SSE-NEXT: psubq %xmm2, %xmm3
-; X32-SSE-NEXT: pand %xmm1, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm4
-; X32-SSE-NEXT: psrlq %xmm2, %xmm4
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm5
-; X32-SSE-NEXT: psrlq %xmm2, %xmm5
-; X32-SSE-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
-; X32-SSE-NEXT: pand %xmm1, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllq %xmm3, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
-; X32-SSE-NEXT: psllq %xmm2, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-SSE-NEXT: orpd %xmm5, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [63,0,63,0]
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = <4,u,14,u>
+; X86-SSE2-NEXT: pxor %xmm3, %xmm3
+; X86-SSE2-NEXT: psubq %xmm2, %xmm3
+; X86-SSE2-NEXT: pand %xmm1, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
+; X86-SSE2-NEXT: psrlq %xmm2, %xmm4
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm5
+; X86-SSE2-NEXT: psrlq %xmm2, %xmm5
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm5 = xmm4[0],xmm5[1]
+; X86-SSE2-NEXT: pand %xmm1, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllq %xmm3, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; X86-SSE2-NEXT: psllq %xmm2, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-SSE2-NEXT: orpd %xmm5, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> <i64 4, i64 14>)
ret <2 x i64> %res
}
@@ -1444,21 +1444,21 @@ define <4 x i32> @constant_funnnel_v4i32(<4 x i32> %x) nounwind {
; XOP-NEXT: vprotd {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [268435456,134217728,67108864,33554432]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [268435456,134217728,67108864,33554432]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
ret <4 x i32> %res
}
@@ -1520,14 +1520,14 @@ define <8 x i16> @constant_funnnel_v8i16(<8 x i16> %x) nounwind {
; XOP-NEXT: vprotw {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [1,32768,16384,8192,4096,2048,1024,512]
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: pmulhuw %xmm1, %xmm2
-; X32-SSE-NEXT: pmullw %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm2, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [1,32768,16384,8192,4096,2048,1024,512]
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: pmulhuw %xmm1, %xmm2
+; X86-SSE2-NEXT: pmullw %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm2, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
ret <8 x i16> %res
}
@@ -1663,29 +1663,29 @@ define <16 x i8> @constant_funnnel_v16i8(<16 x i8> %x) nounwind {
; XOP-NEXT: vprotb {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm1, %xmm1
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: psrlw $8, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm3
-; X32-SSE-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]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm3
-; X32-SSE-NEXT: psrlw $8, %xmm3
-; X32-SSE-NEXT: packuswb %xmm2, %xmm3
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; X32-SSE-NEXT: pand %xmm2, %xmm1
-; X32-SSE-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; X32-SSE-NEXT: pmullw {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: pand %xmm2, %xmm0
-; X32-SSE-NEXT: packuswb %xmm1, %xmm0
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm1, %xmm1
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: psrlw $8, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
+; X86-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]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm3
+; X86-SSE2-NEXT: psrlw $8, %xmm3
+; X86-SSE2-NEXT: packuswb %xmm2, %xmm3
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
+; X86-SSE2-NEXT: pand %xmm2, %xmm1
+; X86-SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; X86-SSE2-NEXT: pmullw {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: pand %xmm2, %xmm0
+; X86-SSE2-NEXT: packuswb %xmm1, %xmm0
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1>)
ret <16 x i8> %res
}
@@ -1741,15 +1741,15 @@ define <2 x i64> @splatconstant_funnnel_v2i64(<2 x i64> %x) nounwind {
; XOP-NEXT: vprotq $50, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i64:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psllq $50, %xmm1
-; X32-SSE-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
-; X32-SSE-NEXT: psrlq $14, %xmm0
-; X32-SSE-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
-; X32-SSE-NEXT: orpd %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i64:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psllq $50, %xmm1
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm1 = xmm1[0,1]
+; X86-SSE2-NEXT: psrlq $14, %xmm0
+; X86-SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm0[0,1]
+; X86-SSE2-NEXT: orpd %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x, <2 x i64> %x, <2 x i64> <i64 14, i64 14>)
ret <2 x i64> %res
}
@@ -1801,13 +1801,13 @@ define <4 x i32> @splatconstant_funnnel_v4i32(<4 x i32> %x) nounwind {
; XOP-NEXT: vprotd $28, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v4i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrld $4, %xmm1
-; X32-SSE-NEXT: pslld $28, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v4i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrld $4, %xmm1
+; X86-SSE2-NEXT: pslld $28, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x, <4 x i32> %x, <4 x i32> <i32 4, i32 4, i32 4, i32 4>)
ret <4 x i32> %res
}
@@ -1840,13 +1840,13 @@ define <8 x i16> @splatconstant_funnnel_v8i16(<8 x i16> %x) nounwind {
; XOP-NEXT: vprotw $9, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v8i16:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlw $7, %xmm1
-; X32-SSE-NEXT: psllw $9, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v8i16:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlw $7, %xmm1
+; X86-SSE2-NEXT: psllw $9, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x, <8 x i16> %x, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
ret <8 x i16> %res
}
@@ -1908,15 +1908,15 @@ define <16 x i8> @splatconstant_funnnel_v16i8(<16 x i8> %x) nounwind {
; XOP-NEXT: vprotb $4, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v16i8:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: psrlw $4, %xmm1
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm1
-; X32-SSE-NEXT: psllw $4, %xmm0
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm0
-; X32-SSE-NEXT: por %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v16i8:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: psrlw $4, %xmm1
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm1
+; X86-SSE2-NEXT: psllw $4, %xmm0
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT: por %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %x, <16 x i8> %x, <16 x i8> <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>)
ret <16 x i8> %res
}
diff --git a/llvm/test/CodeGen/X86/vector-fshr-rot-256.ll b/llvm/test/CodeGen/X86/vector-fshr-rot-256.ll
index 5211be0fcdc4..c9177802a475 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-rot-256.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-rot-256.ll
@@ -1,12 +1,12 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOPAVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOPAVX2
declare <4 x i64> @llvm.fshr.v4i64(<4 x i64>, <4 x i64>, <4 x i64>)
declare <8 x i32> @llvm.fshr.v8i32(<8 x i32>, <8 x i32>, <8 x i32>)
diff --git a/llvm/test/CodeGen/X86/vector-fshr-rot-sub128.ll b/llvm/test/CodeGen/X86/vector-fshr-rot-sub128.ll
index ebb95d6e0410..5e6bf0110f5c 100644
--- a/llvm/test/CodeGen/X86/vector-fshr-rot-sub128.ll
+++ b/llvm/test/CodeGen/X86/vector-fshr-rot-sub128.ll
@@ -1,17 +1,17 @@
; 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=SSE,SSE2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE,SSE41
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VL
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512,AVX512VLBW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE41
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOP,XOPAVX1
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOP,XOPAVX2
; Just one 32-bit run to make sure we do reasonable things for i64 cases.
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X32-SSE,X32-SSE2
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2
declare <2 x i32> @llvm.fshr.v2i32(<2 x i32>, <2 x i32>, <2 x i32>)
@@ -127,26 +127,26 @@ define <2 x i32> @var_funnnel_v2i32(<2 x i32> %x, <2 x i32> %amt) nounwind {
; XOP-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: var_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubd %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: var_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubd %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> %amt)
ret <2 x i32> %res
}
@@ -280,27 +280,27 @@ define <2 x i32> @splatvar_funnnel_v2i32(<2 x i32> %x, <2 x i32> %amt) nounwind
; XOPAVX2-NEXT: vprotd %xmm1, %xmm0, %xmm0
; XOPAVX2-NEXT: retq
;
-; X32-SSE-LABEL: splatvar_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
-; X32-SSE-NEXT: pxor %xmm2, %xmm2
-; X32-SSE-NEXT: psubd %xmm1, %xmm2
-; X32-SSE-NEXT: pand {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: pslld $23, %xmm2
-; X32-SSE-NEXT: paddd {{\.LCPI.*}}, %xmm2
-; X32-SSE-NEXT: cvttps2dq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatvar_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,0,1,1]
+; X86-SSE2-NEXT: pxor %xmm2, %xmm2
+; X86-SSE2-NEXT: psubd %xmm1, %xmm2
+; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: pslld $23, %xmm2
+; X86-SSE2-NEXT: paddd {{\.LCPI.*}}, %xmm2
+; X86-SSE2-NEXT: cvttps2dq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%splat = shufflevector <2 x i32> %amt, <2 x i32> undef, <2 x i32> zeroinitializer
%res = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> %splat)
ret <2 x i32> %res
@@ -395,21 +395,21 @@ define <2 x i32> @constant_funnnel_v2i32(<2 x i32> %x) nounwind {
; XOP-NEXT: vprotd {{.*}}(%rip), %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: constant_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa {{.*#+}} xmm1 = [268435456,134217728,1,1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm1, %xmm0
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; X32-SSE-NEXT: pmuludq %xmm2, %xmm1
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X32-SSE-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; X32-SSE-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X32-SSE-NEXT: por %xmm3, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: constant_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [268435456,134217728,1,1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm1, %xmm0
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; X86-SSE2-NEXT: pmuludq %xmm2, %xmm1
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X86-SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE2-NEXT: por %xmm3, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> <i32 4, i32 5>)
ret <2 x i32> %res
}
@@ -487,16 +487,16 @@ define <2 x i32> @splatconstant_funnnel_v2i32(<2 x i32> %x) nounwind {
; XOP-NEXT: vprotd $28, %xmm0, %xmm0
; XOP-NEXT: retq
;
-; X32-SSE-LABEL: splatconstant_funnnel_v2i32:
-; X32-SSE: # %bb.0:
-; X32-SSE-NEXT: movdqa %xmm0, %xmm2
-; X32-SSE-NEXT: psrld $4, %xmm2
-; X32-SSE-NEXT: movdqa %xmm0, %xmm1
-; X32-SSE-NEXT: pslld $28, %xmm1
-; X32-SSE-NEXT: por %xmm2, %xmm1
-; X32-SSE-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3]
-; X32-SSE-NEXT: movaps %xmm1, %xmm0
-; X32-SSE-NEXT: retl
+; X86-SSE2-LABEL: splatconstant_funnnel_v2i32:
+; X86-SSE2: # %bb.0:
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
+; X86-SSE2-NEXT: psrld $4, %xmm2
+; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
+; X86-SSE2-NEXT: pslld $28, %xmm1
+; X86-SSE2-NEXT: por %xmm2, %xmm1
+; X86-SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3]
+; X86-SSE2-NEXT: movaps %xmm1, %xmm0
+; X86-SSE2-NEXT: retl
%res = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> %x, <2 x i32> %x, <2 x i32> <i32 4, i32 4>)
ret <2 x i32> %res
}
More information about the llvm-commits
mailing list