[llvm] a61272a - [DAG] Fold vector mul(x, 0)/mul(x, 1) to a clearing mask
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Sat Sep 26 06:32:39 PDT 2020
Author: Simon Pilgrim
Date: 2020-09-26T14:31:57+01:00
New Revision: a61272a900296573902dfce99274fd335239a4d3
URL: https://github.com/llvm/llvm-project/commit/a61272a900296573902dfce99274fd335239a4d3
DIFF: https://github.com/llvm/llvm-project/commit/a61272a900296573902dfce99274fd335239a4d3.diff
LOG: [DAG] Fold vector mul(x,0)/mul(x,1) to a clearing mask
If we're multiplying all elements of a vector by '0' or '1' then we can more efficiently perform this as a clearing mask (that is likely to further simplify to a shuffle blend).
This was noticed when reviewing D87502 but seems to help idiv/irem by constant cases even more as '0'/'1' values are often used for 'passthrough' cases.
Differential Revision: https://reviews.llvm.org/D88225
Added:
Modified:
llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
llvm/test/CodeGen/AArch64/srem-seteq-vec-nonsplat.ll
llvm/test/CodeGen/X86/srem-seteq-vec-nonsplat.ll
llvm/test/CodeGen/X86/vector-idiv-sdiv-128.ll
llvm/test/CodeGen/X86/vector-idiv-sdiv-256.ll
llvm/test/CodeGen/X86/vector-idiv-sdiv-512.ll
llvm/test/CodeGen/X86/vector-mul.ll
llvm/test/CodeGen/X86/vector-trunc-math.ll
Removed:
################################################################################
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index c572e3abb564..2b6b430efa29 100644
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -24,6 +24,7 @@
#include "llvm/ADT/Optional.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SetVector.h"
+#include "llvm/ADT/SmallBitVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
@@ -3704,6 +3705,37 @@ SDValue DAGCombiner::visitMUL(SDNode *N) {
return DAG.getVScale(SDLoc(N), VT, C0 * C1);
}
+ // Fold ((mul x, 0/undef) -> 0,
+ // (mul x, 1) -> x) -> x)
+ // -> and(x, mask)
+ // We can replace vectors with '0' and '1' factors with a clearing mask.
+ if (VT.isFixedLengthVector()) {
+ unsigned NumElts = VT.getVectorNumElements();
+ SmallBitVector ClearMask;
+ ClearMask.reserve(NumElts);
+ auto IsClearMask = [&ClearMask](ConstantSDNode *V) {
+ if (!V || V->isNullValue()) {
+ ClearMask.push_back(true);
+ return true;
+ }
+ ClearMask.push_back(false);
+ return V->isOne();
+ };
+ if ((!LegalOperations || TLI.isOperationLegalOrCustom(ISD::AND, VT)) &&
+ ISD::matchUnaryPredicate(N1, IsClearMask, /*AllowUndefs*/ true)) {
+ assert(N1.getOpcode() == ISD::BUILD_VECTOR && "Unknown constant vector");
+ SDLoc DL(N);
+ EVT LegalSVT = N1.getOperand(0).getValueType();
+ SDValue Zero = DAG.getConstant(0, DL, LegalSVT);
+ SDValue AllOnes = DAG.getAllOnesConstant(DL, LegalSVT);
+ SmallVector<SDValue, 16> Mask(NumElts, AllOnes);
+ for (unsigned I = 0; I != NumElts; ++I)
+ if (ClearMask[I])
+ Mask[I] = Zero;
+ return DAG.getNode(ISD::AND, DL, VT, N0, DAG.getBuildVector(VT, DL, Mask));
+ }
+ }
+
// reassociate mul
if (SDValue RMUL = reassociateOps(ISD::MUL, SDLoc(N), N0, N1, N->getFlags()))
return RMUL;
diff --git a/llvm/test/CodeGen/AArch64/srem-seteq-vec-nonsplat.ll b/llvm/test/CodeGen/AArch64/srem-seteq-vec-nonsplat.ll
index 7e78d11f6d8b..669be10d7af8 100644
--- a/llvm/test/CodeGen/AArch64/srem-seteq-vec-nonsplat.ll
+++ b/llvm/test/CodeGen/AArch64/srem-seteq-vec-nonsplat.ll
@@ -15,7 +15,8 @@ define <4 x i32> @test_srem_odd_even(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v3.4s
; CHECK-NEXT: ldr q3, [x8, :lo12:.LCPI0_2]
; CHECK-NEXT: adrp x8, .LCPI0_3
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI0_3]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
@@ -227,7 +228,8 @@ define <4 x i32> @test_srem_odd_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v3.4s
; CHECK-NEXT: ldr q3, [x8, :lo12:.LCPI7_2]
; CHECK-NEXT: adrp x8, .LCPI7_3
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI7_3]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
@@ -282,7 +284,8 @@ define <4 x i32> @test_srem_odd_even_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v3.4s
; CHECK-NEXT: ldr q3, [x8, :lo12:.LCPI9_2]
; CHECK-NEXT: adrp x8, .LCPI9_3
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI9_3]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
@@ -371,7 +374,8 @@ define <4 x i32> @test_srem_odd_even_one(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v4.4s
; CHECK-NEXT: ldr q4, [x8, :lo12:.LCPI12_3]
; CHECK-NEXT: adrp x8, .LCPI12_4
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI12_4]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
@@ -696,7 +700,8 @@ define <4 x i32> @test_srem_odd_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v4.4s
; CHECK-NEXT: ldr q4, [x8, :lo12:.LCPI22_3]
; CHECK-NEXT: adrp x8, .LCPI22_4
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI22_4]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
@@ -762,7 +767,8 @@ define <4 x i32> @test_srem_odd_even_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-NEXT: uzp2 v1.4s, v1.4s, v4.4s
; CHECK-NEXT: ldr q4, [x8, :lo12:.LCPI24_3]
; CHECK-NEXT: adrp x8, .LCPI24_4
-; CHECK-NEXT: mla v1.4s, v0.4s, v2.4s
+; CHECK-NEXT: and v2.16b, v0.16b, v2.16b
+; CHECK-NEXT: add v1.4s, v1.4s, v2.4s
; CHECK-NEXT: ldr q2, [x8, :lo12:.LCPI24_4]
; CHECK-NEXT: neg v3.4s, v3.4s
; CHECK-NEXT: sshl v3.4s, v1.4s, v3.4s
diff --git a/llvm/test/CodeGen/X86/srem-seteq-vec-nonsplat.ll b/llvm/test/CodeGen/X86/srem-seteq-vec-nonsplat.ll
index 9982f34f736d..5435e2d9d74e 100644
--- a/llvm/test/CodeGen/X86/srem-seteq-vec-nonsplat.ll
+++ b/llvm/test/CodeGen/X86/srem-seteq-vec-nonsplat.ll
@@ -9,31 +9,23 @@
define <4 x i32> @test_srem_odd_even(<4 x i32> %X) nounwind {
; CHECK-SSE2-LABEL: test_srem_odd_even:
; CHECK-SSE2: # %bb.0:
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [0,1,0,0]
-; CHECK-SSE2-NEXT: movdqa %xmm0, %xmm2
-; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm2
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,2,2,3]
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm5 = [1717986919,2454267027,1374389535,1374389535]
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [1717986919,2454267027,1374389535,1374389535]
; CHECK-SSE2-NEXT: movdqa %xmm0, %xmm1
-; CHECK-SSE2-NEXT: pmuludq %xmm5, %xmm1
+; CHECK-SSE2-NEXT: pmuludq %xmm3, %xmm1
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm5[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm3[1,1,3,3]
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
+; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm4
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm4[1,3,2,3]
; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
; CHECK-SSE2-NEXT: pxor %xmm1, %xmm1
; CHECK-SSE2-NEXT: pxor %xmm4, %xmm4
; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm4
-; CHECK-SSE2-NEXT: pand %xmm5, %xmm4
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm5 = [0,4294967295,0,0]
-; CHECK-SSE2-NEXT: pand %xmm0, %xmm5
-; CHECK-SSE2-NEXT: paddd %xmm4, %xmm5
-; CHECK-SSE2-NEXT: psubd %xmm5, %xmm2
+; CHECK-SSE2-NEXT: pand %xmm3, %xmm4
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [0,4294967295,0,0]
+; CHECK-SSE2-NEXT: pand %xmm0, %xmm3
+; CHECK-SSE2-NEXT: paddd %xmm3, %xmm4
+; CHECK-SSE2-NEXT: psubd %xmm4, %xmm2
; CHECK-SSE2-NEXT: paddd %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm2, %xmm3
; CHECK-SSE2-NEXT: psrad $5, %xmm3
@@ -68,24 +60,24 @@ define <4 x i32> @test_srem_odd_even(<4 x i32> %X) nounwind {
; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm1
; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,0,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: pxor %xmm3, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1],xmm0[2,3],xmm3[4,5,6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $5, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm3
-; CHECK-SSE41-NEXT: psrad $3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm4
-; CHECK-SSE41-NEXT: psrad $1, %xmm4
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1],xmm1[2,3],xmm4[4,5],xmm1[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm4, %xmm2
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm2
-; CHECK-SSE41-NEXT: psubd %xmm2, %xmm0
-; CHECK-SSE41-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE41-NEXT: pcmpeqd %xmm1, %xmm0
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm4
+; CHECK-SSE41-NEXT: psrad $3, %xmm4
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm5
+; CHECK-SSE41-NEXT: psrad $1, %xmm5
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm4[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1],xmm1[2,3],xmm5[4,5],xmm1[6,7]
+; CHECK-SSE41-NEXT: psrld $31, %xmm3
+; CHECK-SSE41-NEXT: paddd %xmm5, %xmm3
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: psubd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pcmpeqd %xmm2, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
@@ -98,20 +90,20 @@ define <4 x i32> @test_srem_odd_even(<4 x i32> %X) nounwind {
; CHECK-AVX1-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm2[0,1],xmm0[2,3],xmm2[4,5,6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm4
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1],xmm2[2,3],xmm3[4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm5
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
+; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1
; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
+; CHECK-AVX1-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
@@ -124,15 +116,15 @@ define <4 x i32> @test_srem_odd_even(<4 x i32> %X) nounwind {
; CHECK-AVX2-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
+; CHECK-AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm2[0],xmm0[1],xmm2[2,3]
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm3
; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
+; CHECK-AVX2-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
;
@@ -844,18 +836,15 @@ define <4 x i32> @test_srem_odd_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-SSE2-NEXT: pand %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [0,0,4294967295,0]
; CHECK-SSE2-NEXT: pand %xmm0, %xmm4
-; CHECK-SSE2-NEXT: paddd %xmm2, %xmm4
+; CHECK-SSE2-NEXT: paddd %xmm4, %xmm2
; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm3
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[1,3,2,3]
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm3
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,3,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1]
-; CHECK-SSE2-NEXT: psubd %xmm4, %xmm2
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm3 = <0,u,1,u>
-; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm3
-; CHECK-SSE2-NEXT: pand {{.*}}(%rip), %xmm3
-; CHECK-SSE2-NEXT: paddd %xmm2, %xmm3
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
+; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm5
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm5[1,3,2,3]
+; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm5[0],xmm3[1],xmm5[1]
+; CHECK-SSE2-NEXT: psubd %xmm2, %xmm3
+; CHECK-SSE2-NEXT: paddd %xmm4, %xmm3
; CHECK-SSE2-NEXT: movdqa %xmm3, %xmm2
; CHECK-SSE2-NEXT: psrad $1, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm3, %xmm4
@@ -877,66 +866,66 @@ define <4 x i32> @test_srem_odd_poweroftwo(<4 x i32> %X) nounwind {
;
; CHECK-SSE41-LABEL: test_srem_odd_poweroftwo:
; CHECK-SSE41: # %bb.0:
-; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,3,3]
-; CHECK-SSE41-NEXT: pmuldq {{.*}}(%rip), %xmm1
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = <1717986919,u,2147483649,u>
-; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm2
-; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1],xmm1[2,3],xmm2[4,5],xmm1[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm1 = [0,0,1,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm1
-; CHECK-SSE41-NEXT: paddd %xmm2, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm1, %xmm2
+; CHECK-SSE41-NEXT: pxor %xmm1, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm0[4,5],xmm2[6,7]
+; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; CHECK-SSE41-NEXT: pmuldq {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm4 = <1717986919,u,2147483649,u>
+; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm4
+; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm2, %xmm4
+; CHECK-SSE41-NEXT: movdqa %xmm4, %xmm2
; CHECK-SSE41-NEXT: psrad $3, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm4, %xmm3
; CHECK-SSE41-NEXT: psrad $1, %xmm3
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm1
-; CHECK-SSE41-NEXT: paddd %xmm3, %xmm1
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm1
-; CHECK-SSE41-NEXT: psubd %xmm1, %xmm0
-; CHECK-SSE41-NEXT: pxor %xmm1, %xmm1
+; CHECK-SSE41-NEXT: psrld $31, %xmm4
+; CHECK-SSE41-NEXT: paddd %xmm3, %xmm4
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm4
+; CHECK-SSE41-NEXT: psubd %xmm4, %xmm0
; CHECK-SSE41-NEXT: pcmpeqd %xmm1, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
; CHECK-AVX1-LABEL: test_srem_odd_poweroftwo:
; CHECK-AVX1: # %bb.0:
-; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,3,3]
-; CHECK-AVX1-NEXT: vpmuldq {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpmuldq {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm2[0,1],xmm1[2,3],xmm2[4,5],xmm1[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5],xmm3[6,7]
-; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm1[0,1,2,3],xmm0[4,5],xmm1[6,7]
+; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; CHECK-AVX1-NEXT: vpmuldq {{.*}}(%rip), %xmm3, %xmm3
+; CHECK-AVX1-NEXT: vpmuldq {{.*}}(%rip), %xmm0, %xmm4
+; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm3, %xmm2
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm2, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm2, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5],xmm4[6,7]
+; CHECK-AVX1-NEXT: vpsrld $31, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm3, %xmm2
+; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpsubd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
; CHECK-AVX2-LABEL: test_srem_odd_poweroftwo:
; CHECK-AVX2: # %bb.0:
-; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,3,3]
-; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [1717986919,1717986919,1717986919,1717986919]
-; CHECK-AVX2-NEXT: vpmuldq %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpmuldq {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm2[0],xmm1[1],xmm2[2],xmm1[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
-; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm2 = xmm1[0,1],xmm0[2],xmm1[3]
+; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
+; CHECK-AVX2-NEXT: vpbroadcastd {{.*#+}} xmm4 = [1717986919,1717986919,1717986919,1717986919]
+; CHECK-AVX2-NEXT: vpmuldq %xmm4, %xmm3, %xmm3
+; CHECK-AVX2-NEXT: vpmuldq {{.*}}(%rip), %xmm0, %xmm4
+; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm4[0],xmm3[1],xmm4[2],xmm3[3]
+; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm3, %xmm2
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm2, %xmm3
+; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpsubd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
@@ -1065,29 +1054,23 @@ define <4 x i32> @test_srem_even_poweroftwo(<4 x i32> %X) nounwind {
define <4 x i32> @test_srem_odd_even_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-SSE2-LABEL: test_srem_odd_even_poweroftwo:
; CHECK-SSE2: # %bb.0:
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [1717986919,2454267027,2147483649,1374389535]
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [1717986919,2454267027,2147483649,1374389535]
; CHECK-SSE2-NEXT: movdqa %xmm0, %xmm1
-; CHECK-SSE2-NEXT: pmuludq %xmm2, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
+; CHECK-SSE2-NEXT: pmuludq %xmm3, %xmm1
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[1,3,2,3]
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm3[1,1,3,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm4
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm4[1,3,2,3]
+; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
; CHECK-SSE2-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE2-NEXT: pxor %xmm5, %xmm5
-; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm5
-; CHECK-SSE2-NEXT: pand %xmm2, %xmm5
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,4294967295,0]
-; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
-; CHECK-SSE2-NEXT: paddd %xmm5, %xmm2
-; CHECK-SSE2-NEXT: psubd %xmm2, %xmm3
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm2
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm4
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1]
+; CHECK-SSE2-NEXT: pxor %xmm4, %xmm4
+; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm4
+; CHECK-SSE2-NEXT: pand %xmm3, %xmm4
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm3 = [0,4294967295,4294967295,0]
+; CHECK-SSE2-NEXT: pand %xmm0, %xmm3
+; CHECK-SSE2-NEXT: paddd %xmm3, %xmm4
+; CHECK-SSE2-NEXT: psubd %xmm4, %xmm2
; CHECK-SSE2-NEXT: paddd %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm2, %xmm3
; CHECK-SSE2-NEXT: psrad $5, %xmm3
@@ -1122,24 +1105,24 @@ define <4 x i32> @test_srem_odd_even_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm1
; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: movdqa %xmm0, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm2[0,1],xmm3[2,3,4,5],xmm2[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $5, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm3
-; CHECK-SSE41-NEXT: psrad $3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm4
-; CHECK-SSE41-NEXT: psrad $1, %xmm4
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1],xmm1[2,3],xmm4[4,5],xmm1[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm4, %xmm2
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm2
-; CHECK-SSE41-NEXT: psubd %xmm2, %xmm0
-; CHECK-SSE41-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE41-NEXT: pcmpeqd %xmm1, %xmm0
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm4
+; CHECK-SSE41-NEXT: psrad $3, %xmm4
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm5
+; CHECK-SSE41-NEXT: psrad $1, %xmm5
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1,2,3],xmm4[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm5 = xmm5[0,1],xmm1[2,3],xmm5[4,5],xmm1[6,7]
+; CHECK-SSE41-NEXT: psrld $31, %xmm3
+; CHECK-SSE41-NEXT: paddd %xmm5, %xmm3
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: psubd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pcmpeqd %xmm2, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
@@ -1152,20 +1135,20 @@ define <4 x i32> @test_srem_odd_even_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-AVX1-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm2[0,1],xmm0[2,3,4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm4
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1],xmm2[2,3],xmm3[4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm5
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm5[0,1,2,3],xmm4[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
+; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1
; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
+; CHECK-AVX1-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
@@ -1178,15 +1161,15 @@ define <4 x i32> @test_srem_odd_even_poweroftwo(<4 x i32> %X) nounwind {
; CHECK-AVX2-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
+; CHECK-AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm2[0],xmm0[1,2],xmm2[3]
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm3
; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpcmpeqd %xmm1, %xmm0, %xmm0
+; CHECK-AVX2-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
;
@@ -1395,23 +1378,19 @@ define <4 x i32> @test_srem_odd_even_one(<4 x i32> %X) nounwind {
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
+; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm4
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm4[1,3,2,3]
; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
; CHECK-SSE2-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE2-NEXT: pxor %xmm5, %xmm5
-; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm5
-; CHECK-SSE2-NEXT: pand %xmm2, %xmm5
+; CHECK-SSE2-NEXT: pxor %xmm4, %xmm4
+; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm4
+; CHECK-SSE2-NEXT: pand %xmm2, %xmm4
; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,0,0]
; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
-; CHECK-SSE2-NEXT: paddd %xmm5, %xmm2
+; CHECK-SSE2-NEXT: paddd %xmm4, %xmm2
; CHECK-SSE2-NEXT: psubd %xmm2, %xmm3
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm2
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm4
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1]
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,4294967295,0]
+; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
; CHECK-SSE2-NEXT: paddd %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm2, %xmm3
; CHECK-SSE2-NEXT: psrad $5, %xmm3
@@ -1448,25 +1427,25 @@ define <4 x i32> @test_srem_odd_even_one(<4 x i32> %X) nounwind {
; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm1
; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: movdqa %xmm0, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm2[0,1],xmm3[2,3,4,5],xmm2[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $5, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm3
-; CHECK-SSE41-NEXT: psrad $3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm4
+; CHECK-SSE41-NEXT: psrad $3, %xmm4
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $1, %xmm1
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm2
-; CHECK-SSE41-NEXT: pxor %xmm3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm3[4,5],xmm2[6,7]
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm2
-; CHECK-SSE41-NEXT: psubd %xmm2, %xmm0
-; CHECK-SSE41-NEXT: pcmpeqd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm4[2,3],xmm1[4,5],xmm4[6,7]
+; CHECK-SSE41-NEXT: psrld $31, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5],xmm3[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: psubd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pcmpeqd %xmm2, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
@@ -1479,21 +1458,21 @@ define <4 x i32> @test_srem_odd_even_one(<4 x i32> %X) nounwind {
; CHECK-AVX1-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1],xmm2[2,3],xmm3[4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm2[0,1],xmm0[2,3,4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5],xmm1[6,7]
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5],xmm1[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1
; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX1-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX1-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
@@ -1506,16 +1485,16 @@ define <4 x i32> @test_srem_odd_even_one(<4 x i32> %X) nounwind {
; CHECK-AVX2-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
-; CHECK-AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm2 = xmm2[0,1],xmm3[2],xmm2[3]
+; CHECK-AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm2[0],xmm0[1,2],xmm2[3]
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm3
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm3[0,1],xmm2[2],xmm3[3]
; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX2-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX2-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
;
@@ -2696,23 +2675,19 @@ define <4 x i32> @test_srem_odd_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
+; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm4
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm4[1,3,2,3]
; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
; CHECK-SSE2-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE2-NEXT: pxor %xmm5, %xmm5
-; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm5
-; CHECK-SSE2-NEXT: pand %xmm2, %xmm5
+; CHECK-SSE2-NEXT: pxor %xmm4, %xmm4
+; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm4
+; CHECK-SSE2-NEXT: pand %xmm2, %xmm4
; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,0,0]
; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
-; CHECK-SSE2-NEXT: paddd %xmm5, %xmm2
+; CHECK-SSE2-NEXT: paddd %xmm4, %xmm2
; CHECK-SSE2-NEXT: psubd %xmm2, %xmm3
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm2
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm4
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1]
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,4294967295,0]
+; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
; CHECK-SSE2-NEXT: paddd %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm2, %xmm3
; CHECK-SSE2-NEXT: psrad $1, %xmm3
@@ -2747,23 +2722,23 @@ define <4 x i32> @test_srem_odd_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm1
; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: movdqa %xmm0, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm2[0,1],xmm3[2,3,4,5],xmm2[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $1, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm3
-; CHECK-SSE41-NEXT: psrad $3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm2
-; CHECK-SSE41-NEXT: pxor %xmm3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm3[4,5],xmm2[6,7]
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm2
-; CHECK-SSE41-NEXT: psubd %xmm2, %xmm0
-; CHECK-SSE41-NEXT: pcmpeqd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm4
+; CHECK-SSE41-NEXT: psrad $3, %xmm4
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm4[2,3],xmm1[4,5],xmm4[6,7]
+; CHECK-SSE41-NEXT: psrld $31, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5],xmm3[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: psubd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pcmpeqd %xmm2, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
@@ -2776,20 +2751,20 @@ define <4 x i32> @test_srem_odd_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-AVX1-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm3[2,3],xmm2[4,5],xmm3[6,7]
+; CHECK-AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm2[0,1],xmm0[2,3,4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1],xmm4[2,3],xmm3[4,5],xmm4[6,7]
; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5],xmm1[6,7]
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5],xmm1[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1
; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX1-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX1-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
@@ -2802,16 +2777,16 @@ define <4 x i32> @test_srem_odd_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-AVX2-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
-; CHECK-AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm2 = xmm2[0,1],xmm3[2],xmm2[3]
+; CHECK-AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm2[0],xmm0[1,2],xmm2[3]
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm3
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm3[0,1],xmm2[2],xmm3[3]
; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX2-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX2-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
;
@@ -2965,23 +2940,19 @@ define <4 x i32> @test_srem_odd_even_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; CHECK-SSE2-NEXT: pmuludq %xmm4, %xmm1
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,3,2,3]
+; CHECK-SSE2-NEXT: pmuludq %xmm1, %xmm4
+; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm4[1,3,2,3]
; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
; CHECK-SSE2-NEXT: pxor %xmm1, %xmm1
-; CHECK-SSE2-NEXT: pxor %xmm5, %xmm5
-; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm5
-; CHECK-SSE2-NEXT: pand %xmm2, %xmm5
+; CHECK-SSE2-NEXT: pxor %xmm4, %xmm4
+; CHECK-SSE2-NEXT: pcmpgtd %xmm0, %xmm4
+; CHECK-SSE2-NEXT: pand %xmm2, %xmm4
; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,0,0]
; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
-; CHECK-SSE2-NEXT: paddd %xmm5, %xmm2
+; CHECK-SSE2-NEXT: paddd %xmm4, %xmm2
; CHECK-SSE2-NEXT: psubd %xmm2, %xmm3
-; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE2-NEXT: pmuludq %xmm0, %xmm2
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; CHECK-SSE2-NEXT: pmuludq {{.*}}(%rip), %xmm4
-; CHECK-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; CHECK-SSE2-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1]
+; CHECK-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [0,4294967295,4294967295,0]
+; CHECK-SSE2-NEXT: pand %xmm0, %xmm2
; CHECK-SSE2-NEXT: paddd %xmm3, %xmm2
; CHECK-SSE2-NEXT: movdqa %xmm2, %xmm3
; CHECK-SSE2-NEXT: psrad $5, %xmm3
@@ -3018,25 +2989,25 @@ define <4 x i32> @test_srem_odd_even_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-SSE41-NEXT: pmuldq %xmm0, %xmm1
; CHECK-SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,1,0]
-; CHECK-SSE41-NEXT: pmulld %xmm0, %xmm2
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: pxor %xmm2, %xmm2
+; CHECK-SSE41-NEXT: movdqa %xmm0, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm2[0,1],xmm3[2,3,4,5],xmm2[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $5, %xmm1
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm3
-; CHECK-SSE41-NEXT: psrad $3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-SSE41-NEXT: movdqa %xmm2, %xmm1
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm4
+; CHECK-SSE41-NEXT: psrad $3, %xmm4
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-SSE41-NEXT: movdqa %xmm3, %xmm1
; CHECK-SSE41-NEXT: psrad $1, %xmm1
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm3[2,3],xmm1[4,5],xmm3[6,7]
-; CHECK-SSE41-NEXT: psrld $31, %xmm2
-; CHECK-SSE41-NEXT: pxor %xmm3, %xmm3
-; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm3[4,5],xmm2[6,7]
-; CHECK-SSE41-NEXT: paddd %xmm1, %xmm2
-; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm2
-; CHECK-SSE41-NEXT: psubd %xmm2, %xmm0
-; CHECK-SSE41-NEXT: pcmpeqd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm4[2,3],xmm1[4,5],xmm4[6,7]
+; CHECK-SSE41-NEXT: psrld $31, %xmm3
+; CHECK-SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm2[4,5],xmm3[6,7]
+; CHECK-SSE41-NEXT: paddd %xmm1, %xmm3
+; CHECK-SSE41-NEXT: pmulld {{.*}}(%rip), %xmm3
+; CHECK-SSE41-NEXT: psubd %xmm3, %xmm0
+; CHECK-SSE41-NEXT: pcmpeqd %xmm2, %xmm0
; CHECK-SSE41-NEXT: psrld $31, %xmm0
; CHECK-SSE41-NEXT: retq
;
@@ -3049,21 +3020,21 @@ define <4 x i32> @test_srem_odd_even_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-AVX1-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1],xmm2[2,3],xmm1[4,5],xmm2[6,7]
-; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX1-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm2
-; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1,2,3],xmm2[4,5,6,7]
-; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm3[0,1,2,3],xmm1[4,5,6,7]
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm2 = xmm3[0,1],xmm2[2,3],xmm3[4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm2[0,1],xmm0[2,3,4,5],xmm2[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX1-NEXT: vpsrad $5, %xmm1, %xmm3
+; CHECK-AVX1-NEXT: vpsrad $3, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1,2,3],xmm3[4,5,6,7]
+; CHECK-AVX1-NEXT: vpsrad $1, %xmm1, %xmm4
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm4 = xmm4[0,1,2,3],xmm1[4,5,6,7]
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm3 = xmm4[0,1],xmm3[2,3],xmm4[4,5],xmm3[6,7]
; CHECK-AVX1-NEXT: vpsrld $31, %xmm1, %xmm1
-; CHECK-AVX1-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm3[4,5],xmm1[6,7]
-; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm2, %xmm1
+; CHECK-AVX1-NEXT: vpblendw {{.*#+}} xmm1 = xmm1[0,1,2,3],xmm2[4,5],xmm1[6,7]
+; CHECK-AVX1-NEXT: vpaddd %xmm1, %xmm3, %xmm1
; CHECK-AVX1-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX1-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX1-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX1-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX1-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX1-NEXT: retq
;
@@ -3076,16 +3047,16 @@ define <4 x i32> @test_srem_odd_even_poweroftwo_and_one(<4 x i32> %X) nounwind {
; CHECK-AVX2-NEXT: vpmuldq %xmm1, %xmm0, %xmm1
; CHECK-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm1 = xmm1[0],xmm2[1],xmm1[2],xmm2[3]
-; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm0, %xmm2
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm2
-; CHECK-AVX2-NEXT: vpxor %xmm3, %xmm3, %xmm3
-; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm2 = xmm2[0,1],xmm3[2],xmm2[3]
+; CHECK-AVX2-NEXT: vpxor %xmm2, %xmm2, %xmm2
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm2[0],xmm0[1,2],xmm2[3]
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpsrld $31, %xmm1, %xmm3
+; CHECK-AVX2-NEXT: vpblendd {{.*#+}} xmm3 = xmm3[0,1],xmm2[2],xmm3[3]
; CHECK-AVX2-NEXT: vpsravd {{.*}}(%rip), %xmm1, %xmm1
-; CHECK-AVX2-NEXT: vpaddd %xmm2, %xmm1, %xmm1
+; CHECK-AVX2-NEXT: vpaddd %xmm3, %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpmulld {{.*}}(%rip), %xmm1, %xmm1
; CHECK-AVX2-NEXT: vpsubd %xmm1, %xmm0, %xmm0
-; CHECK-AVX2-NEXT: vpcmpeqd %xmm3, %xmm0, %xmm0
+; CHECK-AVX2-NEXT: vpcmpeqd %xmm2, %xmm0, %xmm0
; CHECK-AVX2-NEXT: vpsrld $31, %xmm0, %xmm0
; CHECK-AVX2-NEXT: retq
;
diff --git a/llvm/test/CodeGen/X86/vector-idiv-sdiv-128.ll b/llvm/test/CodeGen/X86/vector-idiv-sdiv-128.ll
index 22a389568735..267ba0a0dd87 100644
--- a/llvm/test/CodeGen/X86/vector-idiv-sdiv-128.ll
+++ b/llvm/test/CodeGen/X86/vector-idiv-sdiv-128.ll
@@ -286,25 +286,16 @@ define <16 x i8> @test_div7_16i8(<16 x i8> %a) nounwind {
define <16 x i8> @test_divconstant_16i8(<16 x i8> %a) nounwind {
; SSE2-LABEL: test_divconstant_16i8:
; SSE2: # %bb.0:
-; SSE2-NEXT: movdqa %xmm0, %xmm1
-; SSE2-NEXT: movdqa %xmm0, %xmm2
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm0[8],xmm3[9],xmm0[9],xmm3[10],xmm0[10],xmm3[11],xmm0[11],xmm3[12],xmm0[12],xmm3[13],xmm0[13],xmm3[14],xmm0[14],xmm3[15],xmm0[15]
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
+; SSE2-NEXT: psraw $8, %xmm1
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm1
-; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255]
-; SSE2-NEXT: pand %xmm4, %xmm1
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: psrlw $8, %xmm1
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
+; SSE2-NEXT: psraw $8, %xmm2
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE2-NEXT: pand %xmm4, %xmm2
+; SSE2-NEXT: psrlw $8, %xmm2
; SSE2-NEXT: packuswb %xmm1, %xmm2
-; SSE2-NEXT: psraw $8, %xmm3
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
-; SSE2-NEXT: psrlw $8, %xmm3
-; SSE2-NEXT: psraw $8, %xmm0
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm0
-; SSE2-NEXT: psrlw $8, %xmm0
-; SSE2-NEXT: packuswb %xmm3, %xmm0
+; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
; SSE2-NEXT: paddb %xmm2, %xmm0
; SSE2-NEXT: movdqa %xmm0, %xmm1
; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
@@ -324,24 +315,16 @@ define <16 x i8> @test_divconstant_16i8(<16 x i8> %a) nounwind {
;
; SSE41-LABEL: test_divconstant_16i8:
; SSE41: # %bb.0:
-; SSE41-NEXT: movdqa %xmm0, %xmm1
-; SSE41-NEXT: pmovzxbw {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; SSE41-NEXT: pmovsxbw %xmm0, %xmm0
-; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm1[2,3,2,3]
-; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm1
-; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255]
-; SSE41-NEXT: pand %xmm4, %xmm1
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE41-NEXT: pand %xmm4, %xmm2
-; SSE41-NEXT: packuswb %xmm1, %xmm2
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm0
-; SSE41-NEXT: psrlw $8, %xmm0
-; SSE41-NEXT: pmovsxbw %xmm3, %xmm1
+; SSE41-NEXT: pmovsxbw %xmm0, %xmm1
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm1
; SSE41-NEXT: psrlw $8, %xmm1
-; SSE41-NEXT: packuswb %xmm1, %xmm0
-; SSE41-NEXT: paddb %xmm2, %xmm0
+; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
+; SSE41-NEXT: pmovsxbw %xmm2, %xmm2
+; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
+; SSE41-NEXT: psrlw $8, %xmm2
+; SSE41-NEXT: packuswb %xmm2, %xmm1
+; SSE41-NEXT: pand {{.*}}(%rip), %xmm0
+; SSE41-NEXT: paddb %xmm1, %xmm0
; SSE41-NEXT: movdqa %xmm0, %xmm1
; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
; SSE41-NEXT: psraw $8, %xmm1
@@ -360,23 +343,16 @@ define <16 x i8> @test_divconstant_16i8(<16 x i8> %a) nounwind {
;
; AVX1-LABEL: test_divconstant_16i8:
; AVX1: # %bb.0:
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpmovsxbw %xmm0, %xmm1
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm3 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpand %xmm2, %xmm3, %xmm2
-; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
-; AVX1-NEXT: vpmovsxbw %xmm0, %xmm2
+; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
+; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
+; AVX1-NEXT: vpmovsxbw %xmm2, %xmm2
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
-; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
-; AVX1-NEXT: vpmovsxbw %xmm0, %xmm0
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm0
-; AVX1-NEXT: vpsrlw $8, %xmm0, %xmm0
-; AVX1-NEXT: vpackuswb %xmm0, %xmm2, %xmm0
-; AVX1-NEXT: vpaddb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX1-NEXT: vpaddb %xmm0, %xmm1, %xmm0
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; AVX1-NEXT: vpsraw $8, %xmm1, %xmm1
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
@@ -393,17 +369,13 @@ define <16 x i8> @test_divconstant_16i8(<16 x i8> %a) nounwind {
;
; AVX2NOBW-LABEL: test_divconstant_16i8:
; AVX2NOBW: # %bb.0:
-; AVX2NOBW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
+; AVX2NOBW-NEXT: vpmovsxbw %xmm0, %ymm1
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm1, %ymm1
; AVX2NOBW-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2NOBW-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
-; AVX2NOBW-NEXT: vpmovsxbw %xmm0, %ymm0
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm0, %ymm0
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm0, %ymm0
-; AVX2NOBW-NEXT: vextracti128 $1, %ymm0, %xmm2
-; AVX2NOBW-NEXT: vpackuswb %xmm2, %xmm0, %xmm0
-; AVX2NOBW-NEXT: vpaddb %xmm1, %xmm0, %xmm0
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX2NOBW-NEXT: vpaddb %xmm0, %xmm1, %xmm0
; AVX2NOBW-NEXT: vpmovsxbw %xmm0, %ymm1
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
; AVX2NOBW-NEXT: vpsrlw $8, %ymm1, %ymm1
@@ -417,21 +389,19 @@ define <16 x i8> @test_divconstant_16i8(<16 x i8> %a) nounwind {
;
; AVX512BW-LABEL: test_divconstant_16i8:
; AVX512BW: # %bb.0:
-; AVX512BW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX512BW-NEXT: vmovdqa {{.*#+}} ymm1 = [2,2,1,2,3,1,2,3,3,2,1,3,2,1,1,2]
+; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm2
+; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX512BW-NEXT: vpsrlw $8, %ymm2, %ymm2
+; AVX512BW-NEXT: vpmovwb %zmm2, %ymm2
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX512BW-NEXT: vpaddb %xmm0, %xmm2, %xmm0
+; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm2
+; AVX512BW-NEXT: vpsravw %zmm1, %zmm2, %zmm1
; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
-; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm0
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm0, %ymm0
-; AVX512BW-NEXT: vpsrlw $8, %ymm0, %ymm0
-; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
-; AVX512BW-NEXT: vpaddb %xmm1, %xmm0, %xmm0
-; AVX512BW-NEXT: vpsrlw $7, %xmm0, %xmm1
-; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
-; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm0
-; AVX512BW-NEXT: vmovdqa {{.*#+}} ymm2 = [2,2,1,2,3,1,2,3,3,2,1,3,2,1,1,2]
-; AVX512BW-NEXT: vpsravw %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
-; AVX512BW-NEXT: vpaddb %xmm1, %xmm0, %xmm0
+; AVX512BW-NEXT: vpsrlw $7, %xmm0, %xmm0
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm0
+; AVX512BW-NEXT: vpaddb %xmm0, %xmm1, %xmm0
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
%res = sdiv <16 x i8> %a, <i8 7, i8 8, i8 9, i8 10,i8 11, i8 12, i8 13, i8 14, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9,i8 9, i8 7>
@@ -777,161 +747,137 @@ define <16 x i8> @test_rem7_16i8(<16 x i8> %a) nounwind {
define <16 x i8> @test_remconstant_16i8(<16 x i8> %a) nounwind {
; SSE2-LABEL: test_remconstant_16i8:
; SSE2: # %bb.0:
-; SSE2-NEXT: movdqa %xmm0, %xmm2
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
-; SSE2-NEXT: pand %xmm1, %xmm2
-; SSE2-NEXT: movdqa %xmm0, %xmm3
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
-; SSE2-NEXT: pand %xmm1, %xmm3
-; SSE2-NEXT: packuswb %xmm2, %xmm3
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm4 = xmm4[8],xmm0[8],xmm4[9],xmm0[9],xmm4[10],xmm0[10],xmm4[11],xmm0[11],xmm4[12],xmm0[12],xmm4[13],xmm0[13],xmm4[14],xmm0[14],xmm4[15],xmm0[15]
-; SSE2-NEXT: psraw $8, %xmm4
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm4
-; SSE2-NEXT: psrlw $8, %xmm4
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
+; SSE2-NEXT: psraw $8, %xmm1
+; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm1
+; SSE2-NEXT: psrlw $8, %xmm1
; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
; SSE2-NEXT: psraw $8, %xmm2
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
; SSE2-NEXT: psrlw $8, %xmm2
-; SSE2-NEXT: packuswb %xmm4, %xmm2
-; SSE2-NEXT: paddb %xmm3, %xmm2
-; SSE2-NEXT: movdqa %xmm2, %xmm3
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
+; SSE2-NEXT: packuswb %xmm1, %xmm2
+; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [255,255,0,0,255,0,0,255,255,0,0,255,0,0,0,255]
+; SSE2-NEXT: pand %xmm0, %xmm1
+; SSE2-NEXT: paddb %xmm2, %xmm1
+; SSE2-NEXT: movdqa %xmm1, %xmm2
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
+; SSE2-NEXT: psraw $8, %xmm2
+; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
+; SSE2-NEXT: psrlw $8, %xmm2
+; SSE2-NEXT: movdqa %xmm1, %xmm3
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3],xmm3[4],xmm1[4],xmm3[5],xmm1[5],xmm3[6],xmm1[6],xmm3[7],xmm1[7]
; SSE2-NEXT: psraw $8, %xmm3
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
; SSE2-NEXT: psrlw $8, %xmm3
-; SSE2-NEXT: movdqa %xmm2, %xmm4
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
-; SSE2-NEXT: psraw $8, %xmm4
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm4
-; SSE2-NEXT: psrlw $8, %xmm4
-; SSE2-NEXT: packuswb %xmm3, %xmm4
-; SSE2-NEXT: psrlw $7, %xmm2
-; SSE2-NEXT: pand {{.*}}(%rip), %xmm2
-; SSE2-NEXT: paddb %xmm4, %xmm2
-; SSE2-NEXT: movdqa %xmm2, %xmm3
-; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm3
-; SSE2-NEXT: pand %xmm1, %xmm3
-; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: packuswb %xmm2, %xmm3
+; SSE2-NEXT: psrlw $7, %xmm1
+; SSE2-NEXT: pand {{.*}}(%rip), %xmm1
+; SSE2-NEXT: paddb %xmm3, %xmm1
+; SSE2-NEXT: movdqa %xmm1, %xmm2
+; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE2-NEXT: pand %xmm1, %xmm2
-; SSE2-NEXT: packuswb %xmm3, %xmm2
-; SSE2-NEXT: psubb %xmm2, %xmm0
+; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
+; SSE2-NEXT: pand %xmm3, %xmm2
+; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; SSE2-NEXT: pmullw {{.*}}(%rip), %xmm1
+; SSE2-NEXT: pand %xmm3, %xmm1
+; SSE2-NEXT: packuswb %xmm2, %xmm1
+; SSE2-NEXT: psubb %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSE41-LABEL: test_remconstant_16i8:
; SSE41: # %bb.0:
-; SSE41-NEXT: movdqa %xmm0, %xmm2
-; SSE41-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
-; SSE41-NEXT: pand %xmm1, %xmm2
-; SSE41-NEXT: pmovzxbw {{.*#+}} xmm3 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm3
-; SSE41-NEXT: pand %xmm1, %xmm3
-; SSE41-NEXT: packuswb %xmm2, %xmm3
; SSE41-NEXT: pmovsxbw %xmm0, %xmm2
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
; SSE41-NEXT: psrlw $8, %xmm2
-; SSE41-NEXT: pshufd {{.*#+}} xmm4 = xmm0[2,3,2,3]
-; SSE41-NEXT: pmovsxbw %xmm4, %xmm4
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm4
-; SSE41-NEXT: psrlw $8, %xmm4
-; SSE41-NEXT: packuswb %xmm4, %xmm2
-; SSE41-NEXT: paddb %xmm3, %xmm2
-; SSE41-NEXT: movdqa %xmm2, %xmm3
-; SSE41-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
+; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
+; SSE41-NEXT: pmovsxbw %xmm1, %xmm1
+; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm1
+; SSE41-NEXT: psrlw $8, %xmm1
+; SSE41-NEXT: packuswb %xmm1, %xmm2
+; SSE41-NEXT: movdqa {{.*#+}} xmm1 = [255,255,0,0,255,0,0,255,255,0,0,255,0,0,0,255]
+; SSE41-NEXT: pand %xmm0, %xmm1
+; SSE41-NEXT: paddb %xmm2, %xmm1
+; SSE41-NEXT: movdqa %xmm1, %xmm2
+; SSE41-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]
+; SSE41-NEXT: psraw $8, %xmm2
+; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
+; SSE41-NEXT: psrlw $8, %xmm2
+; SSE41-NEXT: movdqa %xmm1, %xmm3
+; SSE41-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]
; SSE41-NEXT: psraw $8, %xmm3
; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm3
; SSE41-NEXT: psrlw $8, %xmm3
-; SSE41-NEXT: movdqa %xmm2, %xmm4
-; SSE41-NEXT: punpcklbw {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1],xmm4[2],xmm2[2],xmm4[3],xmm2[3],xmm4[4],xmm2[4],xmm4[5],xmm2[5],xmm4[6],xmm2[6],xmm4[7],xmm2[7]
-; SSE41-NEXT: psraw $8, %xmm4
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm4
-; SSE41-NEXT: psrlw $8, %xmm4
-; SSE41-NEXT: packuswb %xmm3, %xmm4
-; SSE41-NEXT: psrlw $7, %xmm2
-; SSE41-NEXT: pand {{.*}}(%rip), %xmm2
-; SSE41-NEXT: paddb %xmm4, %xmm2
-; SSE41-NEXT: pmovzxbw {{.*#+}} xmm3 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
-; SSE41-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
-; SSE41-NEXT: pand %xmm1, %xmm2
-; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm3
-; SSE41-NEXT: pand %xmm1, %xmm3
; SSE41-NEXT: packuswb %xmm2, %xmm3
-; SSE41-NEXT: psubb %xmm3, %xmm0
+; SSE41-NEXT: psrlw $7, %xmm1
+; SSE41-NEXT: pand {{.*}}(%rip), %xmm1
+; SSE41-NEXT: paddb %xmm3, %xmm1
+; SSE41-NEXT: pmovzxbw {{.*#+}} xmm2 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; SSE41-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm1
+; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
+; SSE41-NEXT: pand %xmm3, %xmm1
+; SSE41-NEXT: pmullw {{.*}}(%rip), %xmm2
+; SSE41-NEXT: pand %xmm3, %xmm2
+; SSE41-NEXT: packuswb %xmm1, %xmm2
+; SSE41-NEXT: psubb %xmm2, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: test_remconstant_16i8:
; AVX1: # %bb.0:
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpmovsxbw %xmm0, %xmm1
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
-; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm3 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpand %xmm2, %xmm3, %xmm3
-; AVX1-NEXT: vpackuswb %xmm1, %xmm3, %xmm1
-; AVX1-NEXT: vpmovsxbw %xmm0, %xmm3
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
-; AVX1-NEXT: vpshufd {{.*#+}} xmm4 = xmm0[2,3,2,3]
-; AVX1-NEXT: vpmovsxbw %xmm4, %xmm4
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpackuswb %xmm4, %xmm3, %xmm3
-; AVX1-NEXT: vpaddb %xmm1, %xmm3, %xmm1
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
+; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
+; AVX1-NEXT: vpmovsxbw %xmm2, %xmm2
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
+; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
+; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm2
+; AVX1-NEXT: vpaddb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpsraw $8, %xmm2, %xmm2
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
+; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
+; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm3 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpsraw $8, %xmm3, %xmm3
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
-; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm4 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
-; AVX1-NEXT: vpsraw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpackuswb %xmm3, %xmm4, %xmm3
+; AVX1-NEXT: vpackuswb %xmm2, %xmm3, %xmm2
; AVX1-NEXT: vpsrlw $7, %xmm1, %xmm1
; AVX1-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
-; AVX1-NEXT: vpaddb %xmm1, %xmm3, %xmm1
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpand %xmm2, %xmm3, %xmm3
+; AVX1-NEXT: vpaddb %xmm1, %xmm2, %xmm1
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
+; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
+; AVX1-NEXT: vpand %xmm3, %xmm2, %xmm2
; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
-; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
-; AVX1-NEXT: vpackuswb %xmm3, %xmm1, %xmm1
+; AVX1-NEXT: vpand %xmm3, %xmm1, %xmm1
+; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpsubb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: retq
;
; AVX2NOBW-LABEL: test_remconstant_16i8:
; AVX2NOBW: # %bb.0:
-; AVX2NOBW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
+; AVX2NOBW-NEXT: vpmovsxbw %xmm0, %ymm1
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX2NOBW-NEXT: vmovdqa {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX2NOBW-NEXT: vpand %ymm2, %ymm1, %ymm1
-; AVX2NOBW-NEXT: vextracti128 $1, %ymm1, %xmm3
-; AVX2NOBW-NEXT: vpackuswb %xmm3, %xmm1, %xmm1
-; AVX2NOBW-NEXT: vpmovsxbw %xmm0, %ymm3
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vextracti128 $1, %ymm3, %xmm4
-; AVX2NOBW-NEXT: vpackuswb %xmm4, %xmm3, %xmm3
-; AVX2NOBW-NEXT: vpaddb %xmm1, %xmm3, %xmm1
-; AVX2NOBW-NEXT: vpmovsxbw %xmm1, %ymm3
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vextracti128 $1, %ymm3, %xmm4
-; AVX2NOBW-NEXT: vpackuswb %xmm4, %xmm3, %xmm3
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vextracti128 $1, %ymm1, %xmm2
+; AVX2NOBW-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm2
+; AVX2NOBW-NEXT: vpaddb %xmm2, %xmm1, %xmm1
+; AVX2NOBW-NEXT: vpmovsxbw %xmm1, %ymm2
+; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vextracti128 $1, %ymm2, %xmm3
+; AVX2NOBW-NEXT: vpackuswb %xmm3, %xmm2, %xmm2
; AVX2NOBW-NEXT: vpsrlw $7, %xmm1, %xmm1
; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %xmm1, %xmm1
-; AVX2NOBW-NEXT: vpaddb %xmm1, %xmm3, %xmm1
+; AVX2NOBW-NEXT: vpaddb %xmm1, %xmm2, %xmm1
; AVX2NOBW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX2NOBW-NEXT: vpand %ymm2, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm1, %ymm1
; AVX2NOBW-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2NOBW-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
; AVX2NOBW-NEXT: vpsubb %xmm1, %xmm0, %xmm0
@@ -940,20 +886,18 @@ define <16 x i8> @test_remconstant_16i8(<16 x i8> %a) nounwind {
;
; AVX512BW-LABEL: test_remconstant_16i8:
; AVX512BW: # %bb.0:
-; AVX512BW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
+; AVX512BW-NEXT: vmovdqa {{.*#+}} ymm1 = [2,2,1,2,3,1,2,3,3,2,1,3,2,1,1,2]
; AVX512BW-NEXT: vpmovsxbw %xmm0, %ymm2
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
; AVX512BW-NEXT: vpsrlw $8, %ymm2, %ymm2
; AVX512BW-NEXT: vpmovwb %zmm2, %ymm2
-; AVX512BW-NEXT: vpaddb %xmm1, %xmm2, %xmm1
-; AVX512BW-NEXT: vpsrlw $7, %xmm1, %xmm2
-; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm2, %xmm2
-; AVX512BW-NEXT: vpmovsxbw %xmm1, %ymm1
-; AVX512BW-NEXT: vmovdqa {{.*#+}} ymm3 = [2,2,1,2,3,1,2,3,3,2,1,3,2,1,1,2]
-; AVX512BW-NEXT: vpsravw %zmm3, %zmm1, %zmm1
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm0, %xmm3
+; AVX512BW-NEXT: vpaddb %xmm3, %xmm2, %xmm2
+; AVX512BW-NEXT: vpmovsxbw %xmm2, %ymm3
+; AVX512BW-NEXT: vpsravw %zmm1, %zmm3, %zmm1
; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
+; AVX512BW-NEXT: vpsrlw $7, %xmm2, %xmm2
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %xmm2, %xmm2
; AVX512BW-NEXT: vpaddb %xmm2, %xmm1, %xmm1
; AVX512BW-NEXT: vpmovzxbw {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
diff --git a/llvm/test/CodeGen/X86/vector-idiv-sdiv-256.ll b/llvm/test/CodeGen/X86/vector-idiv-sdiv-256.ll
index e06140f988d5..d177f0f2a417 100644
--- a/llvm/test/CodeGen/X86/vector-idiv-sdiv-256.ll
+++ b/llvm/test/CodeGen/X86/vector-idiv-sdiv-256.ll
@@ -248,44 +248,31 @@ define <32 x i8> @test_div7_32i8(<32 x i8> %a) nounwind {
define <32 x i8> @test_divconstant_32i8(<32 x i8> %a) nounwind {
; AVX1-LABEL: test_divconstant_32i8:
; AVX1: # %bb.0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm3
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
-; AVX1-NEXT: vpand %xmm1, %xmm3, %xmm3
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm4 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm1, %xmm4, %xmm4
-; AVX1-NEXT: vpackuswb %xmm3, %xmm4, %xmm3
-; AVX1-NEXT: vpmovsxbw %xmm2, %xmm4
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
-; AVX1-NEXT: vpmovsxbw %xmm2, %xmm2
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT: vpmovsxbw %xmm1, %xmm2
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
-; AVX1-NEXT: vpackuswb %xmm2, %xmm4, %xmm2
-; AVX1-NEXT: vpaddb %xmm3, %xmm2, %xmm2
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; AVX1-NEXT: vpmovsxbw %xmm1, %xmm1
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
+; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
+; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
+; AVX1-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm2
+; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm3
+; AVX1-NEXT: vpaddb %xmm3, %xmm1, %xmm1
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; AVX1-NEXT: vpsraw $8, %xmm3, %xmm3
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
-; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm4 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm4 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpsraw $8, %xmm4, %xmm4
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
; AVX1-NEXT: vpackuswb %xmm3, %xmm4, %xmm3
-; AVX1-NEXT: vpsrlw $7, %xmm2, %xmm2
+; AVX1-NEXT: vpsrlw $7, %xmm1, %xmm1
; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-; AVX1-NEXT: vpand %xmm4, %xmm2, %xmm2
-; AVX1-NEXT: vpaddb %xmm2, %xmm3, %xmm2
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpand %xmm1, %xmm3, %xmm3
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm5 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
-; AVX1-NEXT: vpand %xmm1, %xmm5, %xmm1
-; AVX1-NEXT: vpackuswb %xmm3, %xmm1, %xmm1
+; AVX1-NEXT: vpand %xmm4, %xmm1, %xmm1
+; AVX1-NEXT: vpaddb %xmm1, %xmm3, %xmm1
; AVX1-NEXT: vpmovsxbw %xmm0, %xmm3
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
@@ -294,42 +281,35 @@ define <32 x i8> @test_divconstant_32i8(<32 x i8> %a) nounwind {
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm0, %xmm0
; AVX1-NEXT: vpsrlw $8, %xmm0, %xmm0
; AVX1-NEXT: vpackuswb %xmm0, %xmm3, %xmm0
-; AVX1-NEXT: vpaddb %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpsraw $8, %xmm1, %xmm1
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
-; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
+; AVX1-NEXT: vpaddb %xmm2, %xmm0, %xmm0
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpsraw $8, %xmm2, %xmm2
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
+; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm3 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpsraw $8, %xmm3, %xmm3
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm3
-; AVX1-NEXT: vpackuswb %xmm1, %xmm3, %xmm1
+; AVX1-NEXT: vpackuswb %xmm2, %xmm3, %xmm2
; AVX1-NEXT: vpsrlw $7, %xmm0, %xmm0
; AVX1-NEXT: vpand %xmm4, %xmm0, %xmm0
-; AVX1-NEXT: vpaddb %xmm0, %xmm1, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT: vpaddb %xmm0, %xmm2, %xmm0
+; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2NOBW-LABEL: test_divconstant_32i8:
; AVX2NOBW: # %bb.0:
; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX2NOBW-NEXT: vpsraw $8, %ymm1, %ymm1
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX2NOBW-NEXT: vmovdqa {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX2NOBW-NEXT: vpand %ymm2, %ymm1, %ymm1
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpand %ymm2, %ymm3, %ymm2
-; AVX2NOBW-NEXT: vpackuswb %ymm1, %ymm2, %ymm1
-; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2NOBW-NEXT: vpsraw $8, %ymm2, %ymm2
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
; AVX2NOBW-NEXT: vpsrlw $8, %ymm2, %ymm2
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2NOBW-NEXT: vpsraw $8, %ymm0, %ymm0
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm0, %ymm0
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm0, %ymm0
-; AVX2NOBW-NEXT: vpackuswb %ymm2, %ymm0, %ymm0
-; AVX2NOBW-NEXT: vpaddb %ymm1, %ymm0, %ymm0
+; AVX2NOBW-NEXT: vpackuswb %ymm1, %ymm2, %ymm1
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm0, %ymm0
+; AVX2NOBW-NEXT: vpaddb %ymm0, %ymm1, %ymm0
; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2NOBW-NEXT: vpsraw $8, %ymm1, %ymm1
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
@@ -346,9 +326,7 @@ define <32 x i8> @test_divconstant_32i8(<32 x i8> %a) nounwind {
;
; AVX512BW-LABEL: test_divconstant_32i8:
; AVX512BW: # %bb.0:
-; AVX512BW-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %ymm0, %ymm1
; AVX512BW-NEXT: vpmovsxbw %ymm0, %zmm0
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm0, %zmm0
; AVX512BW-NEXT: vpsrlw $8, %zmm0, %zmm0
@@ -676,52 +654,40 @@ define <32 x i8> @test_rem7_32i8(<32 x i8> %a) nounwind {
define <32 x i8> @test_remconstant_32i8(<32 x i8> %a) nounwind {
; AVX1-LABEL: test_remconstant_32i8:
; AVX1: # %bb.0:
-; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm3
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
-; AVX1-NEXT: vpand %xmm1, %xmm3, %xmm3
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm4 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm1, %xmm4, %xmm4
-; AVX1-NEXT: vpackuswb %xmm3, %xmm4, %xmm3
-; AVX1-NEXT: vpmovsxbw %xmm2, %xmm4
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpshufd {{.*#+}} xmm5 = xmm2[2,3,2,3]
-; AVX1-NEXT: vpmovsxbw %xmm5, %xmm5
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
-; AVX1-NEXT: vpsrlw $8, %xmm5, %xmm5
-; AVX1-NEXT: vpackuswb %xmm5, %xmm4, %xmm4
-; AVX1-NEXT: vpaddb %xmm3, %xmm4, %xmm3
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
+; AVX1-NEXT: vpmovsxbw %xmm3, %xmm1
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
+; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
+; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm3[2,3,2,3]
+; AVX1-NEXT: vpmovsxbw %xmm2, %xmm2
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm2
+; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
+; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm2
+; AVX1-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm1
+; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm4
+; AVX1-NEXT: vpaddb %xmm4, %xmm2, %xmm2
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; AVX1-NEXT: vpsraw $8, %xmm4, %xmm4
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
; AVX1-NEXT: vpsrlw $8, %xmm4, %xmm4
-; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm5 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm5 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpsraw $8, %xmm5, %xmm5
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
; AVX1-NEXT: vpsrlw $8, %xmm5, %xmm5
-; AVX1-NEXT: vpackuswb %xmm4, %xmm5, %xmm4
-; AVX1-NEXT: vpsrlw $7, %xmm3, %xmm5
-; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-; AVX1-NEXT: vpand %xmm3, %xmm5, %xmm5
-; AVX1-NEXT: vpaddb %xmm5, %xmm4, %xmm4
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm5 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
-; AVX1-NEXT: vpand %xmm1, %xmm5, %xmm5
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm4 = xmm4[0],zero,xmm4[1],zero,xmm4[2],zero,xmm4[3],zero,xmm4[4],zero,xmm4[5],zero,xmm4[6],zero,xmm4[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm1, %xmm4, %xmm4
-; AVX1-NEXT: vpackuswb %xmm5, %xmm4, %xmm4
-; AVX1-NEXT: vpsubb %xmm4, %xmm2, %xmm2
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm1, %xmm4, %xmm4
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm5 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
+; AVX1-NEXT: vpackuswb %xmm4, %xmm5, %xmm5
+; AVX1-NEXT: vpsrlw $7, %xmm2, %xmm2
+; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; AVX1-NEXT: vpand %xmm4, %xmm2, %xmm2
+; AVX1-NEXT: vpaddb %xmm2, %xmm5, %xmm5
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm2, %xmm6
+; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
+; AVX1-NEXT: vpand %xmm2, %xmm6, %xmm6
+; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm5 = xmm5[0],zero,xmm5[1],zero,xmm5[2],zero,xmm5[3],zero,xmm5[4],zero,xmm5[5],zero,xmm5[6],zero,xmm5[7],zero
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
-; AVX1-NEXT: vpand %xmm1, %xmm5, %xmm5
-; AVX1-NEXT: vpackuswb %xmm4, %xmm5, %xmm4
+; AVX1-NEXT: vpand %xmm2, %xmm5, %xmm5
+; AVX1-NEXT: vpackuswb %xmm6, %xmm5, %xmm5
+; AVX1-NEXT: vpsubb %xmm5, %xmm3, %xmm3
; AVX1-NEXT: vpmovsxbw %xmm0, %xmm5
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
; AVX1-NEXT: vpsrlw $8, %xmm5, %xmm5
@@ -730,77 +696,69 @@ define <32 x i8> @test_remconstant_32i8(<32 x i8> %a) nounwind {
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm6, %xmm6
; AVX1-NEXT: vpsrlw $8, %xmm6, %xmm6
; AVX1-NEXT: vpackuswb %xmm6, %xmm5, %xmm5
-; AVX1-NEXT: vpaddb %xmm4, %xmm5, %xmm4
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm5 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpaddb %xmm1, %xmm5, %xmm1
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm5 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; AVX1-NEXT: vpsraw $8, %xmm5, %xmm5
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm5, %xmm5
; AVX1-NEXT: vpsrlw $8, %xmm5, %xmm5
-; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm6 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
+; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm6 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpsraw $8, %xmm6, %xmm6
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm6, %xmm6
; AVX1-NEXT: vpsrlw $8, %xmm6, %xmm6
; AVX1-NEXT: vpackuswb %xmm5, %xmm6, %xmm5
-; AVX1-NEXT: vpsrlw $7, %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm3, %xmm4, %xmm3
-; AVX1-NEXT: vpaddb %xmm3, %xmm5, %xmm3
-; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
+; AVX1-NEXT: vpsrlw $7, %xmm1, %xmm1
+; AVX1-NEXT: vpand %xmm4, %xmm1, %xmm1
+; AVX1-NEXT: vpaddb %xmm1, %xmm5, %xmm1
+; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm4, %xmm4
-; AVX1-NEXT: vpand %xmm1, %xmm4, %xmm4
-; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm3 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
-; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm3, %xmm3
-; AVX1-NEXT: vpand %xmm1, %xmm3, %xmm1
+; AVX1-NEXT: vpand %xmm2, %xmm4, %xmm4
+; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
+; AVX1-NEXT: vpmullw {{.*}}(%rip), %xmm1, %xmm1
+; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpackuswb %xmm4, %xmm1, %xmm1
; AVX1-NEXT: vpsubb %xmm1, %xmm0, %xmm0
-; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2NOBW-LABEL: test_remconstant_32i8:
; AVX2NOBW: # %bb.0:
; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm2
-; AVX2NOBW-NEXT: vmovdqa {{.*#+}} ymm1 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX2NOBW-NEXT: vpand %ymm1, %ymm2, %ymm2
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpand %ymm1, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpackuswb %ymm2, %ymm3, %ymm2
-; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX2NOBW-NEXT: vpsraw $8, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2NOBW-NEXT: vpsraw $8, %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpackuswb %ymm3, %ymm4, %ymm3
-; AVX2NOBW-NEXT: vpaddb %ymm2, %ymm3, %ymm2
-; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX2NOBW-NEXT: vpsraw $8, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2NOBW-NEXT: vpsraw $8, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpackuswb %ymm1, %ymm2, %ymm1
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm0, %ymm2
+; AVX2NOBW-NEXT: vpaddb %ymm2, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX2NOBW-NEXT: vpsraw $8, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpsrlw $8, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm3 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2NOBW-NEXT: vpsraw $8, %ymm3, %ymm3
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
; AVX2NOBW-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX2NOBW-NEXT: vpsraw $8, %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpsrlw $8, %ymm4, %ymm4
-; AVX2NOBW-NEXT: vpackuswb %ymm3, %ymm4, %ymm3
-; AVX2NOBW-NEXT: vpsrlw $7, %ymm2, %ymm2
-; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm2, %ymm2
-; AVX2NOBW-NEXT: vpaddb %ymm2, %ymm3, %ymm2
-; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpand %ymm1, %ymm3, %ymm3
-; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2NOBW-NEXT: vpackuswb %ymm2, %ymm3, %ymm2
+; AVX2NOBW-NEXT: vpsrlw $7, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpand {{.*}}(%rip), %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpaddb %ymm1, %ymm2, %ymm1
+; AVX2NOBW-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
-; AVX2NOBW-NEXT: vpand %ymm1, %ymm2, %ymm1
-; AVX2NOBW-NEXT: vpackuswb %ymm3, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX2NOBW-NEXT: vpand %ymm3, %ymm2, %ymm2
+; AVX2NOBW-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX2NOBW-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpand %ymm3, %ymm1, %ymm1
+; AVX2NOBW-NEXT: vpackuswb %ymm2, %ymm1, %ymm1
; AVX2NOBW-NEXT: vpsubb %ymm1, %ymm0, %ymm0
; AVX2NOBW-NEXT: retq
;
; AVX512BW-LABEL: test_remconstant_32i8:
; AVX512BW: # %bb.0:
-; AVX512BW-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
+; AVX512BW-NEXT: vpand {{.*}}(%rip), %ymm0, %ymm1
; AVX512BW-NEXT: vpmovsxbw %ymm0, %zmm2
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm2, %zmm2
; AVX512BW-NEXT: vpsrlw $8, %zmm2, %zmm2
diff --git a/llvm/test/CodeGen/X86/vector-idiv-sdiv-512.ll b/llvm/test/CodeGen/X86/vector-idiv-sdiv-512.ll
index c3fa89445c2f..342e83d723c9 100644
--- a/llvm/test/CodeGen/X86/vector-idiv-sdiv-512.ll
+++ b/llvm/test/CodeGen/X86/vector-idiv-sdiv-512.ll
@@ -203,45 +203,32 @@ define <64 x i8> @test_div7_64i8(<64 x i8> %a) nounwind {
define <64 x i8> @test_divconstant_64i8(<64 x i8> %a) nounwind {
; AVX512F-LABEL: test_divconstant_64i8:
; AVX512F: # %bb.0:
-; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm2
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm3
-; AVX512F-NEXT: vmovdqa {{.*#+}} ymm1 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512F-NEXT: vpand %ymm1, %ymm3, %ymm3
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm1, %ymm4, %ymm4
-; AVX512F-NEXT: vpackuswb %ymm3, %ymm4, %ymm3
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpsraw $8, %ymm4, %ymm4
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpsrlw $8, %ymm4, %ymm4
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm1
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm2, %ymm2
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
; AVX512F-NEXT: vpsrlw $8, %ymm2, %ymm2
-; AVX512F-NEXT: vpackuswb %ymm4, %ymm2, %ymm2
-; AVX512F-NEXT: vpaddb %ymm3, %ymm2, %ymm2
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpsraw $8, %ymm1, %ymm1
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX512F-NEXT: vpackuswb %ymm2, %ymm1, %ymm1
+; AVX512F-NEXT: vpandq {{.*}}(%rip), %zmm0, %zmm2
+; AVX512F-NEXT: vextracti64x4 $1, %zmm2, %ymm3
+; AVX512F-NEXT: vpaddb %ymm3, %ymm1, %ymm1
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm3, %ymm3
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
; AVX512F-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT: vpsraw $8, %ymm4, %ymm4
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
; AVX512F-NEXT: vpsrlw $8, %ymm4, %ymm4
; AVX512F-NEXT: vpackuswb %ymm3, %ymm4, %ymm3
-; AVX512F-NEXT: vpsrlw $7, %ymm2, %ymm2
+; AVX512F-NEXT: vpsrlw $7, %ymm1, %ymm1
; AVX512F-NEXT: vmovdqa {{.*#+}} ymm4 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-; AVX512F-NEXT: vpand %ymm4, %ymm2, %ymm2
-; AVX512F-NEXT: vpaddb %ymm2, %ymm3, %ymm2
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX512F-NEXT: vpand %ymm1, %ymm3, %ymm3
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
-; AVX512F-NEXT: vpand %ymm1, %ymm5, %ymm1
-; AVX512F-NEXT: vpackuswb %ymm3, %ymm1, %ymm1
+; AVX512F-NEXT: vpand %ymm4, %ymm1, %ymm1
+; AVX512F-NEXT: vpaddb %ymm1, %ymm3, %ymm1
; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm3, %ymm3
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
@@ -251,42 +238,35 @@ define <64 x i8> @test_divconstant_64i8(<64 x i8> %a) nounwind {
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm0, %ymm0
; AVX512F-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512F-NEXT: vpackuswb %ymm3, %ymm0, %ymm0
-; AVX512F-NEXT: vpaddb %ymm1, %ymm0, %ymm0
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpsraw $8, %ymm1, %ymm1
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
-; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX512F-NEXT: vpaddb %ymm2, %ymm0, %ymm0
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpsraw $8, %ymm2, %ymm2
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX512F-NEXT: vpsrlw $8, %ymm2, %ymm2
; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm3 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT: vpsraw $8, %ymm3, %ymm3
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
; AVX512F-NEXT: vpsrlw $8, %ymm3, %ymm3
-; AVX512F-NEXT: vpackuswb %ymm1, %ymm3, %ymm1
+; AVX512F-NEXT: vpackuswb %ymm2, %ymm3, %ymm2
; AVX512F-NEXT: vpsrlw $7, %ymm0, %ymm0
; AVX512F-NEXT: vpand %ymm4, %ymm0, %ymm0
-; AVX512F-NEXT: vpaddb %ymm0, %ymm1, %ymm0
-; AVX512F-NEXT: vinserti64x4 $1, %ymm2, %zmm0, %zmm0
+; AVX512F-NEXT: vpaddb %ymm0, %ymm2, %ymm0
+; AVX512F-NEXT: vinserti64x4 $1, %ymm1, %zmm0, %zmm0
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: test_divconstant_64i8:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpsraw $8, %zmm1, %zmm1
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1
-; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm3, %zmm3
-; AVX512BW-NEXT: vpandq %zmm2, %zmm3, %zmm2
-; AVX512BW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1
-; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpsrlw $8, %zmm1, %zmm1
+; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT: vpsraw $8, %zmm2, %zmm2
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm2, %zmm2
; AVX512BW-NEXT: vpsrlw $8, %zmm2, %zmm2
-; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
-; AVX512BW-NEXT: vpsraw $8, %zmm0, %zmm0
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm0, %zmm0
-; AVX512BW-NEXT: vpsrlw $8, %zmm0, %zmm0
-; AVX512BW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0
-; AVX512BW-NEXT: vpaddb %zmm1, %zmm0, %zmm0
+; AVX512BW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1
+; AVX512BW-NEXT: vpandq {{.*}}(%rip), %zmm0, %zmm0
+; AVX512BW-NEXT: vpaddb %zmm0, %zmm1, %zmm0
; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
; AVX512BW-NEXT: vpsraw $8, %zmm1, %zmm1
; AVX512BW-NEXT: vpsllvw {{.*}}(%rip), %zmm1, %zmm1
@@ -560,15 +540,19 @@ define <64 x i8> @test_rem7_64i8(<64 x i8> %a) nounwind {
define <64 x i8> @test_remconstant_64i8(<64 x i8> %a) nounwind {
; AVX512F-LABEL: test_remconstant_64i8:
; AVX512F: # %bb.0:
-; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm2
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm3
-; AVX512F-NEXT: vmovdqa {{.*#+}} ymm1 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512F-NEXT: vpand %ymm1, %ymm3, %ymm3
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm1, %ymm4, %ymm4
-; AVX512F-NEXT: vpackuswb %ymm3, %ymm4, %ymm3
+; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm3
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpsraw $8, %ymm1, %ymm1
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpsraw $8, %ymm2, %ymm2
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm2
+; AVX512F-NEXT: vpsrlw $8, %ymm2, %ymm2
+; AVX512F-NEXT: vpackuswb %ymm1, %ymm2, %ymm2
+; AVX512F-NEXT: vpandq {{.*}}(%rip), %zmm0, %zmm1
+; AVX512F-NEXT: vextracti64x4 $1, %zmm1, %ymm4
+; AVX512F-NEXT: vpaddb %ymm4, %ymm2, %ymm2
; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm4, %ymm4
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
@@ -577,36 +561,20 @@ define <64 x i8> @test_remconstant_64i8(<64 x i8> %a) nounwind {
; AVX512F-NEXT: vpsraw $8, %ymm5, %ymm5
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
; AVX512F-NEXT: vpsrlw $8, %ymm5, %ymm5
-; AVX512F-NEXT: vpackuswb %ymm4, %ymm5, %ymm4
-; AVX512F-NEXT: vpaddb %ymm3, %ymm4, %ymm3
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpsraw $8, %ymm4, %ymm4
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpsrlw $8, %ymm4, %ymm4
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpsraw $8, %ymm5, %ymm5
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
-; AVX512F-NEXT: vpsrlw $8, %ymm5, %ymm5
-; AVX512F-NEXT: vpackuswb %ymm4, %ymm5, %ymm4
-; AVX512F-NEXT: vpsrlw $7, %ymm3, %ymm5
-; AVX512F-NEXT: vmovdqa {{.*#+}} ymm3 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
-; AVX512F-NEXT: vpand %ymm3, %ymm5, %ymm5
-; AVX512F-NEXT: vpaddb %ymm5, %ymm4, %ymm4
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm5 = ymm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
-; AVX512F-NEXT: vpand %ymm1, %ymm5, %ymm5
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm4 = ymm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm1, %ymm4, %ymm4
-; AVX512F-NEXT: vpackuswb %ymm5, %ymm4, %ymm4
-; AVX512F-NEXT: vpsubb %ymm4, %ymm2, %ymm2
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm1, %ymm4, %ymm4
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpackuswb %ymm4, %ymm5, %ymm5
+; AVX512F-NEXT: vpsrlw $7, %ymm2, %ymm2
+; AVX512F-NEXT: vmovdqa {{.*#+}} ymm4 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
+; AVX512F-NEXT: vpand %ymm4, %ymm2, %ymm2
+; AVX512F-NEXT: vpaddb %ymm2, %ymm5, %ymm5
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm2, %ymm6
+; AVX512F-NEXT: vmovdqa {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512F-NEXT: vpand %ymm2, %ymm6, %ymm6
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
-; AVX512F-NEXT: vpand %ymm1, %ymm5, %ymm5
-; AVX512F-NEXT: vpackuswb %ymm4, %ymm5, %ymm4
+; AVX512F-NEXT: vpand %ymm2, %ymm5, %ymm5
+; AVX512F-NEXT: vpackuswb %ymm6, %ymm5, %ymm5
+; AVX512F-NEXT: vpsubb %ymm5, %ymm3, %ymm3
; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm5 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm5, %ymm5
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
@@ -616,69 +584,63 @@ define <64 x i8> @test_remconstant_64i8(<64 x i8> %a) nounwind {
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm6, %ymm6
; AVX512F-NEXT: vpsrlw $8, %ymm6, %ymm6
; AVX512F-NEXT: vpackuswb %ymm5, %ymm6, %ymm5
-; AVX512F-NEXT: vpaddb %ymm4, %ymm5, %ymm4
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm5 = ymm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpaddb %ymm1, %ymm5, %ymm1
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm5 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpsraw $8, %ymm5, %ymm5
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm5, %ymm5
; AVX512F-NEXT: vpsrlw $8, %ymm5, %ymm5
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm6 = ymm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm6 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT: vpsraw $8, %ymm6, %ymm6
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm6, %ymm6
; AVX512F-NEXT: vpsrlw $8, %ymm6, %ymm6
; AVX512F-NEXT: vpackuswb %ymm5, %ymm6, %ymm5
-; AVX512F-NEXT: vpsrlw $7, %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm3, %ymm4, %ymm3
-; AVX512F-NEXT: vpaddb %ymm3, %ymm5, %ymm3
-; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
+; AVX512F-NEXT: vpsrlw $7, %ymm1, %ymm1
+; AVX512F-NEXT: vpand %ymm4, %ymm1, %ymm1
+; AVX512F-NEXT: vpaddb %ymm1, %ymm5, %ymm1
+; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm4 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm4, %ymm4
-; AVX512F-NEXT: vpand %ymm1, %ymm4, %ymm4
-; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm3 = ymm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
-; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm3, %ymm3
-; AVX512F-NEXT: vpand %ymm1, %ymm3, %ymm1
+; AVX512F-NEXT: vpand %ymm2, %ymm4, %ymm4
+; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
+; AVX512F-NEXT: vpmullw {{.*}}(%rip), %ymm1, %ymm1
+; AVX512F-NEXT: vpand %ymm2, %ymm1, %ymm1
; AVX512F-NEXT: vpackuswb %ymm4, %ymm1, %ymm1
; AVX512F-NEXT: vpsubb %ymm1, %ymm0, %ymm0
-; AVX512F-NEXT: vinserti64x4 $1, %ymm2, %zmm0, %zmm0
+; AVX512F-NEXT: vinserti64x4 $1, %ymm3, %zmm0, %zmm0
; AVX512F-NEXT: retq
;
; AVX512BW-LABEL: test_remconstant_64i8:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpsraw $8, %zmm1, %zmm1
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
-; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1
-; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm3, %zmm3
-; AVX512BW-NEXT: vpandq %zmm2, %zmm3, %zmm3
-; AVX512BW-NEXT: vpackuswb %zmm1, %zmm3, %zmm1
-; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
-; AVX512BW-NEXT: vpsraw $8, %zmm3, %zmm3
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm3, %zmm3
-; AVX512BW-NEXT: vpsrlw $8, %zmm3, %zmm3
-; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm4 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
-; AVX512BW-NEXT: vpsraw $8, %zmm4, %zmm4
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm4, %zmm4
-; AVX512BW-NEXT: vpsrlw $8, %zmm4, %zmm4
-; AVX512BW-NEXT: vpackuswb %zmm3, %zmm4, %zmm3
-; AVX512BW-NEXT: vpaddb %zmm1, %zmm3, %zmm1
-; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm3 = zmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpsrlw $8, %zmm1, %zmm1
+; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm2 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
+; AVX512BW-NEXT: vpsraw $8, %zmm2, %zmm2
+; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm2, %zmm2
+; AVX512BW-NEXT: vpsrlw $8, %zmm2, %zmm2
+; AVX512BW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1
+; AVX512BW-NEXT: vpandq {{.*}}(%rip), %zmm0, %zmm2
+; AVX512BW-NEXT: vpaddb %zmm2, %zmm1, %zmm1
+; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpsraw $8, %zmm2, %zmm2
+; AVX512BW-NEXT: vpsllvw {{.*}}(%rip), %zmm2, %zmm2
+; AVX512BW-NEXT: vpsrlw $8, %zmm2, %zmm2
+; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT: vpsraw $8, %zmm3, %zmm3
; AVX512BW-NEXT: vpsllvw {{.*}}(%rip), %zmm3, %zmm3
; AVX512BW-NEXT: vpsrlw $8, %zmm3, %zmm3
-; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm4 = zmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
-; AVX512BW-NEXT: vpsraw $8, %zmm4, %zmm4
-; AVX512BW-NEXT: vpsllvw {{.*}}(%rip), %zmm4, %zmm4
-; AVX512BW-NEXT: vpsrlw $8, %zmm4, %zmm4
-; AVX512BW-NEXT: vpackuswb %zmm3, %zmm4, %zmm3
+; AVX512BW-NEXT: vpackuswb %zmm2, %zmm3, %zmm2
; AVX512BW-NEXT: vpsrlw $7, %zmm1, %zmm1
; AVX512BW-NEXT: vpandq {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vpaddb %zmm1, %zmm3, %zmm1
-; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm3 = zmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
-; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm3, %zmm3
-; AVX512BW-NEXT: vpandq %zmm2, %zmm3, %zmm3
+; AVX512BW-NEXT: vpaddb %zmm1, %zmm2, %zmm1
+; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
+; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm2, %zmm2
+; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
+; AVX512BW-NEXT: vpandq %zmm3, %zmm2, %zmm2
; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm1 = zmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT: vpmullw {{.*}}(%rip), %zmm1, %zmm1
-; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1
-; AVX512BW-NEXT: vpackuswb %zmm3, %zmm1, %zmm1
+; AVX512BW-NEXT: vpandq %zmm3, %zmm1, %zmm1
+; AVX512BW-NEXT: vpackuswb %zmm2, %zmm1, %zmm1
; AVX512BW-NEXT: vpsubb %zmm1, %zmm0, %zmm0
; AVX512BW-NEXT: retq
%res = srem <64 x i8> %a, <i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 20, i8 21, i8 22, i8 23, i8 24, i8 25, i8 26, i8 27, i8 28, i8 29, i8 30, i8 31, i8 32, i8 33, i8 34, i8 35, i8 36, i8 37, i8 38, i8 38, i8 37, i8 36, i8 35, i8 34, i8 33, i8 32, i8 31, i8 30, i8 29, i8 28, i8 27, i8 26, i8 25, i8 24, i8 23, i8 22, i8 21, i8 20, i8 19, i8 18, i8 17, i8 16, i8 15, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9, i8 8, i8 7>
diff --git a/llvm/test/CodeGen/X86/vector-mul.ll b/llvm/test/CodeGen/X86/vector-mul.ll
index 805ff9f69ed5..e41f24125f36 100644
--- a/llvm/test/CodeGen/X86/vector-mul.ll
+++ b/llvm/test/CodeGen/X86/vector-mul.ll
@@ -1520,50 +1520,21 @@ define <2 x i64> @mul_v2i64_neg_17_65(<2 x i64> %a0) nounwind {
define <2 x i64> @mul_v2i64_0_1(<2 x i64> %a0) nounwind {
; X86-LABEL: mul_v2i64_0_1:
; X86: # %bb.0:
-; X86-NEXT: movdqa {{.*#+}} xmm1 = [0,0,1,0]
-; X86-NEXT: movdqa %xmm0, %xmm2
-; X86-NEXT: pmuludq %xmm1, %xmm2
-; X86-NEXT: psrlq $32, %xmm0
-; X86-NEXT: pmuludq %xmm1, %xmm0
-; X86-NEXT: psllq $32, %xmm0
-; X86-NEXT: paddq %xmm2, %xmm0
+; X86-NEXT: xorps %xmm1, %xmm1
+; X86-NEXT: blendps {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
; X86-NEXT: retl
;
; X64-LABEL: mul_v2i64_0_1:
; X64: # %bb.0:
-; X64-NEXT: movdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0]
-; X64-NEXT: movdqa %xmm0, %xmm2
-; X64-NEXT: pmuludq %xmm1, %xmm2
-; X64-NEXT: psrlq $32, %xmm0
-; X64-NEXT: pmuludq %xmm1, %xmm0
-; X64-NEXT: psllq $32, %xmm0
-; X64-NEXT: paddq %xmm2, %xmm0
+; X64-NEXT: xorps %xmm1, %xmm1
+; X64-NEXT: blendps {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
; X64-NEXT: retq
;
-; X64-XOP-LABEL: mul_v2i64_0_1:
-; X64-XOP: # %bb.0:
-; X64-XOP-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0]
-; X64-XOP-NEXT: vpmuludq %xmm1, %xmm0, %xmm2
-; X64-XOP-NEXT: vpsrlq $32, %xmm0, %xmm0
-; X64-XOP-NEXT: vpmuludq %xmm1, %xmm0, %xmm0
-; X64-XOP-NEXT: vpsllq $32, %xmm0, %xmm0
-; X64-XOP-NEXT: vpaddq %xmm0, %xmm2, %xmm0
-; X64-XOP-NEXT: retq
-;
-; X64-AVX2-LABEL: mul_v2i64_0_1:
-; X64-AVX2: # %bb.0:
-; X64-AVX2-NEXT: vmovdqa {{.*#+}} xmm1 = [0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0]
-; X64-AVX2-NEXT: vpmuludq %xmm1, %xmm0, %xmm2
-; X64-AVX2-NEXT: vpsrlq $32, %xmm0, %xmm0
-; X64-AVX2-NEXT: vpmuludq %xmm1, %xmm0, %xmm0
-; X64-AVX2-NEXT: vpsllq $32, %xmm0, %xmm0
-; X64-AVX2-NEXT: vpaddq %xmm0, %xmm2, %xmm0
-; X64-AVX2-NEXT: retq
-;
-; X64-AVX512DQ-LABEL: mul_v2i64_0_1:
-; X64-AVX512DQ: # %bb.0:
-; X64-AVX512DQ-NEXT: vpmullq {{.*}}(%rip), %xmm0, %xmm0
-; X64-AVX512DQ-NEXT: retq
+; X64-AVX-LABEL: mul_v2i64_0_1:
+; X64-AVX: # %bb.0:
+; X64-AVX-NEXT: vxorps %xmm1, %xmm1, %xmm1
+; X64-AVX-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
+; X64-AVX-NEXT: retq
%1 = mul <2 x i64> %a0, <i64 0, i64 1>
ret <2 x i64> %1
}
diff --git a/llvm/test/CodeGen/X86/vector-trunc-math.ll b/llvm/test/CodeGen/X86/vector-trunc-math.ll
index f245fc08b9c6..71ffb58a2691 100644
--- a/llvm/test/CodeGen/X86/vector-trunc-math.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-math.ll
@@ -2313,9 +2313,11 @@ define <8 x i16> @trunc_mul_v8i32_v8i16_zext_8i8(<16 x i8> %a0, <8 x i32> %a1) {
define <4 x i32> @trunc_mul_const_v4i64_v4i32(<4 x i64> %a0) nounwind {
; SSE-LABEL: trunc_mul_const_v4i64_v4i32:
; SSE: # %bb.0:
+; SSE-NEXT: xorps %xmm2, %xmm2
+; SSE-NEXT: unpckhpd {{.*#+}} xmm2 = xmm2[1],xmm0[1]
; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm1
-; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm0
-; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
+; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm1[0,2]
+; SSE-NEXT: movaps %xmm2, %xmm0
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc_mul_const_v4i64_v4i32:
@@ -2464,7 +2466,6 @@ define <8 x i16> @trunc_mul_const_v8i32_v8i16(<8 x i32> %a0) nounwind {
define <16 x i8> @trunc_mul_const_v16i64_v16i8(<16 x i64> %a0) nounwind {
; SSE-LABEL: trunc_mul_const_v16i64_v16i8:
; SSE: # %bb.0:
-; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm0
; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm1
; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm2
; SSE-NEXT: pmuludq {{.*}}(%rip), %xmm3
@@ -2484,7 +2485,7 @@ define <16 x i8> @trunc_mul_const_v16i64_v16i8(<16 x i64> %a0) nounwind {
; SSE-NEXT: pand %xmm8, %xmm2
; SSE-NEXT: packuswb %xmm3, %xmm2
; SSE-NEXT: pand %xmm8, %xmm1
-; SSE-NEXT: pand %xmm8, %xmm0
+; SSE-NEXT: pand {{.*}}(%rip), %xmm0
; SSE-NEXT: packuswb %xmm1, %xmm0
; SSE-NEXT: packuswb %xmm2, %xmm0
; SSE-NEXT: packuswb %xmm4, %xmm0
More information about the llvm-commits
mailing list