[x86] avoid cmov in movmsk reduction

This is probably the least important of our movmsk problems, but I'm starting
at the bottom to reduce distractions.

We were creating a select_cc which bypasses the select and bitmask codegen
optimizations that we have now. If we produce a compare+negate instead, we
allow things like neg/sbb carry bit hacks, and in all cases we avoid a cmov.
There's no partial register update danger in these sequences because we always
produce the zero-register xor ahead of the 'set' if needed.

There seems to be a missing fold for sext of a bool bit here:

negl %ecx
movslq %ecx, %rax

...but that's an independent transform.

Differential Revision: https://reviews.llvm.org/D59818

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@357172 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index ba168d35f0a9544864e8301167c8c5a3b95ccf50..50210ad78a2f814c2479750f04e6dcb42510ae3a 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -34302,37 +34302,39 @@ static SDValue combineHorizontalPredicateResult(SDNode *Extract,
     return SDValue();
 
   // For 32/64 bit comparisons use MOVMSKPS/MOVMSKPD, else PMOVMSKB.
-  MVT MaskVT;
+  MVT MaskSrcVT;
   if (64 == BitWidth || 32 == BitWidth)
-    MaskVT = MVT::getVectorVT(MVT::getFloatingPointVT(BitWidth),
-                              MatchSizeInBits / BitWidth);
+    MaskSrcVT = MVT::getVectorVT(MVT::getFloatingPointVT(BitWidth),
+                                 MatchSizeInBits / BitWidth);
   else
-    MaskVT = MVT::getVectorVT(MVT::i8, MatchSizeInBits / 8);
+    MaskSrcVT = MVT::getVectorVT(MVT::i8, MatchSizeInBits / 8);
 
-  APInt CompareBits;
+  SDLoc DL(Extract);
+  SDValue CmpC;
   ISD::CondCode CondCode;
   if (BinOp == ISD::OR) {
     // any_of -> MOVMSK != 0
-    CompareBits = APInt::getNullValue(32);
+    CmpC = DAG.getConstant(0, DL, MVT::i32);
     CondCode = ISD::CondCode::SETNE;
   } else {
     // all_of -> MOVMSK == ((1 << NumElts) - 1)
-    CompareBits = APInt::getLowBitsSet(32, MaskVT.getVectorNumElements());
+    uint64_t NumElts = MaskSrcVT.getVectorNumElements();
+    assert(NumElts <= 32 && "Not expecting more than 32 elements");
+    CmpC = DAG.getConstant((1ULL << NumElts) - 1, DL, MVT::i32);
     CondCode = ISD::CondCode::SETEQ;
   }
 
-  // Perform the select as i32/i64 and then truncate to avoid partial register
-  // stalls.
-  unsigned ResWidth = std::max(BitWidth, 32u);
-  EVT ResVT = EVT::getIntegerVT(*DAG.getContext(), ResWidth);
-  SDLoc DL(Extract);
-  SDValue Zero = DAG.getConstant(0, DL, ResVT);
-  SDValue Ones = DAG.getAllOnesConstant(DL, ResVT);
-  SDValue Res = DAG.getBitcast(MaskVT, Match);
-  Res = DAG.getNode(X86ISD::MOVMSK, DL, MVT::i32, Res);
-  Res = DAG.getSelectCC(DL, Res, DAG.getConstant(CompareBits, DL, MVT::i32),
-                        Ones, Zero, CondCode);
-  return DAG.getSExtOrTrunc(Res, DL, ExtractVT);
+  // The setcc produces an i8 of 0/1, so extend that to the result width and
+  // negate to get the final 0/-1 mask value.
+  SDValue BitcastLogicOp = DAG.getBitcast(MaskSrcVT, Match);
+  SDValue Movmsk = DAG.getNode(X86ISD::MOVMSK, DL, MVT::i32, BitcastLogicOp);
+  const TargetLowering &TLI = DAG.getTargetLoweringInfo();
+  EVT SetccVT = TLI.getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(),
+                                       MVT::i32);
+  SDValue Setcc = DAG.getSetCC(DL, SetccVT, Movmsk, CmpC, CondCode);
+  SDValue Zext = DAG.getZExtOrTrunc(Setcc, DL, ExtractVT);
+  SDValue Zero = DAG.getConstant(0, DL, ExtractVT);
+  return DAG.getNode(ISD::SUB, DL, ExtractVT, Zero, Zext);
 }
 
 static SDValue combineBasicSADPattern(SDNode *Extract, SelectionDAG &DAG,

diff --git a/test/CodeGen/X86/vector-compare-all_of.ll b/test/CodeGen/X86/vector-compare-all_of.ll
index 8b947b7f475a5c19fe533ec000ed063bd746fe64..b067e40b2e95084448832b1c2ed7e82b2cd04ff4 100644 (file)
--- a/test/CodeGen/X86/vector-compare-all_of.ll
+++ b/test/CodeGen/X86/vector-compare-all_of.ll
@@ -50,11 +50,11 @@ define i64 @test_v4f64_sext(<4 x double> %a0, <4 x double> %a1) {
 ; AVX-LABEL: test_v4f64_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltpd %ymm0, %ymm1, %ymm0
-; AVX-NEXT:    vmovmskpd %ymm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $15, %eax
-; AVX-NEXT:    movq $-1, %rax
-; AVX-NEXT:    cmovneq %rcx, %rax
+; AVX-NEXT:    vmovmskpd %ymm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $15, %ecx
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negq %rax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
@@ -87,8 +87,9 @@ define i64 @test_v4f64_legal_sext(<4 x double> %a0, <4 x double> %a1) {
 ; SSE-NEXT:    movmskps %xmm2, %eax
 ; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movq $-1, %rax
-; SSE-NEXT:    cmovneq %rcx, %rax
+; SSE-NEXT:    sete %cl
+; SSE-NEXT:    negl %ecx
+; SSE-NEXT:    movslq %ecx, %rax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4f64_legal_sext:
@@ -99,8 +100,9 @@ define i64 @test_v4f64_legal_sext(<4 x double> %a0, <4 x double> %a1) {
 ; AVX-NEXT:    vmovmskps %xmm0, %eax
 ; AVX-NEXT:    xorl %ecx, %ecx
 ; AVX-NEXT:    cmpl $15, %eax
-; AVX-NEXT:    movq $-1, %rax
-; AVX-NEXT:    cmovneq %rcx, %rax
+; AVX-NEXT:    sete %cl
+; AVX-NEXT:    negl %ecx
+; AVX-NEXT:    movslq %ecx, %rax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
@@ -132,21 +134,21 @@ define i32 @test_v4f32_sext(<4 x float> %a0, <4 x float> %a1) {
 ; SSE-LABEL: test_v4f32_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    cmpltps %xmm0, %xmm1
-; SSE-NEXT:    movmskps %xmm1, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    movmskps %xmm1, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $15, %ecx
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4f32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltps %xmm0, %xmm1, %xmm0
-; AVX-NEXT:    vmovmskps %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $15, %eax
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
+; AVX-NEXT:    vmovmskps %xmm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $15, %ecx
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: test_v4f32_sext:
@@ -174,21 +176,21 @@ define i32 @test_v8f32_sext(<8 x float> %a0, <8 x float> %a1) {
 ; SSE-NEXT:    cmpltps %xmm1, %xmm3
 ; SSE-NEXT:    cmpltps %xmm0, %xmm2
 ; SSE-NEXT:    andps %xmm3, %xmm2
-; SSE-NEXT:    movmskps %xmm2, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    movmskps %xmm2, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $15, %ecx
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v8f32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltps %ymm0, %ymm1, %ymm0
-; AVX-NEXT:    vmovmskps %ymm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $255, %eax
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
+; AVX-NEXT:    vmovmskps %ymm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $255, %ecx
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
@@ -222,11 +224,11 @@ define i32 @test_v8f32_legal_sext(<8 x float> %a0, <8 x float> %a1) {
 ; SSE-NEXT:    cmpltps %xmm1, %xmm3
 ; SSE-NEXT:    cmpltps %xmm0, %xmm2
 ; SSE-NEXT:    packssdw %xmm3, %xmm2
-; SSE-NEXT:    pmovmskb %xmm2, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm2, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v8f32_legal_sext:
@@ -234,11 +236,11 @@ define i32 @test_v8f32_legal_sext(<8 x float> %a0, <8 x float> %a1) {
 ; AVX-NEXT:    vcmpltps %ymm0, %ymm1, %ymm0
 ; AVX-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; AVX-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
+; AVX-NEXT:    vpmovmskb %xmm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
@@ -319,22 +321,22 @@ define i64 @test_v4i64_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX1-NEXT:    vpcmpgtq %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtq %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT:    vmovmskpd %ymm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
-; AVX1-NEXT:    cmpl $15, %eax
-; AVX1-NEXT:    movq $-1, %rax
-; AVX1-NEXT:    cmovneq %rcx, %rax
+; AVX1-NEXT:    vmovmskpd %ymm0, %ecx
+; AVX1-NEXT:    xorl %eax, %eax
+; AVX1-NEXT:    cmpl $15, %ecx
+; AVX1-NEXT:    sete %al
+; AVX1-NEXT:    negq %rax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: test_v4i64_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtq %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vmovmskpd %ymm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
-; AVX2-NEXT:    cmpl $15, %eax
-; AVX2-NEXT:    movq $-1, %rax
-; AVX2-NEXT:    cmovneq %rcx, %rax
+; AVX2-NEXT:    vmovmskpd %ymm0, %ecx
+; AVX2-NEXT:    xorl %eax, %eax
+; AVX2-NEXT:    cmpl $15, %ecx
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negq %rax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -367,8 +369,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; SSE-NEXT:    movmskps %xmm0, %eax
 ; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movq $-1, %rax
-; SSE-NEXT:    cmovneq %rcx, %rax
+; SSE-NEXT:    sete %cl
+; SSE-NEXT:    negl %ecx
+; SSE-NEXT:    movslq %ecx, %rax
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v4i64_legal_sext:
@@ -381,8 +384,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX1-NEXT:    vmovmskps %xmm0, %eax
 ; AVX1-NEXT:    xorl %ecx, %ecx
 ; AVX1-NEXT:    cmpl $15, %eax
-; AVX1-NEXT:    movq $-1, %rax
-; AVX1-NEXT:    cmovneq %rcx, %rax
+; AVX1-NEXT:    sete %cl
+; AVX1-NEXT:    negl %ecx
+; AVX1-NEXT:    movslq %ecx, %rax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -394,8 +398,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX2-NEXT:    vmovmskps %xmm0, %eax
 ; AVX2-NEXT:    xorl %ecx, %ecx
 ; AVX2-NEXT:    cmpl $15, %eax
-; AVX2-NEXT:    movq $-1, %rax
-; AVX2-NEXT:    cmovneq %rcx, %rax
+; AVX2-NEXT:    sete %cl
+; AVX2-NEXT:    negl %ecx
+; AVX2-NEXT:    movslq %ecx, %rax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -427,21 +432,21 @@ define i32 @test_v4i32_sext(<4 x i32> %a0, <4 x i32> %a1) {
 ; SSE-LABEL: test_v4i32_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtd %xmm1, %xmm0
-; SSE-NEXT:    movmskps %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    movmskps %xmm0, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $15, %ecx
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4i32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vmovmskps %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $15, %eax
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
+; AVX-NEXT:    vmovmskps %xmm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $15, %ecx
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: test_v4i32_sext:
@@ -469,11 +474,11 @@ define i32 @test_v8i32_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; SSE-NEXT:    pcmpgtd %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm0
 ; SSE-NEXT:    pand %xmm1, %xmm0
-; SSE-NEXT:    movmskps %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $15, %eax
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    movmskps %xmm0, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $15, %ecx
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v8i32_sext:
@@ -483,22 +488,22 @@ define i32 @test_v8i32_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX1-NEXT:    vpcmpgtd %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT:    vmovmskps %ymm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
-; AVX1-NEXT:    cmpl $255, %eax
-; AVX1-NEXT:    movl $-1, %eax
-; AVX1-NEXT:    cmovnel %ecx, %eax
+; AVX1-NEXT:    vmovmskps %ymm0, %ecx
+; AVX1-NEXT:    xorl %eax, %eax
+; AVX1-NEXT:    cmpl $255, %ecx
+; AVX1-NEXT:    sete %al
+; AVX1-NEXT:    negl %eax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: test_v8i32_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtd %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vmovmskps %ymm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
-; AVX2-NEXT:    cmpl $255, %eax
-; AVX2-NEXT:    movl $-1, %eax
-; AVX2-NEXT:    cmovnel %ecx, %eax
+; AVX2-NEXT:    vmovmskps %ymm0, %ecx
+; AVX2-NEXT:    xorl %eax, %eax
+; AVX2-NEXT:    cmpl $255, %ecx
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -532,11 +537,11 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; SSE-NEXT:    pcmpgtd %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm0
 ; SSE-NEXT:    packssdw %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v8i32_legal_sext:
@@ -546,11 +551,11 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX1-NEXT:    vpcmpgtd %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackssdw %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
-; AVX1-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX1-NEXT:    movl $-1, %eax
-; AVX1-NEXT:    cmovnel %ecx, %eax
+; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
+; AVX1-NEXT:    xorl %eax, %eax
+; AVX1-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; AVX1-NEXT:    sete %al
+; AVX1-NEXT:    negl %eax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -559,11 +564,11 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX2-NEXT:    vpcmpgtd %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpmovmskb %xmm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
-; AVX2-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX2-NEXT:    movl $-1, %eax
-; AVX2-NEXT:    cmovnel %ecx, %eax
+; AVX2-NEXT:    vpmovmskb %xmm0, %ecx
+; AVX2-NEXT:    xorl %eax, %eax
+; AVX2-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -598,22 +603,22 @@ define i16 @test_v8i16_sext(<8 x i16> %a0, <8 x i16> %a1) {
 ; SSE-LABEL: test_v8i16_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtw %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v8i16_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtw %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
+; AVX-NEXT:    vpmovmskb %xmm0, %ecx
+; AVX-NEXT:    xorl %eax, %eax
+; AVX-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
 ;
@@ -647,11 +652,11 @@ define i16 @test_v16i16_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; SSE-NEXT:    pcmpgtw %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtw %xmm2, %xmm0
 ; SSE-NEXT:    pand %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %ecx
+; SSE-NEXT:    xorl %eax, %eax
+; SSE-NEXT:    cmpl $65535, %ecx # imm = 0xFFFF
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
 ;
@@ -679,7 +684,8 @@ define i16 @test_v16i16_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
 ; AVX2-NEXT:    xorl %eax, %eax
 ; AVX2-NEXT:    cmpl $-1, %ecx
-; AVX2-NEXT:    cmovel %ecx, %eax
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -720,10 +726,10 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; SSE-NEXT:    pcmpgtw %xmm2, %xmm0
 ; SSE-NEXT:    packsswb %xmm1, %xmm0
 ; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
+; SSE-NEXT:    movsbl %al, %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
 ;
@@ -735,10 +741,10 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX1-NEXT:    vpcmpgtw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpacksswb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpmovmskb %xmm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
 ; AVX1-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX1-NEXT:    movl $-1, %eax
-; AVX1-NEXT:    cmovnel %ecx, %eax
+; AVX1-NEXT:    sete %al
+; AVX1-NEXT:    negb %al
+; AVX1-NEXT:    movsbl %al, %eax
 ; AVX1-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
@@ -749,10 +755,10 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpacksswb %xmm1, %xmm0, %xmm0
 ; AVX2-NEXT:    vpmovmskb %xmm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
 ; AVX2-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX2-NEXT:    movl $-1, %eax
-; AVX2-NEXT:    cmovnel %ecx, %eax
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negb %al
+; AVX2-NEXT:    movsbl %al, %eax
 ; AVX2-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -794,22 +800,18 @@ define i8 @test_v16i8_sext(<16 x i8> %a0, <16 x i8> %a1) {
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtb %xmm1, %xmm0
 ; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v16i8_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtb %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpmovmskb %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
 ; AVX-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
-; AVX-NEXT:    # kill: def $al killed $al killed $eax
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negb %al
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: test_v16i8_sext:
@@ -847,11 +849,9 @@ define i8 @test_v32i8_sext(<32 x i8> %a0, <32 x i8> %a1) {
 ; SSE-NEXT:    pcmpgtb %xmm2, %xmm0
 ; SSE-NEXT:    pand %xmm1, %xmm0
 ; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v32i8_sext:
@@ -877,11 +877,10 @@ define i8 @test_v32i8_sext(<32 x i8> %a0, <32 x i8> %a1) {
 ; AVX2-LABEL: test_v32i8_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtb %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl $-1, %ecx
-; AVX2-NEXT:    cmovel %ecx, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; AVX2-NEXT:    cmpl $-1, %eax
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negb %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -1272,22 +1271,18 @@ define i1 @bool_reduction_v16i8(<16 x i8> %x, <16 x i8> %y) {
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtb %xmm1, %xmm0
 ; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: bool_reduction_v16i8:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtb %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpmovmskb %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
 ; AVX-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX-NEXT:    movl $-1, %eax
-; AVX-NEXT:    cmovnel %ecx, %eax
-; AVX-NEXT:    # kill: def $al killed $al killed $eax
+; AVX-NEXT:    sete %al
+; AVX-NEXT:    negb %al
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: bool_reduction_v16i8:
@@ -1480,11 +1475,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; SSE-NEXT:    pcmpeqw %xmm2, %xmm0
 ; SSE-NEXT:    packsswb %xmm1, %xmm0
 ; SSE-NEXT:    pmovmskb %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: bool_reduction_v16i16:
@@ -1495,11 +1488,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; AVX1-NEXT:    vpcmpeqw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpacksswb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpmovmskb %xmm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
 ; AVX1-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX1-NEXT:    movl $-1, %eax
-; AVX1-NEXT:    cmovnel %ecx, %eax
-; AVX1-NEXT:    # kill: def $al killed $al killed $eax
+; AVX1-NEXT:    sete %al
+; AVX1-NEXT:    negb %al
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -1509,11 +1500,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpacksswb %xmm1, %xmm0, %xmm0
 ; AVX2-NEXT:    vpmovmskb %xmm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
 ; AVX2-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; AVX2-NEXT:    movl $-1, %eax
-; AVX2-NEXT:    cmovnel %ecx, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negb %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -1552,11 +1541,9 @@ define i1 @bool_reduction_v32i8(<32 x i8> %x, <32 x i8> %y) {
 ; SSE-NEXT:    pcmpeqb %xmm3, %xmm1
 ; SSE-NEXT:    pand %xmm0, %xmm1
 ; SSE-NEXT:    pmovmskb %xmm1, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
 ; SSE-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
-; SSE-NEXT:    movl $-1, %eax
-; SSE-NEXT:    cmovnel %ecx, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    sete %al
+; SSE-NEXT:    negb %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: bool_reduction_v32i8:
@@ -1582,11 +1569,10 @@ define i1 @bool_reduction_v32i8(<32 x i8> %x, <32 x i8> %y) {
 ; AVX2-LABEL: bool_reduction_v32i8:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpeqb %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl $-1, %ecx
-; AVX2-NEXT:    cmovel %ecx, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; AVX2-NEXT:    cmpl $-1, %eax
+; AVX2-NEXT:    sete %al
+; AVX2-NEXT:    negb %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;

diff --git a/test/CodeGen/X86/vector-compare-any_of.ll b/test/CodeGen/X86/vector-compare-any_of.ll
index 65c3052b79eb5037828a5d9d7f35a1abdfdd02a6..4bda687699de9d7fc2291dc0de65e91a5f914bbe 100644 (file)
--- a/test/CodeGen/X86/vector-compare-any_of.ll
+++ b/test/CodeGen/X86/vector-compare-any_of.ll
@@ -50,9 +50,8 @@ define i64 @test_v4f64_sext(<4 x double> %a0, <4 x double> %a1) {
 ; AVX-LABEL: test_v4f64_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltpd %ymm0, %ymm1, %ymm0
-; AVX-NEXT:    vmovmskpd %ymm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vmovmskpd %ymm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbq %rax, %rax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
@@ -84,10 +83,9 @@ define i64 @test_v4f64_legal_sext(<4 x double> %a0, <4 x double> %a1) {
 ; SSE-NEXT:    cmpltpd %xmm0, %xmm2
 ; SSE-NEXT:    packssdw %xmm3, %xmm2
 ; SSE-NEXT:    movmskps %xmm2, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl %eax, %ecx
-; SSE-NEXT:    sbbl %ecx, %ecx
-; SSE-NEXT:    movslq %ecx, %rax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbl %eax, %eax
+; SSE-NEXT:    cltq
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4f64_legal_sext:
@@ -96,10 +94,9 @@ define i64 @test_v4f64_legal_sext(<4 x double> %a0, <4 x double> %a1) {
 ; AVX-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; AVX-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vmovmskps %xmm0, %eax
-; AVX-NEXT:    xorl %ecx, %ecx
-; AVX-NEXT:    cmpl %eax, %ecx
-; AVX-NEXT:    sbbl %ecx, %ecx
-; AVX-NEXT:    movslq %ecx, %rax
+; AVX-NEXT:    negl %eax
+; AVX-NEXT:    sbbl %eax, %eax
+; AVX-NEXT:    cltq
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
@@ -131,18 +128,16 @@ define i32 @test_v4f32_sext(<4 x float> %a0, <4 x float> %a1) {
 ; SSE-LABEL: test_v4f32_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    cmpltps %xmm0, %xmm1
-; SSE-NEXT:    movmskps %xmm1, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    movmskps %xmm1, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4f32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltps %xmm0, %xmm1, %xmm0
-; AVX-NEXT:    vmovmskps %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vmovmskps %xmm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbl %eax, %eax
 ; AVX-NEXT:    retq
 ;
@@ -171,18 +166,16 @@ define i32 @test_v8f32_sext(<8 x float> %a0, <8 x float> %a1) {
 ; SSE-NEXT:    cmpltps %xmm1, %xmm3
 ; SSE-NEXT:    cmpltps %xmm0, %xmm2
 ; SSE-NEXT:    orps %xmm3, %xmm2
-; SSE-NEXT:    movmskps %xmm2, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    movmskps %xmm2, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v8f32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vcmpltps %ymm0, %ymm1, %ymm0
-; AVX-NEXT:    vmovmskps %ymm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vmovmskps %ymm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbl %eax, %eax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
@@ -217,9 +210,8 @@ define i32 @test_v8f32_legal_sext(<8 x float> %a0, <8 x float> %a1) {
 ; SSE-NEXT:    cmpltps %xmm1, %xmm3
 ; SSE-NEXT:    cmpltps %xmm0, %xmm2
 ; SSE-NEXT:    packssdw %xmm3, %xmm2
-; SSE-NEXT:    pmovmskb %xmm2, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm2, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
@@ -228,9 +220,8 @@ define i32 @test_v8f32_legal_sext(<8 x float> %a0, <8 x float> %a1) {
 ; AVX-NEXT:    vcmpltps %ymm0, %ymm1, %ymm0
 ; AVX-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; AVX-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vpmovmskb %xmm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbl %eax, %eax
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
@@ -312,9 +303,8 @@ define i64 @test_v4i64_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX1-NEXT:    vpcmpgtq %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtq %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT:    vmovmskpd %ymm0, %ecx
-; AVX1-NEXT:    xorl %eax, %eax
-; AVX1-NEXT:    cmpl %ecx, %eax
+; AVX1-NEXT:    vmovmskpd %ymm0, %eax
+; AVX1-NEXT:    negl %eax
 ; AVX1-NEXT:    sbbq %rax, %rax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
@@ -322,9 +312,8 @@ define i64 @test_v4i64_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX2-LABEL: test_v4i64_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtq %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vmovmskpd %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
+; AVX2-NEXT:    vmovmskpd %ymm0, %eax
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    sbbq %rax, %rax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -356,10 +345,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; SSE-NEXT:    pcmpgtq %xmm2, %xmm0
 ; SSE-NEXT:    packssdw %xmm1, %xmm0
 ; SSE-NEXT:    movmskps %xmm0, %eax
-; SSE-NEXT:    xorl %ecx, %ecx
-; SSE-NEXT:    cmpl %eax, %ecx
-; SSE-NEXT:    sbbl %ecx, %ecx
-; SSE-NEXT:    movslq %ecx, %rax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbl %eax, %eax
+; SSE-NEXT:    cltq
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v4i64_legal_sext:
@@ -370,10 +358,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX1-NEXT:    vpcmpgtq %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackssdw %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vmovmskps %xmm0, %eax
-; AVX1-NEXT:    xorl %ecx, %ecx
-; AVX1-NEXT:    cmpl %eax, %ecx
-; AVX1-NEXT:    sbbl %ecx, %ecx
-; AVX1-NEXT:    movslq %ecx, %rax
+; AVX1-NEXT:    negl %eax
+; AVX1-NEXT:    sbbl %eax, %eax
+; AVX1-NEXT:    cltq
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -383,10 +370,9 @@ define i64 @test_v4i64_legal_sext(<4 x i64> %a0, <4 x i64> %a1) {
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
 ; AVX2-NEXT:    vmovmskps %xmm0, %eax
-; AVX2-NEXT:    xorl %ecx, %ecx
-; AVX2-NEXT:    cmpl %eax, %ecx
-; AVX2-NEXT:    sbbl %ecx, %ecx
-; AVX2-NEXT:    movslq %ecx, %rax
+; AVX2-NEXT:    negl %eax
+; AVX2-NEXT:    sbbl %eax, %eax
+; AVX2-NEXT:    cltq
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -418,18 +404,16 @@ define i32 @test_v4i32_sext(<4 x i32> %a0, <4 x i32> %a1) {
 ; SSE-LABEL: test_v4i32_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtd %xmm1, %xmm0
-; SSE-NEXT:    movmskps %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    movmskps %xmm0, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v4i32_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vmovmskps %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vmovmskps %xmm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbl %eax, %eax
 ; AVX-NEXT:    retq
 ;
@@ -458,9 +442,8 @@ define i32 @test_v8i32_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; SSE-NEXT:    pcmpgtd %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm0
 ; SSE-NEXT:    por %xmm1, %xmm0
-; SSE-NEXT:    movmskps %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    movmskps %xmm0, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
@@ -471,9 +454,8 @@ define i32 @test_v8i32_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX1-NEXT:    vpcmpgtd %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
-; AVX1-NEXT:    vmovmskps %ymm0, %ecx
-; AVX1-NEXT:    xorl %eax, %eax
-; AVX1-NEXT:    cmpl %ecx, %eax
+; AVX1-NEXT:    vmovmskps %ymm0, %eax
+; AVX1-NEXT:    negl %eax
 ; AVX1-NEXT:    sbbl %eax, %eax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
@@ -481,9 +463,8 @@ define i32 @test_v8i32_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX2-LABEL: test_v8i32_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtd %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vmovmskps %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
+; AVX2-NEXT:    vmovmskps %ymm0, %eax
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    sbbl %eax, %eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -518,9 +499,8 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; SSE-NEXT:    pcmpgtd %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm0
 ; SSE-NEXT:    packssdw %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    retq
 ;
@@ -531,9 +511,8 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX1-NEXT:    vpcmpgtd %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtd %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackssdw %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX1-NEXT:    xorl %eax, %eax
-; AVX1-NEXT:    cmpl %ecx, %eax
+; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    negl %eax
 ; AVX1-NEXT:    sbbl %eax, %eax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
@@ -543,9 +522,8 @@ define i32 @test_v8i32_legal_sext(<8 x i32> %a0, <8 x i32> %a1) {
 ; AVX2-NEXT:    vpcmpgtd %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpackssdw %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
+; AVX2-NEXT:    vpmovmskb %xmm0, %eax
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    sbbl %eax, %eax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -581,9 +559,8 @@ define i16 @test_v8i16_sext(<8 x i16> %a0, <8 x i16> %a1) {
 ; SSE-LABEL: test_v8i16_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtw %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
@@ -591,9 +568,8 @@ define i16 @test_v8i16_sext(<8 x i16> %a0, <8 x i16> %a1) {
 ; AVX-LABEL: test_v8i16_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtw %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
+; AVX-NEXT:    vpmovmskb %xmm0, %eax
+; AVX-NEXT:    negl %eax
 ; AVX-NEXT:    sbbl %eax, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -628,9 +604,8 @@ define i16 @test_v16i16_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; SSE-NEXT:    pcmpgtw %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtw %xmm2, %xmm0
 ; SSE-NEXT:    por %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
 ; SSE-NEXT:    sbbl %eax, %eax
 ; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE-NEXT:    retq
@@ -656,9 +631,8 @@ define i16 @test_v16i16_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX2-LABEL: test_v16i16_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtw %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
+; AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; AVX2-NEXT:    negl %eax
 ; AVX2-NEXT:    sbbl %eax, %eax
 ; AVX2-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX2-NEXT:    vzeroupper
@@ -699,11 +673,9 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; SSE-NEXT:    pcmpgtw %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtw %xmm2, %xmm0
 ; SSE-NEXT:    packsswb %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $ax killed $ax killed $eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbw %ax, %ax
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v16i16_legal_sext:
@@ -713,11 +685,9 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX1-NEXT:    vpcmpgtw %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpgtw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpacksswb %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX1-NEXT:    xorl %eax, %eax
-; AVX1-NEXT:    cmpl %ecx, %eax
-; AVX1-NEXT:    sbbl %eax, %eax
-; AVX1-NEXT:    # kill: def $ax killed $ax killed $eax
+; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    negl %eax
+; AVX1-NEXT:    sbbw %ax, %ax
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -726,11 +696,9 @@ define i16 @test_v16i16_legal_sext(<16 x i16> %a0, <16 x i16> %a1) {
 ; AVX2-NEXT:    vpcmpgtw %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpacksswb %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
-; AVX2-NEXT:    sbbl %eax, %eax
-; AVX2-NEXT:    # kill: def $ax killed $ax killed $eax
+; AVX2-NEXT:    vpmovmskb %xmm0, %eax
+; AVX2-NEXT:    negl %eax
+; AVX2-NEXT:    sbbw %ax, %ax
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -770,21 +738,17 @@ define i8 @test_v16i8_sext(<16 x i8> %a0, <16 x i8> %a1) {
 ; SSE-LABEL: test_v16i8_sext:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtb %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbb %al, %al
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test_v16i8_sext:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtb %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
-; AVX-NEXT:    sbbl %eax, %eax
-; AVX-NEXT:    # kill: def $al killed $al killed $eax
+; AVX-NEXT:    vpmovmskb %xmm0, %eax
+; AVX-NEXT:    negl %eax
+; AVX-NEXT:    sbbb %al, %al
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: test_v16i8_sext:
@@ -821,11 +785,9 @@ define i8 @test_v32i8_sext(<32 x i8> %a0, <32 x i8> %a1) {
 ; SSE-NEXT:    pcmpgtb %xmm3, %xmm1
 ; SSE-NEXT:    pcmpgtb %xmm2, %xmm0
 ; SSE-NEXT:    por %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbb %al, %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: test_v32i8_sext:
@@ -851,11 +813,9 @@ define i8 @test_v32i8_sext(<32 x i8> %a0, <32 x i8> %a1) {
 ; AVX2-LABEL: test_v32i8_sext:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpgtb %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
-; AVX2-NEXT:    sbbl %eax, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; AVX2-NEXT:    negl %eax
+; AVX2-NEXT:    sbbb %al, %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -1258,21 +1218,17 @@ define i1 @bool_reduction_v16i8(<16 x i8> %x, <16 x i8> %y) {
 ; SSE-LABEL: bool_reduction_v16i8:
 ; SSE:       # %bb.0:
 ; SSE-NEXT:    pcmpgtb %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbb %al, %al
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: bool_reduction_v16i8:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpcmpgtb %xmm1, %xmm0, %xmm0
-; AVX-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX-NEXT:    xorl %eax, %eax
-; AVX-NEXT:    cmpl %ecx, %eax
-; AVX-NEXT:    sbbl %eax, %eax
-; AVX-NEXT:    # kill: def $al killed $al killed $eax
+; AVX-NEXT:    vpmovmskb %xmm0, %eax
+; AVX-NEXT:    negl %eax
+; AVX-NEXT:    sbbb %al, %al
 ; AVX-NEXT:    retq
 ;
 ; AVX512-LABEL: bool_reduction_v16i8:
@@ -1469,11 +1425,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; SSE-NEXT:    pcmpeqw %xmm3, %xmm1
 ; SSE-NEXT:    pcmpeqw %xmm2, %xmm0
 ; SSE-NEXT:    packsswb %xmm1, %xmm0
-; SSE-NEXT:    pmovmskb %xmm0, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    pmovmskb %xmm0, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbb %al, %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: bool_reduction_v16i16:
@@ -1483,11 +1437,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; AVX1-NEXT:    vpcmpeqw %xmm2, %xmm3, %xmm2
 ; AVX1-NEXT:    vpcmpeqw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpacksswb %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX1-NEXT:    xorl %eax, %eax
-; AVX1-NEXT:    cmpl %ecx, %eax
-; AVX1-NEXT:    sbbl %eax, %eax
-; AVX1-NEXT:    # kill: def $al killed $al killed $eax
+; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    negl %eax
+; AVX1-NEXT:    sbbb %al, %al
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
@@ -1496,11 +1448,9 @@ define i1 @bool_reduction_v16i16(<16 x i16> %x, <16 x i16> %y) {
 ; AVX2-NEXT:    vpcmpeqw %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
 ; AVX2-NEXT:    vpacksswb %xmm1, %xmm0, %xmm0
-; AVX2-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
-; AVX2-NEXT:    sbbl %eax, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    vpmovmskb %xmm0, %eax
+; AVX2-NEXT:    negl %eax
+; AVX2-NEXT:    sbbb %al, %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
@@ -1538,11 +1488,9 @@ define i1 @bool_reduction_v32i8(<32 x i8> %x, <32 x i8> %y) {
 ; SSE-NEXT:    pcmpeqb %xmm2, %xmm0
 ; SSE-NEXT:    pcmpeqb %xmm3, %xmm1
 ; SSE-NEXT:    por %xmm0, %xmm1
-; SSE-NEXT:    pmovmskb %xmm1, %ecx
-; SSE-NEXT:    xorl %eax, %eax
-; SSE-NEXT:    cmpl %ecx, %eax
-; SSE-NEXT:    sbbl %eax, %eax
-; SSE-NEXT:    # kill: def $al killed $al killed $eax
+; SSE-NEXT:    pmovmskb %xmm1, %eax
+; SSE-NEXT:    negl %eax
+; SSE-NEXT:    sbbb %al, %al
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: bool_reduction_v32i8:
@@ -1568,11 +1516,9 @@ define i1 @bool_reduction_v32i8(<32 x i8> %x, <32 x i8> %y) {
 ; AVX2-LABEL: bool_reduction_v32i8:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vpcmpeqb %ymm1, %ymm0, %ymm0
-; AVX2-NEXT:    vpmovmskb %ymm0, %ecx
-; AVX2-NEXT:    xorl %eax, %eax
-; AVX2-NEXT:    cmpl %ecx, %eax
-; AVX2-NEXT:    sbbl %eax, %eax
-; AVX2-NEXT:    # kill: def $al killed $al killed $eax
+; AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; AVX2-NEXT:    negl %eax
+; AVX2-NEXT:    sbbb %al, %al
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;