[X86] SimplifyDemandedVectorElts - attempt to recombine target shuffle using Demanded...

If we don't demand all elements, then attempt to combine to a simpler shuffle.

At the moment we can only do this if Depth == 0 as combineX86ShufflesRecursively uses Depth to track whether the shuffle has really changed or not - we'll need to change this before we can properly start merging combineX86ShufflesRecursively into SimplifyDemandedVectorElts.

The insertps-combine.ll regression is because XFormVExtractWithShuffleIntoLoad can't see through shuffles of different widths - this will be fixed in a follow-up commit.

Reapplying this as rL368307 had to be reverted as part of rL368660 to revert rL368276

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@368662 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index a82f2138562be7cfb019ebc379b996ed1fec56ee..66d7b76a98de4294b6f40a20407d2606c80055f0 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -33920,6 +33920,23 @@ bool X86TargetLowering::SimplifyDemandedVectorEltsForTargetNode(
       return true;
   }
 
+  // If we don't demand all elements, then attempt to combine to a simpler
+  // shuffle.
+  // TODO: Handle other depths, but first we need to handle the fact that
+  // it might combine to the same shuffle.
+  if (!DemandedElts.isAllOnesValue() && Depth == 0) {
+    SmallVector<int, 32> DemandedMask(NumElts, SM_SentinelUndef);
+    for (int i = 0; i != NumElts; ++i)
+      if (DemandedElts[i])
+        DemandedMask[i] = i;
+
+    SDValue NewShuffle = combineX86ShufflesRecursively(
+        {Op}, 0, Op, DemandedMask, {}, Depth, /*HasVarMask*/ false,
+        /*AllowVarMask*/ true, TLO.DAG, Subtarget);
+    if (NewShuffle)
+      return TLO.CombineTo(Op, NewShuffle);
+  }
+
   // Extract known zero/undef elements.
   // TODO - Propagate input undef/zero elts.
   for (int i = 0; i != NumElts; ++i) {

diff --git a/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll b/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
index d7611d1de0a78418caa2ab76bdfbd09a3b3dd746..4da750e9ecdfb6aad841a617c103da2214684975 100644 (file)
--- a/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
+++ b/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
@@ -8319,7 +8319,7 @@ define float @test_mm512_reduce_max_ps(<16 x float> %__W) {
 ; X86-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vmaxss %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -8336,7 +8336,7 @@ define float @test_mm512_reduce_max_ps(<16 x float> %__W) {
 ; X64-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vmaxss %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
@@ -8445,7 +8445,7 @@ define float @test_mm512_reduce_min_ps(<16 x float> %__W) {
 ; X86-NEXT:    vminps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vminps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vminss %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -8462,7 +8462,7 @@ define float @test_mm512_reduce_min_ps(<16 x float> %__W) {
 ; X64-NEXT:    vminps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vminps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vminss %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
@@ -8623,7 +8623,7 @@ define float @test_mm512_mask_reduce_max_ps(i16 zeroext %__M, <16 x float> %__W)
 ; X86-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vmaxss %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -8643,7 +8643,7 @@ define float @test_mm512_mask_reduce_max_ps(i16 zeroext %__M, <16 x float> %__W)
 ; X64-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vmaxps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vmaxss %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
@@ -8808,7 +8808,7 @@ define float @test_mm512_mask_reduce_min_ps(i16 zeroext %__M, <16 x float> %__W)
 ; X86-NEXT:    vminps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vminps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vminss %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -8828,7 +8828,7 @@ define float @test_mm512_mask_reduce_min_ps(i16 zeroext %__M, <16 x float> %__W)
 ; X64-NEXT:    vminps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vminps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vminss %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq

diff --git a/test/CodeGen/X86/insertps-combine.ll b/test/CodeGen/X86/insertps-combine.ll
index 6bef76ee9df8b6afa0eb055b2059a090e21f3847..98ed157405c12758fe35300066a60dce7814c887 100644 (file)
--- a/test/CodeGen/X86/insertps-combine.ll
+++ b/test/CodeGen/X86/insertps-combine.ll
@@ -285,12 +285,13 @@ define float @extract_lane_insertps_5123(<4 x float> %a0, <4 x float> *%p1) {
 define float @extract_lane_insertps_6123(<4 x float> %a0, <4 x float> *%p1) {
 ; SSE-LABEL: extract_lane_insertps_6123:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; SSE-NEXT:    movaps (%rdi), %xmm0
+; SSE-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: extract_lane_insertps_6123:
 ; AVX:       # %bb.0:
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = mem[1,0]
 ; AVX-NEXT:    retq
   %a1 = load <4 x float>, <4 x float> *%p1
   %res = call <4 x float> @llvm.x86.sse41.insertps(<4 x float> %a0, <4 x float> %a1, i8 128)

diff --git a/test/CodeGen/X86/shrink_vmul.ll b/test/CodeGen/X86/shrink_vmul.ll
index 5ceb299befc1a0df041f00cf353fc6c0666f4bfa..1c1032a4c321af9fa035eb8ba4a2f7ca8d2a413e 100644 (file)
--- a/test/CodeGen/X86/shrink_vmul.ll
+++ b/test/CodeGen/X86/shrink_vmul.ll
@@ -2085,85 +2085,88 @@ define void @PR34947(<9 x i16>* %p0, <9 x i32>* %p1) nounwind {
 ; X86-SSE-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-SSE-NEXT:    movdqa (%eax), %xmm5
-; X86-SSE-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-SSE-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; X86-SSE-NEXT:    movdqa (%ecx), %xmm2
 ; X86-SSE-NEXT:    movdqa 16(%ecx), %xmm6
-; X86-SSE-NEXT:    pxor %xmm0, %xmm0
-; X86-SSE-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X86-SSE-NEXT:    pxor %xmm1, %xmm1
+; X86-SSE-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
 ; X86-SSE-NEXT:    movdqa %xmm5, %xmm4
-; X86-SSE-NEXT:    punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm0[0],xmm4[1],xmm0[1],xmm4[2],xmm0[2],xmm4[3],xmm0[3]
-; X86-SSE-NEXT:    punpckhwd {{.*#+}} xmm5 = xmm5[4],xmm0[4],xmm5[5],xmm0[5],xmm5[6],xmm0[6],xmm5[7],xmm0[7]
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[3,1,2,3]
-; X86-SSE-NEXT:    movd %xmm0, %eax
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[3,1,2,3]
-; X86-SSE-NEXT:    movd %xmm0, %esi
+; X86-SSE-NEXT:    punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3]
+; X86-SSE-NEXT:    movdqa %xmm5, %xmm3
+; X86-SSE-NEXT:    punpckhwd {{.*#+}} xmm3 = xmm3[4],xmm1[4],xmm3[5],xmm1[5],xmm3[6],xmm1[6],xmm3[7],xmm1[7]
+; X86-SSE-NEXT:    movdqa %xmm5, %xmm1
+; X86-SSE-NEXT:    psrldq {{.*#+}} xmm1 = xmm1[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X86-SSE-NEXT:    movd %xmm1, %eax
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[3,1,2,3]
+; X86-SSE-NEXT:    movd %xmm1, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
-; X86-SSE-NEXT:    movd %edx, %xmm0
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[2,3,0,1]
-; X86-SSE-NEXT:    movd %xmm3, %eax
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm6[2,3,0,1]
-; X86-SSE-NEXT:    movd %xmm3, %esi
+; X86-SSE-NEXT:    movd %edx, %xmm1
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm3[2,3,0,1]
+; X86-SSE-NEXT:    movd %xmm7, %eax
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm6[2,3,0,1]
+; X86-SSE-NEXT:    movd %xmm7, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
 ; X86-SSE-NEXT:    movd %edx, %xmm7
-; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm0[0],xmm7[1],xmm0[1]
-; X86-SSE-NEXT:    movd %xmm5, %eax
+; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm1[0],xmm7[1],xmm1[1]
+; X86-SSE-NEXT:    movd %xmm3, %eax
 ; X86-SSE-NEXT:    movd %xmm6, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,2,3]
+; X86-SSE-NEXT:    movd %xmm3, %eax
 ; X86-SSE-NEXT:    movd %edx, %xmm3
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,2,3]
-; X86-SSE-NEXT:    movd %xmm5, %eax
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm6[1,1,2,3]
-; X86-SSE-NEXT:    movd %xmm5, %esi
-; X86-SSE-NEXT:    xorl %edx, %edx
-; X86-SSE-NEXT:    divl %esi
-; X86-SSE-NEXT:    movd %edx, %xmm5
-; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm5[0],xmm3[1],xmm5[1]
-; X86-SSE-NEXT:    punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm7[0]
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[3,1,2,3]
-; X86-SSE-NEXT:    movd %xmm6, %eax
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm2[3,1,2,3]
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,2,3]
 ; X86-SSE-NEXT:    movd %xmm6, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
 ; X86-SSE-NEXT:    movd %edx, %xmm6
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm4[2,3,0,1]
+; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm6[0],xmm3[1],xmm6[1]
+; X86-SSE-NEXT:    punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm7[0]
+; X86-SSE-NEXT:    movdqa %xmm5, %xmm7
+; X86-SSE-NEXT:    psrld $16, %xmm7
 ; X86-SSE-NEXT:    movd %xmm7, %eax
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm2[2,3,0,1]
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm2[1,1,2,3]
 ; X86-SSE-NEXT:    movd %xmm7, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
 ; X86-SSE-NEXT:    movd %edx, %xmm7
-; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm6[0],xmm7[1],xmm6[1]
 ; X86-SSE-NEXT:    movd %xmm4, %eax
 ; X86-SSE-NEXT:    movd %xmm2, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,2,3]
+; X86-SSE-NEXT:    psrlq $48, %xmm5
+; X86-SSE-NEXT:    movd %xmm5, %eax
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[3,1,2,3]
+; X86-SSE-NEXT:    movd %xmm5, %esi
+; X86-SSE-NEXT:    movd %edx, %xmm5
+; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm5 = xmm5[0],xmm7[0],xmm5[1],xmm7[1]
+; X86-SSE-NEXT:    xorl %edx, %edx
+; X86-SSE-NEXT:    divl %esi
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[2,3,0,1]
 ; X86-SSE-NEXT:    movd %xmm4, %eax
 ; X86-SSE-NEXT:    movd %edx, %xmm4
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,2,3]
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
 ; X86-SSE-NEXT:    movd %xmm2, %esi
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl %esi
 ; X86-SSE-NEXT:    movd %edx, %xmm2
+; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1]
+; X86-SSE-NEXT:    punpcklqdq {{.*#+}} xmm5 = xmm5[0],xmm2[0]
+; X86-SSE-NEXT:    movd %xmm0, %eax
+; X86-SSE-NEXT:    shufps {{.*#+}} xmm7 = xmm7[0,0],xmm4[0,0]
+; X86-SSE-NEXT:    movdqa {{.*#+}} xmm0 = [8199,8199,8199,8199]
+; X86-SSE-NEXT:    pmuludq %xmm0, %xmm7
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm7[0,2,2,3]
+; X86-SSE-NEXT:    pmuludq %xmm0, %xmm5
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm5[0,2,2,3]
 ; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1]
-; X86-SSE-NEXT:    punpcklqdq {{.*#+}} xmm4 = xmm4[0],xmm7[0]
-; X86-SSE-NEXT:    movd %xmm1, %eax
-; X86-SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,0],xmm6[0,0]
-; X86-SSE-NEXT:    movdqa {{.*#+}} xmm1 = [8199,8199,8199,8199]
-; X86-SSE-NEXT:    pmuludq %xmm1, %xmm4
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[0,2,2,3]
-; X86-SSE-NEXT:    pmuludq %xmm1, %xmm2
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm2[0],xmm4[1],xmm2[1]
-; X86-SSE-NEXT:    shufps {{.*#+}} xmm5 = xmm5[0,0],xmm0[0,0]
-; X86-SSE-NEXT:    pmuludq %xmm1, %xmm3
-; X86-SSE-NEXT:    pmuludq %xmm1, %xmm5
+; X86-SSE-NEXT:    shufps {{.*#+}} xmm6 = xmm6[0,0],xmm1[0,0]
+; X86-SSE-NEXT:    pmuludq %xmm0, %xmm3
+; X86-SSE-NEXT:    pmuludq %xmm0, %xmm6
 ; X86-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
-; X86-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[0,2,2,3]
+; X86-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[0,2,2,3]
 ; X86-SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; X86-SSE-NEXT:    xorl %edx, %edx
 ; X86-SSE-NEXT:    divl 32(%ecx)
@@ -2324,92 +2327,95 @@ define void @PR34947(<9 x i16>* %p0, <9 x i32>* %p1) nounwind {
 ; X64-SSE-LABEL: PR34947:
 ; X64-SSE:       # %bb.0:
 ; X64-SSE-NEXT:    movdqa (%rdi), %xmm5
-; X64-SSE-NEXT:    movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X64-SSE-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; X64-SSE-NEXT:    movdqa (%rsi), %xmm2
 ; X64-SSE-NEXT:    movdqa 16(%rsi), %xmm6
-; X64-SSE-NEXT:    pxor %xmm0, %xmm0
-; X64-SSE-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
+; X64-SSE-NEXT:    pxor %xmm1, %xmm1
+; X64-SSE-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
 ; X64-SSE-NEXT:    movdqa %xmm5, %xmm3
-; X64-SSE-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3]
-; X64-SSE-NEXT:    punpckhwd {{.*#+}} xmm5 = xmm5[4],xmm0[4],xmm5[5],xmm0[5],xmm5[6],xmm0[6],xmm5[7],xmm0[7]
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[3,1,2,3]
-; X64-SSE-NEXT:    movd %xmm0, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[3,1,2,3]
-; X64-SSE-NEXT:    movd %xmm0, %ecx
+; X64-SSE-NEXT:    punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1],xmm3[2],xmm1[2],xmm3[3],xmm1[3]
+; X64-SSE-NEXT:    movdqa %xmm5, %xmm7
+; X64-SSE-NEXT:    punpckhwd {{.*#+}} xmm7 = xmm7[4],xmm1[4],xmm7[5],xmm1[5],xmm7[6],xmm1[6],xmm7[7],xmm1[7]
+; X64-SSE-NEXT:    movdqa %xmm5, %xmm1
+; X64-SSE-NEXT:    psrldq {{.*#+}} xmm1 = xmm1[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
+; X64-SSE-NEXT:    movd %xmm1, %eax
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[3,1,2,3]
+; X64-SSE-NEXT:    movd %xmm1, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
 ; X64-SSE-NEXT:    movd %edx, %xmm8
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm5[2,3,0,1]
-; X64-SSE-NEXT:    movd %xmm4, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm6[2,3,0,1]
-; X64-SSE-NEXT:    movd %xmm4, %ecx
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm7[2,3,0,1]
+; X64-SSE-NEXT:    movd %xmm1, %eax
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[2,3,0,1]
+; X64-SSE-NEXT:    movd %xmm1, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
-; X64-SSE-NEXT:    movd %edx, %xmm7
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm8[0],xmm7[1],xmm8[1]
-; X64-SSE-NEXT:    movd %xmm5, %eax
+; X64-SSE-NEXT:    movd %edx, %xmm1
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm8[0],xmm1[1],xmm8[1]
+; X64-SSE-NEXT:    movd %xmm7, %eax
 ; X64-SSE-NEXT:    movd %xmm6, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
 ; X64-SSE-NEXT:    movd %edx, %xmm4
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,2,3]
-; X64-SSE-NEXT:    movd %xmm5, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm6[1,1,2,3]
-; X64-SSE-NEXT:    movd %xmm5, %ecx
-; X64-SSE-NEXT:    xorl %edx, %edx
-; X64-SSE-NEXT:    divl %ecx
-; X64-SSE-NEXT:    movd %edx, %xmm5
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm5[0],xmm4[1],xmm5[1]
-; X64-SSE-NEXT:    punpcklqdq {{.*#+}} xmm4 = xmm4[0],xmm7[0]
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm3[3,1,2,3]
-; X64-SSE-NEXT:    movd %xmm6, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm2[3,1,2,3]
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[1,1,2,3]
+; X64-SSE-NEXT:    movd %xmm7, %eax
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[1,1,2,3]
 ; X64-SSE-NEXT:    movd %xmm6, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
 ; X64-SSE-NEXT:    movd %edx, %xmm6
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm3[2,3,0,1]
-; X64-SSE-NEXT:    movd %xmm7, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm2[2,3,0,1]
-; X64-SSE-NEXT:    movd %xmm7, %ecx
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm6[0],xmm4[1],xmm6[1]
+; X64-SSE-NEXT:    punpcklqdq {{.*#+}} xmm4 = xmm4[0],xmm1[0]
+; X64-SSE-NEXT:    movdqa %xmm5, %xmm1
+; X64-SSE-NEXT:    psrld $16, %xmm1
+; X64-SSE-NEXT:    movd %xmm1, %eax
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,2,3]
+; X64-SSE-NEXT:    movd %xmm1, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
 ; X64-SSE-NEXT:    movd %edx, %xmm7
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm7 = xmm7[0],xmm6[0],xmm7[1],xmm6[1]
 ; X64-SSE-NEXT:    movd %xmm3, %eax
 ; X64-SSE-NEXT:    movd %xmm2, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
-; X64-SSE-NEXT:    movd %edx, %xmm0
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,2,3]
+; X64-SSE-NEXT:    movd %edx, %xmm1
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm7[0],xmm1[1],xmm7[1]
+; X64-SSE-NEXT:    psrlq $48, %xmm5
+; X64-SSE-NEXT:    movd %xmm5, %eax
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[3,1,2,3]
+; X64-SSE-NEXT:    movd %xmm5, %ecx
+; X64-SSE-NEXT:    xorl %edx, %edx
+; X64-SSE-NEXT:    divl %ecx
+; X64-SSE-NEXT:    movd %edx, %xmm5
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[2,3,0,1]
 ; X64-SSE-NEXT:    movd %xmm3, %eax
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,2,3]
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1]
 ; X64-SSE-NEXT:    movd %xmm2, %ecx
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl %ecx
 ; X64-SSE-NEXT:    movd %edx, %xmm2
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X64-SSE-NEXT:    punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm7[0]
-; X64-SSE-NEXT:    movd %xmm1, %eax
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm5[0],xmm2[1],xmm5[1]
+; X64-SSE-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
+; X64-SSE-NEXT:    movd %xmm0, %eax
 ; X64-SSE-NEXT:    xorl %edx, %edx
 ; X64-SSE-NEXT:    divl 32(%rsi)
-; X64-SSE-NEXT:    movdqa {{.*#+}} xmm1 = [8199,8199,8199,8199]
-; X64-SSE-NEXT:    pmuludq %xmm1, %xmm0
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; X64-SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[0,0],xmm6[0,0]
-; X64-SSE-NEXT:    pmuludq %xmm1, %xmm2
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; X64-SSE-NEXT:    pmuludq %xmm1, %xmm4
+; X64-SSE-NEXT:    shufps {{.*#+}} xmm7 = xmm7[0,0],xmm5[0,0]
+; X64-SSE-NEXT:    movdqa {{.*#+}} xmm0 = [8199,8199,8199,8199]
+; X64-SSE-NEXT:    pmuludq %xmm0, %xmm7
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm7[0,2,2,3]
+; X64-SSE-NEXT:    pmuludq %xmm0, %xmm1
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
+; X64-SSE-NEXT:    pmuludq %xmm0, %xmm4
 ; X64-SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm4[0,2,2,3]
-; X64-SSE-NEXT:    shufps {{.*#+}} xmm5 = xmm5[0,0],xmm8[0,0]
-; X64-SSE-NEXT:    pmuludq %xmm1, %xmm5
-; X64-SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[0,2,2,3]
-; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; X64-SSE-NEXT:    shufps {{.*#+}} xmm6 = xmm6[0,0],xmm8[0,0]
+; X64-SSE-NEXT:    pmuludq %xmm0, %xmm6
+; X64-SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm6[0,2,2,3]
+; X64-SSE-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1]
 ; X64-SSE-NEXT:    imull $8199, %edx, %eax # imm = 0x2007
 ; X64-SSE-NEXT:    movl %eax, (%rax)
 ; X64-SSE-NEXT:    movdqa %xmm2, (%rax)
-; X64-SSE-NEXT:    movdqa %xmm0, (%rax)
+; X64-SSE-NEXT:    movdqa %xmm1, (%rax)
 ; X64-SSE-NEXT:    retq
 ;
 ; X64-AVX1-LABEL: PR34947:

diff --git a/test/CodeGen/X86/vec_smulo.ll b/test/CodeGen/X86/vec_smulo.ll
index 385c556e969048e1e666ddb3a2d2bb8c2c5a412d..a34824dd10622d6f645cbd559911db867ba6abb4 100644 (file)
--- a/test/CodeGen/X86/vec_smulo.ll
+++ b/test/CodeGen/X86/vec_smulo.ll
@@ -195,17 +195,17 @@ define <3 x i32> @smulo_v3i32(<3 x i32> %a0, <3 x i32> %a1, <3 x i32>* %p2) noun
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
 ; SSE2-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1]
 ; SSE2-NEXT:    psubd %xmm2, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[0,2,2,3]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE2-NEXT:    movq %xmm0, (%rdi)
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
-; SSE2-NEXT:    movdqa %xmm0, %xmm2
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; SSE2-NEXT:    movq %xmm2, (%rdi)
 ; SSE2-NEXT:    psrad $31, %xmm2
 ; SSE2-NEXT:    pcmpeqd %xmm4, %xmm2
-; SSE2-NEXT:    pcmpeqd %xmm0, %xmm0
-; SSE2-NEXT:    pxor %xmm2, %xmm0
-; SSE2-NEXT:    movd %xmm1, 8(%rdi)
+; SSE2-NEXT:    pcmpeqd %xmm1, %xmm1
+; SSE2-NEXT:    pxor %xmm2, %xmm1
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
+; SSE2-NEXT:    movd %xmm0, 8(%rdi)
+; SSE2-NEXT:    movdqa %xmm1, %xmm0
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: smulo_v3i32:
@@ -225,17 +225,17 @@ define <3 x i32> @smulo_v3i32(<3 x i32> %a0, <3 x i32> %a1, <3 x i32>* %p2) noun
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[1,3,2,3]
 ; SSSE3-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1]
 ; SSSE3-NEXT:    psubd %xmm2, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[0,2,2,3]
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSSE3-NEXT:    movq %xmm0, (%rdi)
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
-; SSSE3-NEXT:    movdqa %xmm0, %xmm2
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; SSSE3-NEXT:    movq %xmm2, (%rdi)
 ; SSSE3-NEXT:    psrad $31, %xmm2
 ; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm2
-; SSSE3-NEXT:    pcmpeqd %xmm0, %xmm0
-; SSSE3-NEXT:    pxor %xmm2, %xmm0
-; SSSE3-NEXT:    movd %xmm1, 8(%rdi)
+; SSSE3-NEXT:    pcmpeqd %xmm1, %xmm1
+; SSSE3-NEXT:    pxor %xmm2, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
+; SSSE3-NEXT:    movd %xmm0, 8(%rdi)
+; SSSE3-NEXT:    movdqa %xmm1, %xmm0
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: smulo_v3i32:
@@ -1767,52 +1767,52 @@ define <4 x i32> @smulo_v4i24(<4 x i24> %a0, <4 x i24> %a1, <4 x i24>* %p2) noun
 ; SSE2-NEXT:    psrad $8, %xmm0
 ; SSE2-NEXT:    pslld $8, %xmm1
 ; SSE2-NEXT:    psrad $8, %xmm1
-; SSE2-NEXT:    pxor %xmm4, %xmm4
+; SSE2-NEXT:    pxor %xmm3, %xmm3
 ; SSE2-NEXT:    pxor %xmm2, %xmm2
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm2
 ; SSE2-NEXT:    pand %xmm0, %xmm2
-; SSE2-NEXT:    pcmpgtd %xmm0, %xmm4
-; SSE2-NEXT:    pand %xmm1, %xmm4
-; SSE2-NEXT:    paddd %xmm2, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm3
+; SSE2-NEXT:    pand %xmm1, %xmm3
+; SSE2-NEXT:    paddd %xmm2, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
 ; SSE2-NEXT:    pmuludq %xmm1, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[1,3,2,3]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,1,3,3]
-; SSE2-NEXT:    pmuludq %xmm5, %xmm2
+; SSE2-NEXT:    pmuludq %xmm4, %xmm2
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,3,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
-; SSE2-NEXT:    psubd %xmm4, %xmm3
-; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,2,2,3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm5 = xmm5[0],xmm1[0],xmm5[1],xmm1[1]
+; SSE2-NEXT:    psubd %xmm3, %xmm5
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[0,2,2,3]
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[0,2,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSE2-NEXT:    movdqa %xmm4, %xmm1
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; SSE2-NEXT:    movdqa %xmm3, %xmm1
 ; SSE2-NEXT:    pslld $8, %xmm1
 ; SSE2-NEXT:    psrad $8, %xmm1
-; SSE2-NEXT:    pcmpeqd %xmm4, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[3,1,2,3]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[2,3,0,1]
-; SSE2-NEXT:    psrad $31, %xmm4
-; SSE2-NEXT:    pcmpeqd %xmm3, %xmm4
-; SSE2-NEXT:    pcmpeqd %xmm3, %xmm3
-; SSE2-NEXT:    pxor %xmm3, %xmm4
-; SSE2-NEXT:    pxor %xmm3, %xmm1
-; SSE2-NEXT:    por %xmm4, %xmm1
+; SSE2-NEXT:    pcmpeqd %xmm3, %xmm1
+; SSE2-NEXT:    psrad $31, %xmm3
+; SSE2-NEXT:    pcmpeqd %xmm5, %xmm3
+; SSE2-NEXT:    pcmpeqd %xmm4, %xmm4
+; SSE2-NEXT:    pxor %xmm4, %xmm3
+; SSE2-NEXT:    pxor %xmm4, %xmm1
+; SSE2-NEXT:    por %xmm3, %xmm1
 ; SSE2-NEXT:    movd %xmm0, %eax
 ; SSE2-NEXT:    movw %ax, (%rdi)
-; SSE2-NEXT:    movd %xmm2, %ecx
-; SSE2-NEXT:    movw %cx, 3(%rdi)
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
+; SSE2-NEXT:    movd %xmm0, %ecx
+; SSE2-NEXT:    movw %cx, 6(%rdi)
+; SSE2-NEXT:    movd %xmm2, %edx
+; SSE2-NEXT:    movw %dx, 3(%rdi)
 ; SSE2-NEXT:    shrl $16, %eax
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    shrl $16, %ecx
-; SSE2-NEXT:    movb %cl, 5(%rdi)
-; SSE2-NEXT:    movd %xmm5, %eax
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[2,3,2,3]
+; SSE2-NEXT:    movd %xmm0, %eax
 ; SSE2-NEXT:    movw %ax, 9(%rdi)
-; SSE2-NEXT:    movd %xmm6, %ecx
-; SSE2-NEXT:    movw %cx, 6(%rdi)
-; SSE2-NEXT:    shrl $16, %eax
-; SSE2-NEXT:    movb %al, 11(%rdi)
 ; SSE2-NEXT:    shrl $16, %ecx
 ; SSE2-NEXT:    movb %cl, 8(%rdi)
+; SSE2-NEXT:    shrl $16, %edx
+; SSE2-NEXT:    movb %dl, 5(%rdi)
+; SSE2-NEXT:    shrl $16, %eax
+; SSE2-NEXT:    movb %al, 11(%rdi)
 ; SSE2-NEXT:    movdqa %xmm1, %xmm0
 ; SSE2-NEXT:    retq
 ;
@@ -1822,52 +1822,52 @@ define <4 x i32> @smulo_v4i24(<4 x i24> %a0, <4 x i24> %a1, <4 x i24>* %p2) noun
 ; SSSE3-NEXT:    psrad $8, %xmm0
 ; SSSE3-NEXT:    pslld $8, %xmm1
 ; SSSE3-NEXT:    psrad $8, %xmm1
-; SSSE3-NEXT:    pxor %xmm4, %xmm4
+; SSSE3-NEXT:    pxor %xmm3, %xmm3
 ; SSSE3-NEXT:    pxor %xmm2, %xmm2
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm2
 ; SSSE3-NEXT:    pand %xmm0, %xmm2
-; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
-; SSSE3-NEXT:    pand %xmm1, %xmm4
-; SSSE3-NEXT:    paddd %xmm2, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[1,1,3,3]
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm3
+; SSSE3-NEXT:    pand %xmm1, %xmm3
+; SSSE3-NEXT:    paddd %xmm2, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
 ; SSSE3-NEXT:    pmuludq %xmm1, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[1,3,2,3]
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    pmuludq %xmm5, %xmm2
+; SSSE3-NEXT:    pmuludq %xmm4, %xmm2
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,3,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
-; SSSE3-NEXT:    psubd %xmm4, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,2,2,3]
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm5 = xmm5[0],xmm1[0],xmm5[1],xmm1[1]
+; SSSE3-NEXT:    psubd %xmm3, %xmm5
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[0,2,2,3]
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
-; SSSE3-NEXT:    movdqa %xmm4, %xmm1
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; SSSE3-NEXT:    movdqa %xmm3, %xmm1
 ; SSSE3-NEXT:    pslld $8, %xmm1
 ; SSSE3-NEXT:    psrad $8, %xmm1
-; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[3,1,2,3]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[2,3,0,1]
-; SSSE3-NEXT:    psrad $31, %xmm4
-; SSSE3-NEXT:    pcmpeqd %xmm3, %xmm4
-; SSSE3-NEXT:    pcmpeqd %xmm3, %xmm3
-; SSSE3-NEXT:    pxor %xmm3, %xmm4
-; SSSE3-NEXT:    pxor %xmm3, %xmm1
-; SSSE3-NEXT:    por %xmm4, %xmm1
+; SSSE3-NEXT:    pcmpeqd %xmm3, %xmm1
+; SSSE3-NEXT:    psrad $31, %xmm3
+; SSSE3-NEXT:    pcmpeqd %xmm5, %xmm3
+; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm4
+; SSSE3-NEXT:    pxor %xmm4, %xmm3
+; SSSE3-NEXT:    pxor %xmm4, %xmm1
+; SSSE3-NEXT:    por %xmm3, %xmm1
 ; SSSE3-NEXT:    movd %xmm0, %eax
 ; SSSE3-NEXT:    movw %ax, (%rdi)
-; SSSE3-NEXT:    movd %xmm2, %ecx
-; SSSE3-NEXT:    movw %cx, 3(%rdi)
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
+; SSSE3-NEXT:    movd %xmm0, %ecx
+; SSSE3-NEXT:    movw %cx, 6(%rdi)
+; SSSE3-NEXT:    movd %xmm2, %edx
+; SSSE3-NEXT:    movw %dx, 3(%rdi)
 ; SSSE3-NEXT:    shrl $16, %eax
 ; SSSE3-NEXT:    movb %al, 2(%rdi)
-; SSSE3-NEXT:    shrl $16, %ecx
-; SSSE3-NEXT:    movb %cl, 5(%rdi)
-; SSSE3-NEXT:    movd %xmm5, %eax
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[2,3,2,3]
+; SSSE3-NEXT:    movd %xmm0, %eax
 ; SSSE3-NEXT:    movw %ax, 9(%rdi)
-; SSSE3-NEXT:    movd %xmm6, %ecx
-; SSSE3-NEXT:    movw %cx, 6(%rdi)
-; SSSE3-NEXT:    shrl $16, %eax
-; SSSE3-NEXT:    movb %al, 11(%rdi)
 ; SSSE3-NEXT:    shrl $16, %ecx
 ; SSSE3-NEXT:    movb %cl, 8(%rdi)
+; SSSE3-NEXT:    shrl $16, %edx
+; SSSE3-NEXT:    movb %dl, 5(%rdi)
+; SSSE3-NEXT:    shrl $16, %eax
+; SSSE3-NEXT:    movb %al, 11(%rdi)
 ; SSSE3-NEXT:    movdqa %xmm1, %xmm0
 ; SSSE3-NEXT:    retq
 ;

diff --git a/test/CodeGen/X86/vec_umulo.ll b/test/CodeGen/X86/vec_umulo.ll
index c8146e34468f69e5da9fd99d73994c71a6e98864..52b920557857a5a66146a2674340afff0ff81d58 100644 (file)
--- a/test/CodeGen/X86/vec_umulo.ll
+++ b/test/CodeGen/X86/vec_umulo.ll
@@ -179,12 +179,10 @@ define <3 x i32> @umulo_v3i32(<3 x i32> %a0, <3 x i32> %a1, <3 x i32>* %p2) noun
 ; SSE2-NEXT:    pcmpeqd %xmm3, %xmm2
 ; SSE2-NEXT:    pcmpeqd %xmm1, %xmm1
 ; SSE2-NEXT:    pxor %xmm2, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm4[0,2,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1]
+; SSE2-NEXT:    movd %xmm2, 8(%rdi)
 ; SSE2-NEXT:    movq %xmm0, (%rdi)
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; SSE2-NEXT:    movd %xmm0, 8(%rdi)
 ; SSE2-NEXT:    movdqa %xmm1, %xmm0
 ; SSE2-NEXT:    retq
 ;
@@ -201,12 +199,10 @@ define <3 x i32> @umulo_v3i32(<3 x i32> %a0, <3 x i32> %a1, <3 x i32>* %p2) noun
 ; SSSE3-NEXT:    pcmpeqd %xmm3, %xmm2
 ; SSSE3-NEXT:    pcmpeqd %xmm1, %xmm1
 ; SSSE3-NEXT:    pxor %xmm2, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm4[0,2,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1]
+; SSSE3-NEXT:    movd %xmm2, 8(%rdi)
 ; SSSE3-NEXT:    movq %xmm0, (%rdi)
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
-; SSSE3-NEXT:    movd %xmm0, 8(%rdi)
 ; SSSE3-NEXT:    movdqa %xmm1, %xmm0
 ; SSSE3-NEXT:    retq
 ;
@@ -1567,90 +1563,90 @@ define <2 x i32> @umulo_v2i64(<2 x i64> %a0, <2 x i64> %a1, <2 x i64>* %p2) noun
 define <4 x i32> @umulo_v4i24(<4 x i24> %a0, <4 x i24> %a1, <4 x i24>* %p2) nounwind {
 ; SSE2-LABEL: umulo_v4i24:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
-; SSE2-NEXT:    pand %xmm2, %xmm1
-; SSE2-NEXT:    pand %xmm2, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pmuludq %xmm1, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,1,3,3]
-; SSE2-NEXT:    pmuludq %xmm4, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,3,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; SSE2-NEXT:    movdqa %xmm0, %xmm2
+; SSE2-NEXT:    movdqa {{.*#+}} xmm0 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
+; SSE2-NEXT:    pand %xmm0, %xmm1
+; SSE2-NEXT:    pand %xmm0, %xmm2
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[1,1,3,3]
+; SSE2-NEXT:    pmuludq %xmm1, %xmm2
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[1,3,2,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSE2-NEXT:    pmuludq %xmm0, %xmm1
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,3,2,3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1]
 ; SSE2-NEXT:    pxor %xmm4, %xmm4
 ; SSE2-NEXT:    pcmpeqd %xmm4, %xmm3
-; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[0,2,2,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm5[0],xmm1[1],xmm5[1]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[3,1,2,3]
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[2,3,0,1]
-; SSE2-NEXT:    psrld $24, %xmm1
-; SSE2-NEXT:    pcmpeqd %xmm4, %xmm1
-; SSE2-NEXT:    pcmpeqd %xmm4, %xmm4
-; SSE2-NEXT:    pxor %xmm4, %xmm3
-; SSE2-NEXT:    pxor %xmm4, %xmm1
-; SSE2-NEXT:    por %xmm3, %xmm1
-; SSE2-NEXT:    movd %xmm0, %eax
+; SSE2-NEXT:    pcmpeqd %xmm5, %xmm5
+; SSE2-NEXT:    pxor %xmm5, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[0,2,2,3]
+; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1]
+; SSE2-NEXT:    psrld $24, %xmm0
+; SSE2-NEXT:    pcmpeqd %xmm4, %xmm0
+; SSE2-NEXT:    pxor %xmm5, %xmm0
+; SSE2-NEXT:    por %xmm3, %xmm0
+; SSE2-NEXT:    movd %xmm2, %eax
 ; SSE2-NEXT:    movw %ax, (%rdi)
+; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
 ; SSE2-NEXT:    movd %xmm2, %ecx
-; SSE2-NEXT:    movw %cx, 3(%rdi)
+; SSE2-NEXT:    movw %cx, 6(%rdi)
+; SSE2-NEXT:    movd %xmm1, %edx
+; SSE2-NEXT:    movw %dx, 3(%rdi)
 ; SSE2-NEXT:    shrl $16, %eax
 ; SSE2-NEXT:    movb %al, 2(%rdi)
-; SSE2-NEXT:    shrl $16, %ecx
-; SSE2-NEXT:    movb %cl, 5(%rdi)
-; SSE2-NEXT:    movd %xmm5, %eax
+; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; SSE2-NEXT:    movd %xmm1, %eax
 ; SSE2-NEXT:    movw %ax, 9(%rdi)
-; SSE2-NEXT:    movd %xmm6, %ecx
-; SSE2-NEXT:    movw %cx, 6(%rdi)
-; SSE2-NEXT:    shrl $16, %eax
-; SSE2-NEXT:    movb %al, 11(%rdi)
 ; SSE2-NEXT:    shrl $16, %ecx
 ; SSE2-NEXT:    movb %cl, 8(%rdi)
-; SSE2-NEXT:    movdqa %xmm1, %xmm0
+; SSE2-NEXT:    shrl $16, %edx
+; SSE2-NEXT:    movb %dl, 5(%rdi)
+; SSE2-NEXT:    shrl $16, %eax
+; SSE2-NEXT:    movb %al, 11(%rdi)
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: umulo_v4i24:
 ; SSSE3:       # %bb.0:
-; SSSE3-NEXT:    movdqa {{.*#+}} xmm2 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
-; SSSE3-NEXT:    pand %xmm2, %xmm1
-; SSSE3-NEXT:    pand %xmm2, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pmuludq %xmm1, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,3,2,3]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    pmuludq %xmm4, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,3,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; SSSE3-NEXT:    movdqa %xmm0, %xmm2
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm0 = [255,255,255,0,255,255,255,0,255,255,255,0,255,255,255,0]
+; SSSE3-NEXT:    pand %xmm0, %xmm1
+; SSSE3-NEXT:    pand %xmm0, %xmm2
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[1,1,3,3]
+; SSSE3-NEXT:    pmuludq %xmm1, %xmm2
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[1,3,2,3]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
+; SSSE3-NEXT:    pmuludq %xmm0, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,3,2,3]
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1]
 ; SSSE3-NEXT:    pxor %xmm4, %xmm4
 ; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm2[0,2,2,3]
-; SSSE3-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm5[0],xmm1[1],xmm5[1]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[3,1,2,3]
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[2,3,0,1]
-; SSSE3-NEXT:    psrld $24, %xmm1
-; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm1
-; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm4
-; SSSE3-NEXT:    pxor %xmm4, %xmm3
-; SSSE3-NEXT:    pxor %xmm4, %xmm1
-; SSSE3-NEXT:    por %xmm3, %xmm1
-; SSSE3-NEXT:    movd %xmm0, %eax
+; SSSE3-NEXT:    pcmpeqd %xmm5, %xmm5
+; SSSE3-NEXT:    pxor %xmm5, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[0,2,2,3]
+; SSSE3-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm6[0],xmm0[1],xmm6[1]
+; SSSE3-NEXT:    psrld $24, %xmm0
+; SSSE3-NEXT:    pcmpeqd %xmm4, %xmm0
+; SSSE3-NEXT:    pxor %xmm5, %xmm0
+; SSSE3-NEXT:    por %xmm3, %xmm0
+; SSSE3-NEXT:    movd %xmm2, %eax
 ; SSSE3-NEXT:    movw %ax, (%rdi)
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,3,2,3]
 ; SSSE3-NEXT:    movd %xmm2, %ecx
-; SSSE3-NEXT:    movw %cx, 3(%rdi)
+; SSSE3-NEXT:    movw %cx, 6(%rdi)
+; SSSE3-NEXT:    movd %xmm1, %edx
+; SSSE3-NEXT:    movw %dx, 3(%rdi)
 ; SSSE3-NEXT:    shrl $16, %eax
 ; SSSE3-NEXT:    movb %al, 2(%rdi)
-; SSSE3-NEXT:    shrl $16, %ecx
-; SSSE3-NEXT:    movb %cl, 5(%rdi)
-; SSSE3-NEXT:    movd %xmm5, %eax
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
+; SSSE3-NEXT:    movd %xmm1, %eax
 ; SSSE3-NEXT:    movw %ax, 9(%rdi)
-; SSSE3-NEXT:    movd %xmm6, %ecx
-; SSSE3-NEXT:    movw %cx, 6(%rdi)
-; SSSE3-NEXT:    shrl $16, %eax
-; SSSE3-NEXT:    movb %al, 11(%rdi)
 ; SSSE3-NEXT:    shrl $16, %ecx
 ; SSSE3-NEXT:    movb %cl, 8(%rdi)
-; SSSE3-NEXT:    movdqa %xmm1, %xmm0
+; SSSE3-NEXT:    shrl $16, %edx
+; SSSE3-NEXT:    movb %dl, 5(%rdi)
+; SSSE3-NEXT:    shrl $16, %eax
+; SSSE3-NEXT:    movb %al, 11(%rdi)
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: umulo_v4i24: