[X86][SSE] Added support for combining target shuffles to (V)PSHUFD/VPERMILPD/VPERMILPS immediate permute

[RKSimon]

Summary

This patch allows target shuffles to be combined to single input immediate permute instructions - (V)PSHUFD/VPERMILPD/VPERMILPS - allowing more general pattern matching than what we current do and improves the likelihood of memory folding compared to existing patterns which tend to reuse the input in multiple arguments.

Further permute instructions (V)PSHUFLW/(V)PSHUFHW/(V)PERMQ/(V)PERMPD may be added in the future but its proven tricky to create tests cases for them so far. (V)PSHUFLW/(V)PSHUFHW is already handled quite well in combineTargetShuffle so it may be that removing some of that code may allow us to perform more of the combining in one place without duplication.

[RKSimon]

ping?

[RKSimon]

Rebased. Added an extra check that forces use of 256-bit float/double unary shuffles on AVX1 targets (since it doesn't have any 256-bit integer shuffles).

[RKSimon]

Fixed issue causing an unnecessary assertion on some AVX2 cross-lane shuffles.

[ab]

Code LGTM, but I'm not sure I have a clear picture; questions inline.

lib/Target/X86/X86ISelLowering.cpp

96–1

This looks like independent goodness; maybe extract that out?

test/CodeGen/X86/vector-shuffle-128-v2.ll

2

I'm surprised by this and other changes; isn't the combine for shuffle chains? (it does look better for folding though; just trying to understand)

test/CodeGen/X86/vector-shuffle-128-v4.ll

3

In particular, this looks slightly more expensive according to Agner's Intel tables (for the folded variants)

[RKSimon]

Updated to prefer binary shuffle (unpck mainly) over permutes - although this will prevent some folding its shouldn't affect register pressure. We don't handle i32 unpcks as pshufd typically has similar performance. The other changes you see from unary shuffle (e.g. movddup to pshufd) are typically because the target shuffle combine is a bit more ruthless at looking through bitcasts, hopefully reducing domain stalls.

[RKSimon]

lib/Target/X86/X86ISelLowering.cpp

24576–24578

I can't separate this easily - annoyingly this is only of use once we add support for permutes in this patch, and without it we see regressions with this patch.

[ab]

In D21148#465731, @RKSimon wrote:

Updated to prefer binary shuffle (unpck mainly) over permutes - although this will prevent some folding its shouldn't affect register pressure.

The permutes are indeed better because they either enable folding, or have similar performance to the binary shuffles (at least on the recent Intel CPUs I look at). If there's no other reason, revert to the previous patch? That one LGTM.
Very sorry I wasn't clear; was just curious!

We don't handle i32 unpcks as pshufd typically has similar performance. The other changes you see from unary shuffle (e.g. movddup to pshufd) are typically because the target shuffle combine is a bit more ruthless at looking through bitcasts, hopefully reducing domain stalls.

Sounds good, thanks!

test/CodeGen/X86/vector-shuffle-128-v2.ll

0

Extra NOTE

[RKSimon]

Updated based on Ahmed's comments. OK to commit?

[ab]

Yup yup, LGTM; thanks again.

lib/Target/X86/X86ISelLowering.cpp

 /// \brief Get a 4-lane 8-bit shuffle immediate for a mask.
 ///
 /// This helper function produces an 8-bit shuffle immediate corresponding to
 /// the ubiquitous shuffle encoding scheme used in x86 instructions for
 /// shuffling 4 lanes. It can be used with most of the PSHUF instructions for
 /// example.
 ///
 /// NB: We rely heavily on "undef" masks preserving the input lane.
-static SDValue getV4X86ShuffleImm8ForMask(ArrayRef<int> Mask, const SDLoc &DL,
-                                          SelectionDAG &DAG) {
+static unsigned getV4X86ShuffleImm(ArrayRef<int> Mask) {
   assert(Mask.size() == 4 && "Only 4-lane shuffle masks");
   assert(Mask[0] >= -1 && Mask[0] < 4 && "Out of bound mask element!");
   assert(Mask[1] >= -1 && Mask[1] < 4 && "Out of bound mask element!");
   assert(Mask[2] >= -1 && Mask[2] < 4 && "Out of bound mask element!");
   assert(Mask[3] >= -1 && Mask[3] < 4 && "Out of bound mask element!");
 
   unsigned Imm = 0;
   Imm |= (Mask[0] == -1 ? 0 : Mask[0]) << 0;
   Imm |= (Mask[1] == -1 ? 1 : Mask[1]) << 2;
   Imm |= (Mask[2] == -1 ? 2 : Mask[2]) << 4;
   Imm |= (Mask[3] == -1 ? 3 : Mask[3]) << 6;
-  return DAG.getConstant(Imm, DL, MVT::i8);
+  return Imm;
+}
+
+static SDValue getV4X86ShuffleImm8ForMask(ArrayRef<int> Mask, SDLoc DL,
+                                          SelectionDAG &DAG) {
+  return DAG.getConstant(getV4X86ShuffleImm(Mask), DL, MVT::i8);
 }
 
 /// \brief Compute whether each element of a shuffle is zeroable.
 ///
 /// A "zeroable" vector shuffle element is one which can be lowered to zero.
 /// Either it is an undef element in the shuffle mask, the element of the input
 /// referenced is undef, or the element of the input referenced is known to be
 /// zero. Many x86 shuffles can zero lanes cheaply and we often want to handle

 }
 
 // Attempt to match a combined shuffle mask against supported unary shuffle
 // instructions.
 // TODO: Investigate sharing more of this with shuffle lowering.
 static bool matchUnaryVectorShuffle(MVT SrcVT, ArrayRef<int> Mask,
                                     const X86Subtarget &Subtarget,
                                     unsigned &Shuffle, MVT &ShuffleVT) {
-  bool FloatDomain = SrcVT.isFloatingPoint();
+  bool FloatDomain = SrcVT.isFloatingPoint() ||
+                     (!Subtarget.hasAVX2() && SrcVT.is256BitVector());
 
   // Match a 128-bit integer vector against a VZEXT_MOVL (MOVQ) instruction.
   if (!FloatDomain && SrcVT.is128BitVector() &&
       isTargetShuffleEquivalent(Mask, {0, SM_SentinelZero})) {
     Shuffle = X86ISD::VZEXT_MOVL;
     ShuffleVT = MVT::v2i64;
     return true;
   }
 
if (isTargetShuffleEquivalent(
       ShuffleVT = MVT::v16f32;
       return true;
     }
   }
 
   return false;
 }
 
+// Attempt to match a combined shuffle mask against supported unary immediate
+// permute instructions.
+// TODO: Investigate sharing more of this with shuffle lowering.
+static bool matchPermuteVectorShuffle(MVT SrcVT, ArrayRef<int> Mask,
+                                      const X86Subtarget &Subtarget,
+                                      unsigned &Shuffle, MVT &ShuffleVT,
+                                      unsigned &PermuteImm) {
+  // Ensure we don't contain any zero elements.
+  for (int M : Mask) {
+    if (M == SM_SentinelZero)
+      return false;
+    assert(SM_SentinelUndef <= M && M < Mask.size() &&
+           "Expected unary shuffle");
+  }
+
+  // We only support permutation of 32/64 bit elements.
+  // TODO - support PSHUFLW/PSHUFHW.
+  unsigned MaskScalarSizeInBits = SrcVT.getSizeInBits() / Mask.size();
+  if (MaskScalarSizeInBits != 32 && MaskScalarSizeInBits != 64)
+    return false;
+  MVT MaskEltVT = MVT::getIntegerVT(MaskScalarSizeInBits);
+
+  // AVX introduced the VPERMILPD/VPERMILPS float permutes, before then we
+  // had to use 2-input SHUFPD/SHUFPS shuffles (not handled here).
+  bool FloatDomain = SrcVT.isFloatingPoint();
+  if (FloatDomain && !Subtarget.hasAVX())
+    return false;
+
+  // Pre-AVX2 we must use float shuffles on 256-bit vectors.
+  if (SrcVT.is256BitVector() && !Subtarget.hasAVX2())
+    FloatDomain = true;
+
+  // TODO - support LaneCrossing for AVX2 PERMQ/PERMPD
+  if (is128BitLaneCrossingShuffleMask(MaskEltVT, Mask))
+    return false;
+
+  // VPERMILPD can permute with a non-repeating shuffle.
+  if (FloatDomain && MaskScalarSizeInBits == 64) {
+    Shuffle = X86ISD::VPERMILPI;
+    ShuffleVT = MVT::getVectorVT(MVT::f64, Mask.size());
+    PermuteImm = 0;
+    for (int i = 0, e = Mask.size(); i != e; ++i) {
+      int M = Mask[i];
+      if (M == SM_SentinelUndef)
+        continue;
+      assert(((M / 2) == (i / 2)) && "Out of range shuffle mask index");
+      PermuteImm |= (M & 1) << i;
+    }
+    return true;
+  }
+
+  // We need a repeating shuffle mask for VPERMILPS/PSHUFD.
+  SmallVector<int, 4> RepeatedMask;
+  if (!is128BitLaneRepeatedShuffleMask(MaskEltVT, Mask, RepeatedMask))
+    return false;
+
+  // Narrow the repeated mask for 32-bit element permutes.
+  SmallVector<int, 4> WordMask = RepeatedMask;
+  if (MaskScalarSizeInBits == 64) {
+    WordMask.clear();
+    for (int M : Mask) {
+      if (M == SM_SentinelUndef) {
+        WordMask.append(2, SM_SentinelUndef);
+        continue;
+      }
+      WordMask.push_back((M * 2) + 0);
+      WordMask.push_back((M * 2) + 1);
+    }
+  }
+
+  Shuffle = (FloatDomain ? X86ISD::VPERMILPI : X86ISD::PSHUFD);
+  ShuffleVT = (FloatDomain ? MVT::f32 : MVT::i32);
+  ShuffleVT = MVT::getVectorVT(ShuffleVT, SrcVT.getSizeInBits() / 32);
+  PermuteImm = getV4X86ShuffleImm(WordMask);
+  return true;
+}
+
 // Attempt to match a combined unary shuffle mask against supported binary
 // shuffle instructions.
 // TODO: Investigate sharing more of this with shuffle lowering.
 static bool matchBinaryVectorShuffle(MVT SrcVT, ArrayRef<int> Mask,
                                      unsigned &Shuffle, MVT &ShuffleVT) {
   bool FloatDomain = SrcVT.isFloatingPoint();
 
   if (SrcVT.is128BitVector()) {
static bool combineX86ShuffleChain(SDValue Input, SDValue Root,
   if (RootVT.getScalarSizeInBits() != MaskEltSizeInBits &&
       (RootSizeInBits == 512 ||
        (Subtarget.hasVLX() && RootSizeInBits >= 128))) {
     return false;
   }
 
   // Attempt to match the mask against known shuffle patterns.
   MVT ShuffleVT;
-  unsigned Shuffle;
+  unsigned Shuffle, PermuteImm;
 
   if (matchUnaryVectorShuffle(VT, Mask, Subtarget, Shuffle, ShuffleVT)) {
     if (Depth == 1 && Root.getOpcode() == Shuffle)
       return false; // Nothing to do!
     Res = DAG.getBitcast(ShuffleVT, Input);
     DCI.AddToWorklist(Res.getNode());
     Res = DAG.getNode(Shuffle, DL, ShuffleVT, Res);
     DCI.AddToWorklist(Res.getNode());
     DCI.CombineTo(Root.getNode(), DAG.getBitcast(RootVT, Res),
                   /*AddTo*/ true);
     return true;
   }
 
+  if (matchPermuteVectorShuffle(VT, Mask, Subtarget, Shuffle, ShuffleVT,
+                                PermuteImm)) {
+    if (Depth == 1 && Root.getOpcode() == Shuffle)
+      return false; // Nothing to do!
+    Res = DAG.getBitcast(ShuffleVT, Input);
+    DCI.AddToWorklist(Res.getNode());
+    Res = DAG.getNode(Shuffle, DL, ShuffleVT, Res,
+                      DAG.getConstant(PermuteImm, DL, MVT::i8));
+    DCI.AddToWorklist(Res.getNode());
+    DCI.CombineTo(Root.getNode(), DAG.getBitcast(RootVT, Res),
+                  /*AddTo*/ true);
+    return true;
+  }
+
   if (matchBinaryVectorShuffle(VT, Mask, Shuffle, ShuffleVT)) {
     if (Depth == 1 && Root.getOpcode() == Shuffle)
       return false; // Nothing to do!
     Res = DAG.getBitcast(ShuffleVT, Input);
     DCI.AddToWorklist(Res.getNode());
     Res = DAG.getNode(Shuffle, DL, ShuffleVT, Res, Res);
     DCI.AddToWorklist(Res.getNode());
     DCI.CombineTo(Root.getNode(), DAG.getBitcast(RootVT, Res),

test/CodeGen/X86/2012-01-12-extract-sv.ll

 ; RUN: llc < %s -mattr=+avx -mtriple=i686-pc-win32 | FileCheck %s
 
 define void @endless_loop() {
 ; CHECK-LABEL: endless_loop:
 ; CHECK-NEXT:  # BB#0:
 ; CHECK-NEXT:    vmovaps (%eax), %ymm0
 ; CHECK-NEXT:    vextractf128 $1, %ymm0, %xmm0
 ; CHECK-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,1,1]
-; CHECK-NEXT:    vmovddup {{.*#+}} xmm1 = xmm0[0,0]
+; CHECK-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,1,0,1]
 ; CHECK-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm1
 ; CHECK-NEXT:    vxorps %ymm2, %ymm2, %ymm2
 ; CHECK-NEXT:    vblendps {{.*#+}} ymm1 = ymm2[0,1,2,3,4,5,6],ymm1[7]
 ; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm2[1,2,3,4,5,6,7]
 ; CHECK-NEXT:    vmovaps %ymm0, (%eax)
 ; CHECK-NEXT:    vmovaps %ymm1, (%eax)
 ; CHECK-NEXT:    vzeroupper
 ; CHECK-NEXT:    retl
 entry:
   %0 = load <8 x i32>, <8 x i32> addrspace(1)* undef, align 32
   %1 = shufflevector <8 x i32> %0, <8 x i32> undef, <16 x i32> <i32 4, i32 4, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
   %2 = shufflevector <16 x i32> <i32 undef, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 undef>, <16 x i32> %1, <16 x i32> <i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 17>
   store <16 x i32> %2, <16 x i32> addrspace(1)* undef, align 64
   ret void
 }

test/CodeGen/X86/2012-04-26-sdglue.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 | FileCheck %s
 
 ; rdar://11314175: SD Scheduler, BuildSchedUnits assert:
 ;                  N->getNodeId() == -1 && "Node already inserted!
 
 define void @func() nounwind ssp {
 ; CHECK-LABEL: func:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    vmovups 0, %xmm0
 ; CHECK-NEXT:    vxorps %ymm1, %ymm1, %ymm1
 ; CHECK-NEXT:    vblendps {{.*#+}} ymm2 = ymm0[0,1,2,3],ymm1[4,5,6,7]
-; CHECK-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,2,3,3]
-; CHECK-NEXT:    vpbroadcastd 32, %xmm3
-; CHECK-NEXT:    vinserti128 $1, %xmm3, %ymm0, %ymm0
+; CHECK-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[1,2,3,3]
+; CHECK-NEXT:    vbroadcastss 32, %xmm3
+; CHECK-NEXT:    vinsertf128 $1, %xmm3, %ymm0, %ymm0
 ; CHECK-NEXT:    vmulps %ymm0, %ymm2, %ymm2
 ; CHECK-NEXT:    vmulps %ymm0, %ymm0, %ymm0
 ; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
 ; CHECK-NEXT:    vaddps %ymm0, %ymm0, %ymm0
 ; CHECK-NEXT:    vmulps %xmm0, %xmm0, %xmm0
 ; CHECK-NEXT:    vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
 ; CHECK-NEXT:    vaddps %ymm0, %ymm0, %ymm0
 ; CHECK-NEXT:    vhaddps %ymm0, %ymm0, %ymm0

test/CodeGen/X86/avx-intrinsics-fast-isel.ll

; X64-NEXT:    retq
   %res = bitcast <16 x i16> %res15 to <4 x i64>
   ret <4 x i64> %res
 }
 
 define <4 x i64> @test_mm256_set1_epi32(i32 %a0) nounwind {
 ; X32-LABEL: test_mm256_set1_epi32:
 ; X32:       # BB#0:
 ; X32-NEXT:    vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
 ; X32-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X32-NEXT:    retl
 ;
 ; X64-LABEL: test_mm256_set1_epi32:
 ; X64:       # BB#0:
 ; X64-NEXT:    vmovd %edi, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
 ; X64-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X64-NEXT:    retq
   %res0 = insertelement <8 x i32> undef, i32 %a0, i32 0
   %res1 = insertelement <8 x i32> %res0, i32 %a0, i32 1
   %res2 = insertelement <8 x i32> %res1, i32 %a0, i32 2
   %res3 = insertelement <8 x i32> %res2, i32 %a0, i32 3
   %res4 = insertelement <8 x i32> %res3, i32 %a0, i32 4
   %res5 = insertelement <8 x i32> %res4, i32 %a0, i32 5

 ; X32-NEXT:    vpinsrd $2, %ecx, %xmm0, %xmm0
 ; X32-NEXT:    vpinsrd $3, %eax, %xmm0, %xmm0
 ; X32-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X32-NEXT:    retl
 ;
 ; X64-LABEL: test_mm256_set1_epi64x:
 ; X64:       # BB#0:
 ; X64-NEXT:    vmovq %rdi, %xmm0
-; X64-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
 ; X64-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X64-NEXT:    retq
   %res0 = insertelement <4 x i64> undef, i64 %a0, i32 0
   %res1 = insertelement <4 x i64> %res0, i64 %a0, i32 1
   %res2 = insertelement <4 x i64> %res1, i64 %a0, i32 2
   %res3 = insertelement <4 x i64> %res2, i64 %a0, i32 3
   ret <4 x i64> %res3
 }

test/CodeGen/X86/avx-intrinsics-x86.ll

 define <4 x double> @test_x86_avx_vpermilvar_pd_256_2(<4 x double> %a0) {
 ; AVX-LABEL: test_x86_avx_vpermilvar_pd_256_2:
 ; AVX:       ## BB#0:
 ; AVX-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3]
 ; AVX-NEXT:    retl
 ;
 ; AVX512VL-LABEL: test_x86_avx_vpermilvar_pd_256_2:
 ; AVX512VL:       ## BB#0:
-; AVX512VL-NEXT:    vpermilpd LCPI227_0, %ymm0, %ymm0
+; AVX512VL-NEXT:    vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3]
 ; AVX512VL-NEXT:    retl
   %res = call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> <i64 2, i64 0, i64 0, i64 2>) ; <<4 x double>> [#uses=1]
   ret <4 x double> %res
 }
 
 define <4 x float> @test_x86_avx_vpermilvar_ps(<4 x float> %a0, <4 x i32> %a1) {
 ; AVX-LABEL: test_x86_avx_vpermilvar_ps:
 ; AVX:       ## BB#0:

test/CodeGen/X86/avx-splat.ll

entry:
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> undef, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
   ret <16 x i16> %shuffle
 }
 
 define <4 x i64> @funcC(i64 %q) nounwind uwtable readnone ssp {
 ; CHECK-LABEL: funcC:
 ; CHECK:       ## BB#0: ## %entry
 ; CHECK-NEXT:    vmovq %rdi, %xmm0
-; CHECK-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
+; CHECK-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
 ; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
 entry:
   %vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
   %vecinit2.i = insertelement <4 x i64> %vecinit.i, i64 %q, i32 1
   %vecinit4.i = insertelement <4 x i64> %vecinit2.i, i64 %q, i32 2
   %vecinit6.i = insertelement <4 x i64> %vecinit4.i, i64 %q, i32 3
   ret <4 x i64> %vecinit6.i

test/CodeGen/X86/avx-vbroadcast.ll

entry:
   %ret = shufflevector <4 x i32> %ld, <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
   ret <4 x i32> %ret
 }
 
 define <8 x i32> @load_splat_8i32_4i32_33333333(<4 x i32>* %ptr) nounwind uwtable readnone ssp {
 ; X32-LABEL: load_splat_8i32_4i32_33333333:
 ; X32:       ## BB#0: ## %entry
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT:    vpermilps {{.*#+}} xmm0 = mem[3,3,3,3]
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,3,3,3]
 ; X32-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X32-NEXT:    retl
 ;
 ; X64-LABEL: load_splat_8i32_4i32_33333333:
 ; X64:       ## BB#0: ## %entry
-; X64-NEXT:    vpermilps {{.*#+}} xmm0 = mem[3,3,3,3]
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = mem[3,3,3,3]
 ; X64-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X64-NEXT:    retq
 entry:
   %ld = load <4 x i32>, <4 x i32>* %ptr
   %ret = shufflevector <4 x i32> %ld, <4 x i32> undef, <8 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
   ret <8 x i32> %ret
 }
 
entry:
   %ret = shufflevector <2 x i64> %ld, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
   ret <2 x i64> %ret
 }
 
 define <4 x i64> @load_splat_4i64_2i64_1111(<2 x i64>* %ptr) nounwind uwtable readnone ssp {
 ; X32-LABEL: load_splat_4i64_2i64_1111:
 ; X32:       ## BB#0: ## %entry
 ; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT:    vmovaps (%eax), %xmm0
-; X32-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = mem[2,3,2,3]
 ; X32-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X32-NEXT:    retl
 ;
 ; X64-LABEL: load_splat_4i64_2i64_1111:
 ; X64:       ## BB#0: ## %entry
-; X64-NEXT:    vmovaps (%rdi), %xmm0
-; X64-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = mem[2,3,2,3]
 ; X64-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; X64-NEXT:    retq
 entry:
   %ld = load <2 x i64>, <2 x i64>* %ptr
   %ret = shufflevector <2 x i64> %ld, <2 x i64> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
   ret <4 x i64> %ret
 }
 

test/CodeGen/X86/merge-consecutive-loads-128.ll

 ; SSE-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0,0,1,1]
 ; SSE-NEXT:    unpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: merge_4f32_f32_X0YY:
 ; AVX:       # BB#0:
 ; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; AVX-NEXT:    vunpcklps {{.*#+}} xmm1 = xmm1[0,0,1,1]
+; AVX-NEXT:    vpermilps {{.*#+}} xmm1 = xmm1[0,0,1,1]
 ; AVX-NEXT:    vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX-NEXT:    retq
 ;
 ; X32-SSE-LABEL: merge_4f32_f32_X0YY:
 ; X32-SSE:       # BB#0:
 ; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X32-SSE-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X32-SSE-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero

test/CodeGen/X86/pshufb-mask-comments.ll

; CHECK-NEXT:    retq
   ret <16 x i8> %1
 }
 
 ; Test that we won't crash when the constant was reused for another instruction.
 
 define <16 x i8> @test4(<16 x i8> %V, <2 x i64>* %P) {
 ; CHECK-LABEL: test4:
 ; CHECK:       # BB#0:
-; CHECK-NEXT:    movdqa {{.*#+}} xmm1 = [1084818905618843912,506097522914230528]
-; CHECK-NEXT:    movdqa %xmm1, (%rdi)
-; CHECK-NEXT:    pshufb %xmm1, %xmm0
+; CHECK-NEXT:    movaps {{.*#+}} xmm1 = [1084818905618843912,506097522914230528]
+; CHECK-NEXT:    movaps %xmm1, (%rdi)
+; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
 ; CHECK-NEXT:    retq
   %1 = insertelement <2 x i64> undef, i64 1084818905618843912, i32 0
   %2 = insertelement <2 x i64>    %1, i64  506097522914230528, i32 1
   store <2 x i64> %2, <2 x i64>* %P, align 16
   %3 = bitcast <2 x i64> %2 to <16 x i8>
   %4 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %V, <16 x i8> %3)
   ret <16 x i8> %4
 }

test/CodeGen/X86/sse3.ll

; X64-NEXT:    retq
   store <2 x i64> %tmp15.upgrd.2, <2 x i64>* %res
   ret void
 }
 
 define void @t9(<4 x float>* %r, <2 x i32>* %A) nounwind {
 ; X64-LABEL: t9:
 ; X64:       ## BB#0:
 ; X64-NEXT:    movapd (%rdi), %xmm0
-; X64-NEXT:    movhpd (%rsi), %xmm0
+; X64-NEXT:    movhpd {{.*#+}} xmm0 = xmm0[0],mem[0]
 ; X64-NEXT:    movapd %xmm0, (%rdi)
 ; X64-NEXT:    retq
   %tmp = load <4 x float>, <4 x float>* %r
   %tmp.upgrd.3 = bitcast <2 x i32>* %A to double*
   %tmp.upgrd.4 = load double, double* %tmp.upgrd.3
   %tmp.upgrd.5 = insertelement <2 x double> undef, double %tmp.upgrd.4, i32 0
   %tmp5 = insertelement <2 x double> %tmp.upgrd.5, double undef, i32 1
   %tmp6 = bitcast <2 x double> %tmp5 to <4 x float>

 
 }
 
 define <8 x i16> @t12(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
 ; X64-LABEL: t12:
 ; X64:       ## BB#0: ## %entry
 ; X64-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
 ; X64-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; X64-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,7,6,7]
+; X64-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,3,3]
 ; X64-NEXT:    retq
 entry:
   %tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 0, i32 1, i32 undef, i32 undef, i32 3, i32 11, i32 undef , i32 undef >
   ret <8 x i16> %tmp9
 
 }
 
 define <8 x i16> @t13(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
 ; X64-LABEL: t13:
 ; X64:       ## BB#0: ## %entry
 ; X64-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3]
 ; X64-NEXT:    pshuflw {{.*#+}} xmm0 = xmm1[0,2,2,3,4,5,6,7]
-; X64-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,6,7,6,7]
+; X64-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,3,3]
 ; X64-NEXT:    retq
 entry:
   %tmp9 = shufflevector <8 x i16> %T0, <8 x i16> %T1, <8 x i32> < i32 8, i32 9, i32 undef, i32 undef, i32 11, i32 3, i32 undef , i32 undef >
   ret <8 x i16> %tmp9
 }
 
 define <8 x i16> @t14(<8 x i16> %T0, <8 x i16> %T1) nounwind readnone {
 ; X64-LABEL: t14:

test/CodeGen/X86/vector-compare-results.ll

-; NOTE: Assertions have been autogenerated by update_llc_test_checks.py
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE42
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX --check-prefix=AVX512 --check-prefix=AVX512F
 
 ;

 ; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
 ; AVX1-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: test_cmp_v4f64:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vcmpltpd %ymm0, %ymm1, %ymm0
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,3,2,3]
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
 ; AVX512-LABEL: test_cmp_v4f64:
 ; AVX512:       # BB#0:
 ; AVX512-NEXT:    vcmpltpd %ymm0, %ymm1, %ymm0
 ; AVX512-NEXT:    vpmovqd %zmm0, %ymm0

 ; AVX1-NEXT:    vpackusdw %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackusdw %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: test_cmp_v8f64:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vcmpltpd %ymm0, %ymm2, %ymm0
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,3,2,3]
 ; AVX2-NEXT:    vcmpltpd %ymm1, %ymm3, %ymm1
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm1 = ymm1[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,3,2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,1,4,5,8,9,12,13],zero,zero,zero,zero,zero,zero,zero,zero,ymm0[16,17,20,21,24,25,28,29],zero,zero,zero,zero,zero,zero,zero,zero
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
 ; AVX512-LABEL: test_cmp_v8f64:

 ; AVX1-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
 ; AVX1-NEXT:    vpackuswb %xmm2, %xmm0, %xmm0
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: test_cmp_v16f64:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vcmpltpd %ymm2, %ymm6, %ymm2
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm2 = ymm2[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm2 = ymm2[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,3,2,3]
 ; AVX2-NEXT:    vcmpltpd %ymm3, %ymm7, %ymm3
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm3 = ymm3[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm3 = ymm3[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,3,2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm3, %ymm2, %ymm2
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [0,1,4,5,8,9,12,13,128,128,128,128,128,128,128,128,0,1,4,5,8,9,12,13,128,128,128,128,128,128,128,128]
 ; AVX2-NEXT:    vpshufb %ymm3, %ymm2, %ymm2
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,2,2,3]
 ; AVX2-NEXT:    vmovdqa {{.*#+}} xmm6 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
 ; AVX2-NEXT:    vpshufb %xmm6, %xmm2, %xmm2
 ; AVX2-NEXT:    vcmpltpd %ymm0, %ymm4, %ymm0
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,3,2,3]
 ; AVX2-NEXT:    vcmpltpd %ymm1, %ymm5, %ymm1
-; AVX2-NEXT:    vpshufd {{.*#+}} ymm1 = ymm1[0,2,0,2,4,6,4,6]
+; AVX2-NEXT:    vpermilps {{.*#+}} ymm1 = ymm1[0,2,0,2,4,6,4,6]
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,3,2,3]
 ; AVX2-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
 ; AVX2-NEXT:    vpshufb %ymm3, %ymm0, %ymm0
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
 ; AVX2-NEXT:    vpshufb %xmm6, %xmm0, %xmm0
 ; AVX2-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq

test/CodeGen/X86/vector-shuffle-128-v16.ll

; AVX-NEXT:    retq
   ret <16 x i8> %shuffle
 }
 
 define <16 x i8> @shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu(<16 x i8> %a) {
 ; SSE2-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
 ; SSE2:       # BB#0:
 ; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
 ; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,1,3]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:
 ; SSSE3:       # BB#0:
 ; SSSE3-NEXT:    pshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: shuffle_v16i8_00_uu_uu_uu_uu_uu_uu_uu_01_uu_uu_uu_uu_uu_uu_uu:

test/CodeGen/X86/vector-shuffle-128-v2.ll

-; NOTE: Assertions have been autogenerated by update_llc_test_checks.py
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE3
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+ssse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSSE3
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=knl -mattr=+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX512VL
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"

 define <2 x double> @shuffle_v2f64_11(<2 x double> %a, <2 x double> %b) {
 ; SSE-LABEL: shuffle_v2f64_11:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: shuffle_v2f64_11:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,1]
 ; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 1, i32 1>
   ret <2 x double> %shuffle
 }
 define <2 x double> @shuffle_v2f64_22(<2 x double> %a, <2 x double> %b) {
 ; SSE2-LABEL: shuffle_v2f64_22:
 ; SSE2:       # BB#0:
 ; SSE2-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0,0]

 ; SSE-LABEL: shuffle_v2f64_33:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    movhlps {{.*#+}} xmm1 = xmm1[1,1]
 ; SSE-NEXT:    movaps %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: shuffle_v2f64_33:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm1[1,1]
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm1[1,1]
 ; AVX-NEXT:    retq
   %shuffle = shufflevector <2 x double> %a, <2 x double> %b, <2 x i32> <i32 3, i32 3>
   ret <2 x double> %shuffle
 }
 define <2 x double> @shuffle_v2f64_03(<2 x double> %a, <2 x double> %b) {
 ; SSE2-LABEL: shuffle_v2f64_03:
 ; SSE2:       # BB#0:
 ; SSE2-NEXT:    movsd {{.*#+}} xmm1 = xmm0[0],xmm1[1]

test/CodeGen/X86/vector-shuffle-128-v4.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE3
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+ssse3 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSSE3
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE41
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64-unknown-unknown"
 

 define <4 x float> @shuffle_v4f32_0011(<4 x float> %a, <4 x float> %b) {
 ; SSE-LABEL: shuffle_v4f32_0011:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0,0,1,1]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: shuffle_v4f32_0011:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
 ; AVX-NEXT:    retq
   %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 0, i32 0, i32 1, i32 1>
   ret <4 x float> %shuffle
 }
 define <4 x float> @shuffle_v4f32_2233(<4 x float> %a, <4 x float> %b) {
 ; SSE-LABEL: shuffle_v4f32_2233:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    unpckhps {{.*#+}} xmm0 = xmm0[2,2,3,3]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: shuffle_v4f32_2233:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2,2,3,3]
+; AVX-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
 ; AVX-NEXT:    retq
   %shuffle = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 2, i32 2, i32 3, i32 3>
   ret <4 x float> %shuffle
 }
 define <4 x float> @shuffle_v4f32_0022(<4 x float> %a, <4 x float> %b) {
 ; SSE2-LABEL: shuffle_v4f32_0022:
 ; SSE2:       # BB#0:
 ; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,0,2,2]

test/CodeGen/X86/vector-shuffle-256-v16.ll

; AVX2-NEXT:    retq
   ret <16 x i16> %shuffle
 }
 
 define <16 x i16> @shuffle_v16i16_02_03_zz_zz_06_07_zz_zz_10_11_zz_zz_14_15_zz_zz(<16 x i16> %a) {
 ; AVX1-LABEL: shuffle_v16i16_02_03_zz_zz_06_07_zz_zz_10_11_zz_zz_14_15_zz_zz:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vxorps %ymm1, %ymm1, %ymm1
 ; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[1,3],ymm1[1,3],ymm0[5,7],ymm1[5,7]
-; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
+; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v16i16_02_03_zz_zz_06_07_zz_zz_10_11_zz_zz_14_15_zz_zz:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpsrlq $32, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <16 x i16> %a, <16 x i16> zeroinitializer, <16 x i32> <i32 2, i32 3, i32 16, i32 16, i32 6, i32 7, i32 16, i32 16, i32 10, i32 11, i32 16, i32 16, i32 14, i32 15, i32 16, i32 16>
   ret <16 x i16> %shuffle

test/CodeGen/X86/vector-shuffle-256-v4.ll

 ; ALL-NEXT:    retq
   %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 5, i32 undef, i32 undef>
   ret <4 x double> %shuffle
 }
 
 define <4 x double> @shuffle_v4f64_11uu(<4 x double> %a, <4 x double> %b) {
 ; ALL-LABEL: shuffle_v4f64_11uu:
 ; ALL:       # BB#0:
-; ALL-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; ALL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,1]
 ; ALL-NEXT:    retq
   %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 1, i32 undef, i32 undef>
   ret <4 x double> %shuffle
 }
 
 define <4 x double> @shuffle_v4f64_22uu(<4 x double> %a, <4 x double> %b) {
 ; AVX1-LABEL: shuffle_v4f64_22uu:
 ; AVX1:       # BB#0:

 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
   ret <4 x i64> %shuffle
 }
 
 define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {
 ; AVX1-LABEL: shuffle_v4i64_0001:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm1 = xmm0[0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,1,0,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v4i64_0001:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
 ; AVX2-NEXT:    retq
 ;
 ; AVX512VL-LABEL: shuffle_v4i64_0001:
 ; AVX512VL:       # BB#0:
 ; AVX512VL-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
   ret <4 x i64> %shuffle
 }
 
 define <4 x i64> @shuffle_v4i64_0020(<4 x i64> %a, <4 x i64> %b) {
 ; AVX1-LABEL: shuffle_v4i64_0020:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; AVX1-NEXT:    vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v4i64_0020:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
 ; AVX2-NEXT:    retq
 ;

 ; AVX512VL-NEXT:    retq
   %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
   ret <4 x i64> %shuffle
 }
 
 define <4 x i64> @shuffle_v4i64_1000(<4 x i64> %a, <4 x i64> %b) {
 ; AVX1-LABEL: shuffle_v4i64_1000:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[2,3,0,1]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v4i64_1000:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
 ; AVX2-NEXT:    retq
 ;
; AVX512VL-NEXT:    retq
   ret <4 x i64> %shuffle
 }
 
 define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) {
 ; AVX1-LABEL: shuffle_v4i64_4012:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
 ; AVX1-NEXT:    vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v4i64_4012:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,2]
 ; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5,6,7]

 ; AVX512VL-NEXT:    retq
   %1 = shufflevector <2 x double> %r, <2 x double> undef, <4 x i32> zeroinitializer
   ret <4 x double> %1
 }
 
 define <4 x i64> @splat_mem_v4i64_from_v2i64(<2 x i64>* %ptr) {
 ; AVX1-LABEL: splat_mem_v4i64_from_v2i64:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = mem[0,1,0,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: splat_mem_v4i64_from_v2i64:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vbroadcastsd (%rdi), %ymm0
 ; AVX2-NEXT:    retq
 ;

test/CodeGen/X86/vector-shuffle-256-v8.ll

 ; ALL-NEXT:    retq
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>
   ret <8 x float> %shuffle
 }
 
 define <8 x float> @shuffle_v8f32_00112233(<8 x float> %a, <8 x float> %b) {
 ; AVX1-LABEL: shuffle_v8f32_00112233:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vunpcklps {{.*#+}} xmm1 = xmm0[0,0,1,1]
-; AVX1-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2,2,3,3]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,1,1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8f32_00112233:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovaps {{.*#+}} ymm1 = [0,0,1,1,2,2,3,3]
 ; AVX2-NEXT:    vpermps %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
; ALL-NEXT:    retq
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 8, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 7>
   ret <8 x float> %shuffle
 }
 
 define <8 x float> @shuffle_v8f32_08080808(<8 x float> %a, <8 x float> %b) {
 ; AVX1-LABEL: shuffle_v8f32_08080808:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[0,0]
-; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,2,1,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8f32_08080808:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0],xmm1[0],zero,zero
 ; AVX2-NEXT:    vbroadcastsd %xmm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 0, i32 8, i32 0, i32 8, i32 0, i32 8>
   ret <8 x float> %shuffle
 }
 
 define <8 x float> @shuffle_v8f32_08084c4c(<8 x float> %a, <8 x float> %b) {
 ; ALL-LABEL: shuffle_v8f32_08084c4c:
 ; ALL:       # BB#0:
 ; ALL-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,0],ymm1[0,0],ymm0[4,4],ymm1[4,4]
-; ALL-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
+; ALL-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
 ; ALL-NEXT:    retq
   %shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 0, i32 8, i32 4, i32 12, i32 4, i32 12>
   ret <8 x float> %shuffle
 }
 
 define <8 x float> @shuffle_v8f32_8823cc67(<8 x float> %a, <8 x float> %b) {
 ; ALL-LABEL: shuffle_v8f32_8823cc67:
 ; ALL:       # BB#0:

 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_00000010(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_00000010:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,1,0]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_00000010:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,1,0]
 ; AVX2-NEXT:    vpermd %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_00000200(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_00000200:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,2,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,2,0,0]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_00000200:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,2,0,0]
 ; AVX2-NEXT:    vpermd %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_00003000(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_00003000:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[3,0,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[3,0,0,0]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_00003000:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,0,0,0,3,0,0,0]
 ; AVX2-NEXT:    vpermd %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq

 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 5, i32 4, i32 5>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_00112233(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_00112233:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,1,1]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,0,1,1]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_00112233:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,0,1,1,2,2,3,3]
 ; AVX2-NEXT:    vpermd %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 1, i32 1, i32 2, i32 2, i32 3, i32 3>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_00001111(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_00001111:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[0,0,0,0]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,1,1]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_00001111:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm1 = [0,0,0,0,1,1,1,1]
 ; AVX2-NEXT:    vpermd %ymm0, %ymm1, %ymm0
 ; AVX2-NEXT:    retq
; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 8, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 7>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_08080808(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_08080808:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[0,0]
-; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,2,1,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_08080808:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
 ; AVX2-NEXT:    vpbroadcastq %xmm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 8, i32 0, i32 8, i32 0, i32 8, i32 0, i32 8>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_08084c4c(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_08084c4c:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,0],ymm1[0,0],ymm0[4,4],ymm1[4,4]
-; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
+; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_08084c4c:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpshufd {{.*#+}} ymm1 = ymm1[0,0,2,0,4,4,6,4]
 ; AVX2-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,1,0,1,4,5,4,5]
 ; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
 ; AVX2-NEXT:    retq

 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 8, i32 9, i32 9, i32 1, i32 10, i32 11, i32 11>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_091b2d3f(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_091b2d3f:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm2 = xmm0[0,1,1,3]
-; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,1,3,3]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[2,1,3,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm2, %ymm0
 ; AVX1-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_091b2d3f:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
 ; AVX2-NEXT:    vpblendd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3],ymm0[4],ymm1[5],ymm0[6],ymm1[7]
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 9, i32 1, i32 11, i32 2, i32 13, i32 3, i32 15>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_09ab1def(<8 x i32> %a, <8 x i32> %b) {
 ; AVX1-LABEL: shuffle_v8i32_09ab1def:
 ; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovshdup {{.*#+}} xmm2 = xmm0[1,1,3,3]
+; AVX1-NEXT:    vpshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_09ab1def:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vmovdqa {{.*#+}} ymm2 = <0,u,u,u,1,u,u,u>
 ; AVX2-NEXT:    vpermd %ymm0, %ymm2, %ymm0

 ; Shuffle to logical bit shifts
 ;
 
 define <8 x i32> @shuffle_v8i32_z0U2zUz6(<8 x i32> %a) {
 ; AVX1-LABEL: shuffle_v8i32_z0U2zUz6:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vxorps %ymm1, %ymm1, %ymm1
 ; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,2],ymm1[0,2],ymm0[4,6],ymm1[4,6]
-; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
+; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_z0U2zUz6:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpsllq $32, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> zeroinitializer, <8 x i32> <i32 8, i32 0, i32 undef, i32 2, i32 8, i32 undef, i32 8, i32 6>
   ret <8 x i32> %shuffle
 }
 
 define <8 x i32> @shuffle_v8i32_1U3z5zUU(<8 x i32> %a) {
 ; AVX1-LABEL: shuffle_v8i32_1U3z5zUU:
 ; AVX1:       # BB#0:
 ; AVX1-NEXT:    vxorps %ymm1, %ymm1, %ymm1
 ; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[1,3],ymm1[1,3],ymm0[5,7],ymm1[5,7]
-; AVX1-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
+; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-LABEL: shuffle_v8i32_1U3z5zUU:
 ; AVX2:       # BB#0:
 ; AVX2-NEXT:    vpsrlq $32, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
   %shuffle = shufflevector <8 x i32> %a, <8 x i32> zeroinitializer, <8 x i32> <i32 1, i32 undef, i32 3, i32 8, i32 5, i32 8, i32 undef, i32 undef>
   ret <8 x i32> %shuffle

test/CodeGen/X86/vector-shuffle-combining-avx.ll

 ; ALL-NEXT:    retq
   %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 0, i32 0, i32 2, i32 undef>)
   ret <4 x float> %1
 }
 
 define <4 x float> @combine_vpermilvar_4f32_unpckh(<4 x float> %a0) {
 ; ALL-LABEL: combine_vpermilvar_4f32_unpckh:
 ; ALL:       # BB#0:
-; ALL-NEXT:    vunpckhps {{.*#+}} xmm0 = xmm0[2,2,3,3]
+; ALL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,2,3,3]
 ; ALL-NEXT:    retq
   %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 2, i32 2, i32 3, i32 3>)
   ret <4 x float> %1
 }
 
 define <4 x float> @combine_vpermilvar_4f32_unpckl(<4 x float> %a0) {
 ; ALL-LABEL: combine_vpermilvar_4f32_unpckl:
 ; ALL:       # BB#0:
-; ALL-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; ALL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
 ; ALL-NEXT:    retq
   %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 0, i32 0, i32 1, i32 1>)
   ret <4 x float> %1
 }
 
 define <8 x float> @combine_vpermilvar_8f32_identity(<8 x float> %a0) {
 ; ALL-LABEL: combine_vpermilvar_8f32_identity:
 ; ALL:       # BB#0:

 ; ALL-NEXT:    retq
   %1 = tail call <4 x double> @llvm.x86.avx.vpermilvar.pd.256(<4 x double> %a0, <4 x i64> <i64 0, i64 0, i64 4, i64 4>)
   ret <4 x double> %1
 }
 
 define <4 x float> @combine_vpermilvar_4f32_4stage(<4 x float> %a0) {
 ; ALL-LABEL: combine_vpermilvar_4f32_4stage:
 ; ALL:       # BB#0:
-; ALL-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[8,9,10,11,0,1,2,3,12,13,14,15,4,5,6,7]
+; ALL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[2,0,3,1]
 ; ALL-NEXT:    retq
   %1 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float> %a0, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
   %2 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %1, <4 x i32> <i32 2, i32 3, i32 0, i32 1>)
   %3 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %2, <4 x i32> <i32 0, i32 2, i32 1, i32 3>)
   %4 = tail call <4 x float> @llvm.x86.avx.vpermilvar.ps(<4 x float>  %3, <4 x i32> <i32 3, i32 2, i32 1, i32 0>)
   ret <4 x float> %4
 }
 
 define <8 x float> @combine_vpermilvar_8f32_4stage(<8 x float> %a0) {
-; AVX1-LABEL: combine_vpermilvar_8f32_4stage:
-; AVX1:       # BB#0:
-; AVX1-NEXT:    vmovaps {{.*#+}} ymm1 = [3,2,1,0,3,2,1,0]
-; AVX1-NEXT:    vpermilps %ymm1, %ymm0, %ymm0
-; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
-; AVX1-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,2,1,3,4,6,5,7]
-; AVX1-NEXT:    vpermilps %ymm1, %ymm0, %ymm0
-; AVX1-NEXT:    retq
-;
-; AVX2-LABEL: combine_vpermilvar_8f32_4stage:
-; AVX2:       # BB#0:
-; AVX2-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[8,9,10,11,0,1,2,3,12,13,14,15,4,5,6,7,24,25,26,27,16,17,18,19,28,29,30,31,20,21,22,23]
-; AVX2-NEXT:    retq
-;
-; AVX512F-LABEL: combine_vpermilvar_8f32_4stage:
-; AVX512F:       # BB#0:
-; AVX512F-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[8,9,10,11,0,1,2,3,12,13,14,15,4,5,6,7,24,25,26,27,16,17,18,19,28,29,30,31,20,21,22,23]
-; AVX512F-NEXT:    retq
+; ALL-LABEL: combine_vpermilvar_8f32_4stage:
+; ALL:       # BB#0:
+; ALL-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,0,3,1,6,4,7,5]
+; ALL-NEXT:    retq
   %1 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float> %a0, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
   %2 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %1, <8 x i32> <i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1>)
   %3 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %2, <8 x i32> <i32 0, i32 2, i32 1, i32 3, i32 0, i32 2, i32 1, i32 3>)
   %4 = tail call <8 x float> @llvm.x86.avx.vpermilvar.ps.256(<8 x float>  %3, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>)
   ret <8 x float> %4
 }

test/CodeGen/X86/vector-shuffle-combining-avx2.ll

; CHECK-NEXT:    retq
   %tmp2 = shufflevector <32 x i8> %tmp1, <32 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 30>
   ret <32 x i8> %tmp2
 }
 
 define <4 x i64> @combine_permq_pshufb(<4 x i64> %a0) {
 ; CHECK-LABEL: combine_permq_pshufb:
 ; CHECK:       # BB#0:
 ; CHECK-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[3,2,1,0]
-; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[8,9,10,11,12,13,14,15,0,1,2,3,4,5,6,7,24,25,26,27,28,29,30,31,16,17,18,19,20,21,22,23]
+; CHECK-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
 ; CHECK-NEXT:    retq
   %1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
   %2 = bitcast <4 x i64> %1 to <32 x i8>
   %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
   %4 = bitcast <32 x i8> %3 to <4 x i64>
   ret <4 x i64> %4
 }

test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll

 ; CHECK-NEXT:    kmovw %esi, %k1
 ; CHECK-NEXT:    vmovsldup {{.*#+}} zmm0 = mem[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
 ; CHECK-NEXT:    retq
   %x0 = load <16 x float>, <16 x float> *%p0
   %res0 = call <16 x float> @llvm.x86.avx512.maskz.vpermt2var.ps.512(<16 x i32> <i32 undef, i32 0, i32 undef, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14>, <16 x float> %x0, <16 x float> %x1, i16 %m)
   ret <16 x float> %res0
 }
 
+define <16 x float> @combine_vpermt2var_16f32_vpermilps(<16 x float> %x0, <16 x float> %x1) {
+; CHECK-LABEL: combine_vpermt2var_16f32_vpermilps:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vpermilps {{.*#+}} zmm0 = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-NEXT:    retq
+  %res0 = call <16 x float> @llvm.x86.avx512.maskz.vpermt2var.ps.512(<16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>, <16 x float> %x0, <16 x float> %x1, i16 -1)
+  ret <16 x float> %res0
+}
+define <16 x float> @combine_vpermt2var_16f32_vpermilps_load(<16 x float> *%p0, <16 x float> %x1) {
+; CHECK-LABEL: combine_vpermt2var_16f32_vpermilps_load:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    vpermilps {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-NEXT:    retq
+  %x0 = load <16 x float>, <16 x float> *%p0
+  %res0 = call <16 x float> @llvm.x86.avx512.maskz.vpermt2var.ps.512(<16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>, <16 x float> %x0, <16 x float> %x1, i16 -1)
+  ret <16 x float> %res0
+}
+define <16 x float> @combine_vpermt2var_16f32_vpermilps_mask(<16 x float> %x0, <16 x float> %x1, i16 %m) {
+; CHECK-LABEL: combine_vpermt2var_16f32_vpermilps_mask:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    kmovw %edi, %k1
+; CHECK-NEXT:    vpermilps {{.*#+}} zmm0 = zmm0[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-NEXT:    retq
+  %res0 = call <16 x float> @llvm.x86.avx512.maskz.vpermt2var.ps.512(<16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>, <16 x float> %x0, <16 x float> %x1, i16 %m)
+  ret <16 x float> %res0
+}
+define <16 x float> @combine_vpermt2var_16f32_vpermilps_mask_load(<16 x float> *%p0, <16 x float> %x1, i16 %m) {
+; CHECK-LABEL: combine_vpermt2var_16f32_vpermilps_mask_load:
+; CHECK:       # BB#0:
+; CHECK-NEXT:    kmovw %esi, %k1
+; CHECK-NEXT:    vpermilps {{.*#+}} zmm0 = mem[3,2,1,0,7,6,5,4,11,10,9,8,15,14,13,12]
+; CHECK-NEXT:    retq
+  %x0 = load <16 x float>, <16 x float> *%p0
+  %res0 = call <16 x float> @llvm.x86.avx512.maskz.vpermt2var.ps.512(<16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4, i32 11, i32 10, i32 9, i32 8, i32 15, i32 14, i32 13, i32 12>, <16 x float> %x0, <16 x float> %x1, i16 %m)
+  ret <16 x float> %res0
+}
+
 define <16 x i32> @combine_vpermt2var_16i32_identity(<16 x i32> %x0, <16 x i32> %x1) {
 ; CHECK-LABEL: combine_vpermt2var_16i32_identity:
 ; CHECK:       # BB#0:
 ; CHECK-NEXT:    retq
   %res0 = call <16 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.512(<16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 undef>, <16 x i32> %x0, <16 x i32> %x1, i16 -1)
   %res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.512(<16 x i32> <i32 15, i32 30, i32 13, i32 28, i32 undef, i32 26, i32 9, i32 24, i32 7, i32 22, i32 5, i32 20, i32 3, i32 18, i32 1, i32 16>, <16 x i32> %res0, <16 x i32> %res0, i16 -1)
   ret <16 x i32> %res1
 }

test/CodeGen/X86/vector-shuffle-combining-ssse3.ll

; AVX-NEXT:    retq
   %3 = bitcast <16 x i8> %2 to <4 x float>
   %4 = shufflevector <4 x float> %3, <4 x float> undef, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
   ret <4 x float> %4
 }
 
 define <16 x i8> @combine_pshufb_palignr(<16 x i8> %a0, <16 x i8> %a1) {
 ; SSE-LABEL: combine_pshufb_palignr:
 ; SSE:       # BB#0:
-; SSE-NEXT:    pshufb {{.*#+}} xmm1 = xmm1[8,9,10,11,12,13,14,15,8,9,10,11,12,13,14,15]
-; SSE-NEXT:    movdqa %xmm1, %xmm0
+; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: combine_pshufb_palignr:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vpshufb {{.*#+}} xmm0 = xmm1[8,9,10,11,12,13,14,15,8,9,10,11,12,13,14,15]
+; AVX-NEXT:    vpshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
 ; AVX-NEXT:    retq
   %1 = shufflevector <16 x i8> %a0, <16 x i8> %a1, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
   %2 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %1, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
   ret <16 x i8> %2
 }
 
 define <16 x i8> @combine_pshufb_pslldq(<16 x i8> %a0) {
 ; SSE-LABEL: combine_pshufb_pslldq:

test/CodeGen/X86/vector-shuffle-combining.ll

 define <4 x float> @combine_undef_input_test9(<4 x float> %a) {
 ; SSE-LABEL: combine_undef_input_test9:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: combine_undef_input_test9:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,1]
 ; AVX-NEXT:    retq
   %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
   %2 = shufflevector <4 x float> %1, <4 x float> %a, <4 x i32> <i32 6, i32 7, i32 0, i32 1>
   ret <4 x float> %2
 }
 
 define <4 x float> @combine_undef_input_test10(<4 x float> %a) {
 ; ALL-LABEL: combine_undef_input_test10:

 define <4 x float> @combine_undef_input_test19(<4 x float> %a) {
 ; SSE-LABEL: combine_undef_input_test19:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: combine_undef_input_test19:
 ; AVX:       # BB#0:
-; AVX-NEXT:    vmovhlps {{.*#+}} xmm0 = xmm0[1,1]
+; AVX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,1]
 ; AVX-NEXT:    retq
   %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 5, i32 5>
   %2 = shufflevector <4 x float> %a, <4 x float> %1, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
   ret <4 x float> %2
 }
 
 define <4 x float> @combine_undef_input_test20(<4 x float> %a) {
 ; ALL-LABEL: combine_undef_input_test20: