[CostModel] Replace ShuffleKind::SK_Alternate with ShuffleKind::SK_Select (PR33744)

[RKSimon]

Summary

As discussed on PR33744, this patch replaces ShuffleKind::SK_Alternate which requires shuffle masks to only match an alternating pattern from its 2 sources:

e.g. v4f32: <0,5,2,7> or <4,1,6,3>

This seems far too restrictive as most SIMD hardware which will implement it using a general blend/bit-select instruction (permitting elements from either source as long as they are inline):

e.g. v4f32: <0,5,2,7>, <4,1,6,3>, <0,1,6,7>, <4,1,2,3> etc.

This initial patch just updates the name and cost model shuffle mask analysis, later patch reviews will update SLP to better utilise this - it still limits itself to SK_Alternate patterns.

It also raised an interesting issue in that Identity shuffles are being matched by analysis to a mixture of SK_PermuteSingleSrc/SK_Alternate - shouldn't we just return cost = 0 for these cases?

[RKSimon]

rebase

[spatel]

LGTM.

include/llvm/Analysis/TargetTransformInfo.h

644

A bit more description would be good here. Something like:
Selects elements from the corresponding lane of one source operand or the other. This is equivalent to a vector select with a constant condition operand.

lib/Analysis/TargetTransformInfo.cpp

678

Formatting: IsSelect.

include/llvm/Analysis/TargetTransformInfo.h

public:
 
   /// \name Vector Target Information
   /// @{
 
   /// The various kinds of shuffle patterns for vector queries.
   enum ShuffleKind {
     SK_Broadcast,       ///< Broadcast element 0 to all other elements.
     SK_Reverse,         ///< Reverse the order of the vector.
-    SK_Alternate,       ///< Choose alternate elements from vector.
+    SK_Select,          ///< Select either vectors' element from that lane.
     SK_Transpose,       ///< Transpose two vectors.
     SK_InsertSubvector, ///< InsertSubvector. Index indicates start offset.
     SK_ExtractSubvector,///< ExtractSubvector Index indicates start offset.
     SK_PermuteTwoSrc,   ///< Merge elements from two source vectors into one
                         ///< with any shuffle mask.
     SK_PermuteSingleSrc ///< Shuffle elements of single source vector with any
                         ///< shuffle mask.
   };

include/llvm/CodeGen/BasicTTIImpl.h

unsigned getArithmeticInstrCost(
 
     // We don't know anything about this scalar instruction.
     return OpCost;
   }
 
   unsigned getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
                           Type *SubTp) {
     switch (Kind) {
-    case TTI::SK_Alternate:
+    case TTI::SK_Select:
     case TTI::SK_Transpose:
     case TTI::SK_PermuteSingleSrc:
     case TTI::SK_PermuteTwoSrc:
       return getPermuteShuffleOverhead(Tp);
     default:
       return 1;
     }
   }

lib/Analysis/TargetTransformInfo.cpp

for (unsigned i = 0; i < MaskSize && (IdentityLHS || IdentityRHS); ++i) {
     if (Mask[i] < 0)
       continue;
     IdentityLHS &= (Mask[i] == i);
     IdentityRHS &= (Mask[i] == (i + MaskSize));
   }
   return IdentityLHS || IdentityRHS;
 }
 
-static bool isAlternateVectorMask(ArrayRef<int> Mask) {
-  bool isAlternate = true;
+static bool isSelectVectorMask(ArrayRef<int> Mask) {
+  bool isSelect = true;
+  bool FoundLHS = false;
+  bool FoundRHS = false;
   unsigned MaskSize = Mask.size();
 
-  // Example: shufflevector A, B, <0,5,2,7>
-  for (unsigned i = 0; i < MaskSize && isAlternate; ++i) {
-    if (Mask[i] < 0)
-      continue;
-    isAlternate = Mask[i] == (int)((i & 1) ? MaskSize + i : i);
-  }
-
-  if (isAlternate)
-    return true;
-
-  isAlternate = true;
+  // Example: shufflevector A, B, <0,1,6,3>
   // Example: shufflevector A, B, <4,1,6,3>
-  for (unsigned i = 0; i < MaskSize && isAlternate; ++i) {
+  for (unsigned i = 0; i < MaskSize && isSelect; ++i) {
     if (Mask[i] < 0)
       continue;
-    isAlternate = Mask[i] == (int)((i & 1) ? i : MaskSize + i);
+    FoundLHS |= (Mask[i] == i);
+    FoundRHS |= (Mask[i] == (i + MaskSize));
+    isSelect = (Mask[i] == i) || (Mask[i] == (i + MaskSize));
   }
-
-  return isAlternate;
+  // If we don't use both vectors this is really an Identity mask.
+  return isSelect && FoundLHS && FoundRHS;
 }
 
 static bool isTransposeVectorMask(ArrayRef<int> Mask) {
   // Transpose vector masks transpose a 2xn matrix. They read corresponding
   // even- or odd-numbered vector elements from two n-dimensional source
   // vectors and write each result into consecutive elements of an
   // n-dimensional destination vector. Two shuffles are necessary to complete
   // the transpose, one for the even elements and another for the odd elements.
case Instruction::ShuffleVector: {
     if (NumVecElems == Mask.size()) {
       if (isIdentityVectorMask(Mask))
         return 0;
 
       if (isReverseVectorMask(Mask))
         return TTIImpl->getShuffleCost(TargetTransformInfo::SK_Reverse,
                                        VecTypOp0, 0, nullptr);
 
-      if (isAlternateVectorMask(Mask))
-        return TTIImpl->getShuffleCost(TargetTransformInfo::SK_Alternate,
+      if (isSelectVectorMask(Mask))
+        return TTIImpl->getShuffleCost(TargetTransformInfo::SK_Select,
                                        VecTypOp0, 0, nullptr);
 
       if (isTransposeVectorMask(Mask))
         return TTIImpl->getShuffleCost(TargetTransformInfo::SK_Transpose,
                                        VecTypOp0, 0, nullptr);
 
       if (isZeroEltBroadcastVectorMask(Mask))
         return TTIImpl->getShuffleCost(TargetTransformInfo::SK_Broadcast,

lib/Target/ARM/ARMTargetTransformInfo.cpp

int ARMTTIImpl::getAddressComputationCost(Type *Ty, ScalarEvolution *SE,
 
   // In many cases the address computation is not merged into the instruction
   // addressing mode.
   return 1;
 }
 
 int ARMTTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
                                Type *SubTp) {
-  // We only handle costs of reverse and alternate shuffles for now.
-  if (Kind != TTI::SK_Reverse && Kind != TTI::SK_Alternate)
+  // We only handle costs of reverse and select shuffles for now.
+  if (Kind != TTI::SK_Reverse && Kind != TTI::SK_Select)
     return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
 
   if (Kind == TTI::SK_Reverse) {
     static const CostTblEntry NEONShuffleTbl[] = {
         // Reverse shuffle cost one instruction if we are shuffling within a
         // double word (vrev) or two if we shuffle a quad word (vrev, vext).
         {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1},
 
         {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2},
         {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2},
         {ISD::VECTOR_SHUFFLE, MVT::v8i16, 2},
         {ISD::VECTOR_SHUFFLE, MVT::v16i8, 2}};
 
     std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Tp);
 
     if (const auto *Entry = CostTableLookup(NEONShuffleTbl, ISD::VECTOR_SHUFFLE,
                                             LT.second))
       return LT.first * Entry->Cost;
 
     return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
   }
-  if (Kind == TTI::SK_Alternate) {
-    static const CostTblEntry NEONAltShuffleTbl[] = {
-        // Alt shuffle cost table for ARM. Cost is the number of instructions
+  if (Kind == TTI::SK_Select) {
+    static const CostTblEntry NEONSelShuffleTbl[] = {
+        // Select shuffle cost table for ARM. Cost is the number of instructions
         // required to create the shuffled vector.
 
         {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1},
         {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1},
 
         {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2},
         {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2},
         {ISD::VECTOR_SHUFFLE, MVT::v4i16, 2},
 
         {ISD::VECTOR_SHUFFLE, MVT::v8i16, 16},
 
         {ISD::VECTOR_SHUFFLE, MVT::v16i8, 32}};
 
     std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Tp);
-    if (const auto *Entry = CostTableLookup(NEONAltShuffleTbl,
+    if (const auto *Entry = CostTableLookup(NEONSelShuffleTbl,
                                             ISD::VECTOR_SHUFFLE, LT.second))
       return LT.first * Entry->Cost;
     return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
   }
   return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
 }
 
 int ARMTTIImpl::getArithmeticInstrCost(

lib/Target/X86/X86TargetTransformInfo.cpp

static const CostTblEntry AVX2ShuffleTbl[] = {
 
     { TTI::SK_Reverse,   MVT::v4f64,  1 }, // vpermpd
     { TTI::SK_Reverse,   MVT::v8f32,  1 }, // vpermps
     { TTI::SK_Reverse,   MVT::v4i64,  1 }, // vpermq
     { TTI::SK_Reverse,   MVT::v8i32,  1 }, // vpermd
     { TTI::SK_Reverse,   MVT::v16i16, 2 }, // vperm2i128 + pshufb
     { TTI::SK_Reverse,   MVT::v32i8,  2 }, // vperm2i128 + pshufb
 
-    { TTI::SK_Alternate, MVT::v16i16, 1 }, // vpblendw
-    { TTI::SK_Alternate, MVT::v32i8,  1 }, // vpblendvb
+    { TTI::SK_Select,    MVT::v16i16, 1 }, // vpblendvb
+    { TTI::SK_Select,    MVT::v32i8,  1 }, // vpblendvb
 
     { TTI::SK_PermuteSingleSrc, MVT::v4f64,  1 }, // vpermpd
     { TTI::SK_PermuteSingleSrc, MVT::v8f32,  1 }, // vpermps
     { TTI::SK_PermuteSingleSrc, MVT::v4i64,  1 }, // vpermq
     { TTI::SK_PermuteSingleSrc, MVT::v8i32,  1 }, // vpermd
     { TTI::SK_PermuteSingleSrc, MVT::v16i16, 4 }, // vperm2i128 + 2*vpshufb
                                                   // + vpblendvb
     { TTI::SK_PermuteSingleSrc, MVT::v32i8,  4 }, // vperm2i128 + 2*vpshufb
static const CostTblEntry AVX1ShuffleTbl[] = {
     { TTI::SK_Reverse,   MVT::v8f32,  2 }, // vperm2f128 + vpermilps
     { TTI::SK_Reverse,   MVT::v4i64,  2 }, // vperm2f128 + vpermilpd
     { TTI::SK_Reverse,   MVT::v8i32,  2 }, // vperm2f128 + vpermilps
     { TTI::SK_Reverse,   MVT::v16i16, 4 }, // vextractf128 + 2*pshufb
                                            // + vinsertf128
     { TTI::SK_Reverse,   MVT::v32i8,  4 }, // vextractf128 + 2*pshufb
                                            // + vinsertf128
 
-    { TTI::SK_Alternate, MVT::v4i64,  1 }, // vblendpd
-    { TTI::SK_Alternate, MVT::v4f64,  1 }, // vblendpd
-    { TTI::SK_Alternate, MVT::v8i32,  1 }, // vblendps
-    { TTI::SK_Alternate, MVT::v8f32,  1 }, // vblendps
-    { TTI::SK_Alternate, MVT::v16i16, 3 }, // vpand + vpandn + vpor
-    { TTI::SK_Alternate, MVT::v32i8,  3 }, // vpand + vpandn + vpor
+    { TTI::SK_Select,    MVT::v4i64,  1 }, // vblendpd
+    { TTI::SK_Select,    MVT::v4f64,  1 }, // vblendpd
+    { TTI::SK_Select,    MVT::v8i32,  1 }, // vblendps
+    { TTI::SK_Select,    MVT::v8f32,  1 }, // vblendps
+    { TTI::SK_Select,    MVT::v16i16, 3 }, // vpand + vpandn + vpor
+    { TTI::SK_Select,    MVT::v32i8,  3 }, // vpand + vpandn + vpor
 
     { TTI::SK_PermuteSingleSrc, MVT::v4f64,  3 }, // 2*vperm2f128 + vshufpd
     { TTI::SK_PermuteSingleSrc, MVT::v4i64,  3 }, // 2*vperm2f128 + vshufpd
     { TTI::SK_PermuteSingleSrc, MVT::v8f32,  4 }, // 2*vperm2f128 + 2*vshufps
     { TTI::SK_PermuteSingleSrc, MVT::v8i32,  4 }, // 2*vperm2f128 + 2*vshufps
     { TTI::SK_PermuteSingleSrc, MVT::v16i16, 8 }, // vextractf128 + 4*pshufb
                                                   // + 2*por + vinsertf128
     { TTI::SK_PermuteSingleSrc, MVT::v32i8,  8 }, // vextractf128 + 4*pshufb
static const CostTblEntry AVX1ShuffleTbl[] = {
                                                    // + 4*por + vinsertf128
   };
 
   if (ST->hasAVX())
     if (const auto *Entry = CostTableLookup(AVX1ShuffleTbl, Kind, LT.second))
       return LT.first * Entry->Cost;
 
   static const CostTblEntry SSE41ShuffleTbl[] = {
-    { TTI::SK_Alternate, MVT::v2i64,  1 }, // pblendw
-    { TTI::SK_Alternate, MVT::v2f64,  1 }, // movsd
-    { TTI::SK_Alternate, MVT::v4i32,  1 }, // pblendw
-    { TTI::SK_Alternate, MVT::v4f32,  1 }, // blendps
-    { TTI::SK_Alternate, MVT::v8i16,  1 }, // pblendw
-    { TTI::SK_Alternate, MVT::v16i8,  1 }  // pblendvb
+    { TTI::SK_Select,    MVT::v2i64,  1 }, // pblendw
+    { TTI::SK_Select,    MVT::v2f64,  1 }, // movsd
+    { TTI::SK_Select,    MVT::v4i32,  1 }, // pblendw
+    { TTI::SK_Select,    MVT::v4f32,  1 }, // blendps
+    { TTI::SK_Select,    MVT::v8i16,  1 }, // pblendw
+    { TTI::SK_Select,    MVT::v16i8,  1 }  // pblendvb
   };
 
   if (ST->hasSSE41())
     if (const auto *Entry = CostTableLookup(SSE41ShuffleTbl, Kind, LT.second))
       return LT.first * Entry->Cost;
 
   static const CostTblEntry SSSE3ShuffleTbl[] = {
     { TTI::SK_Broadcast, MVT::v8i16,  1 }, // pshufb
     { TTI::SK_Broadcast, MVT::v16i8,  1 }, // pshufb
 
     { TTI::SK_Reverse,   MVT::v8i16,  1 }, // pshufb
     { TTI::SK_Reverse,   MVT::v16i8,  1 }, // pshufb
 
-    { TTI::SK_Alternate, MVT::v8i16,  3 }, // 2*pshufb + por
-    { TTI::SK_Alternate, MVT::v16i8,  3 }, // 2*pshufb + por
+    { TTI::SK_Select,    MVT::v8i16,  3 }, // 2*pshufb + por
+    { TTI::SK_Select,    MVT::v16i8,  3 }, // 2*pshufb + por
 
     { TTI::SK_PermuteSingleSrc, MVT::v8i16, 1 }, // pshufb
     { TTI::SK_PermuteSingleSrc, MVT::v16i8, 1 }, // pshufb
 
     { TTI::SK_PermuteTwoSrc,    MVT::v8i16, 3 }, // 2*pshufb + por
     { TTI::SK_PermuteTwoSrc,    MVT::v16i8, 3 }, // 2*pshufb + por
   };
 
static const CostTblEntry SSE2ShuffleTbl[] = {
 
     { TTI::SK_Reverse,   MVT::v2f64,  1 }, // shufpd
     { TTI::SK_Reverse,   MVT::v2i64,  1 }, // pshufd
     { TTI::SK_Reverse,   MVT::v4i32,  1 }, // pshufd
     { TTI::SK_Reverse,   MVT::v8i16,  3 }, // pshuflw + pshufhw + pshufd
     { TTI::SK_Reverse,   MVT::v16i8,  9 }, // 2*pshuflw + 2*pshufhw
                                            // + 2*pshufd + 2*unpck + packus
 
-    { TTI::SK_Alternate, MVT::v2i64,  1 }, // movsd
-    { TTI::SK_Alternate, MVT::v2f64,  1 }, // movsd
-    { TTI::SK_Alternate, MVT::v4i32,  2 }, // 2*shufps
-    { TTI::SK_Alternate, MVT::v8i16,  3 }, // pand + pandn + por
-    { TTI::SK_Alternate, MVT::v16i8,  3 }, // pand + pandn + por
+    { TTI::SK_Select,    MVT::v2i64,  1 }, // movsd
+    { TTI::SK_Select,    MVT::v2f64,  1 }, // movsd
+    { TTI::SK_Select,    MVT::v4i32,  2 }, // 2*shufps
+    { TTI::SK_Select,    MVT::v8i16,  3 }, // pand + pandn + por
+    { TTI::SK_Select,    MVT::v16i8,  3 }, // pand + pandn + por
 
     { TTI::SK_PermuteSingleSrc, MVT::v2f64,  1 }, // shufpd
     { TTI::SK_PermuteSingleSrc, MVT::v2i64,  1 }, // pshufd
     { TTI::SK_PermuteSingleSrc, MVT::v4i32,  1 }, // pshufd
     { TTI::SK_PermuteSingleSrc, MVT::v8i16,  5 }, // 2*pshuflw + 2*pshufhw
                                                   // + pshufd/unpck
     { TTI::SK_PermuteSingleSrc, MVT::v16i8, 10 }, // 2*pshuflw + 2*pshufhw
                                                   // + 2*pshufd + 2*unpck + 2*packus
 
     { TTI::SK_PermuteTwoSrc,    MVT::v2f64,  1 }, // shufpd
     { TTI::SK_PermuteTwoSrc,    MVT::v2i64,  1 }, // shufpd
     { TTI::SK_PermuteTwoSrc,    MVT::v4i32,  2 }, // 2*{unpck,movsd,pshufd}
     { TTI::SK_PermuteTwoSrc,    MVT::v8i16,  8 }, // blend+permute
     { TTI::SK_PermuteTwoSrc,    MVT::v16i8, 13 }, // blend+permute
   };
 
   if (ST->hasSSE2())
     if (const auto *Entry = CostTableLookup(SSE2ShuffleTbl, Kind, LT.second))
       return LT.first * Entry->Cost;
 
   static const CostTblEntry SSE1ShuffleTbl[] = {
     { TTI::SK_Broadcast,        MVT::v4f32, 1 }, // shufps
     { TTI::SK_Reverse,          MVT::v4f32, 1 }, // shufps
-    { TTI::SK_Alternate,        MVT::v4f32, 2 }, // 2*shufps
+    { TTI::SK_Select,           MVT::v4f32, 2 }, // 2*shufps
     { TTI::SK_PermuteSingleSrc, MVT::v4f32, 1 }, // shufps
     { TTI::SK_PermuteTwoSrc,    MVT::v4f32, 2 }, // 2*shufps
   };
 
   if (ST->hasSSE1())
     if (const auto *Entry = CostTableLookup(SSE1ShuffleTbl, Kind, LT.second))
       return LT.first * Entry->Cost;
 
int X86TTIImpl::getMaskedMemoryOpCost(unsigned Opcode, Type *SrcTy,
 
   // Legalize the type.
   std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, SrcVTy);
   auto VT = TLI->getValueType(DL, SrcVTy);
   int Cost = 0;
   if (VT.isSimple() && LT.second != VT.getSimpleVT() &&
       LT.second.getVectorNumElements() == NumElem)
     // Promotion requires expand/truncate for data and a shuffle for mask.
-    Cost += getShuffleCost(TTI::SK_Alternate, SrcVTy, 0, nullptr) +
-            getShuffleCost(TTI::SK_Alternate, MaskTy, 0, nullptr);
+    Cost += getShuffleCost(TTI::SK_Select, SrcVTy, 0, nullptr) +
+            getShuffleCost(TTI::SK_Select, MaskTy, 0, nullptr);
 
   else if (LT.second.getVectorNumElements() > NumElem) {
     VectorType *NewMaskTy = VectorType::get(MaskTy->getVectorElementType(),
                                             LT.second.getVectorNumElements());
     // Expanding requires fill mask with zeroes
     Cost += getShuffleCost(TTI::SK_InsertSubvector, NewMaskTy, 0, MaskTy);
   }
   if (!ST->hasAVX512())

lib/Transforms/Vectorize/SLPVectorizer.cpp

for (unsigned I = 0, E = VL.size(); I < E; ++I) {
     // operation - permutation.
     if (CommonShuffleMode != CurrentShuffleMode)
       CommonShuffleMode = Permute;
   }
   // If we're not crossing lanes in different vectors, consider it as blending.
   if ((CommonShuffleMode == FirstAlternate ||
        CommonShuffleMode == SecondAlternate) &&
       Vec2)
-    return TargetTransformInfo::SK_Alternate;
+    return TargetTransformInfo::SK_Select;
   // If Vec2 was never used, we have a permutation of a single vector, otherwise
   // we have permutation of 2 vectors.
   return Vec2 ? TargetTransformInfo::SK_PermuteTwoSrc
               : TargetTransformInfo::SK_PermuteSingleSrc;
 }
 
 ///\returns Opcode that can be clubbed with \p Op to create an alternate
 /// sequence which can later be merged as a ShuffleVector instruction.
case Instruction::ShuffleVector: {
       // VecCost is equal to sum of the cost of creating 2 vectors
       // and the cost of creating shuffle.
       Instruction *I0 = cast<Instruction>(VL[0]);
       VecCost =
           TTI->getArithmeticInstrCost(I0->getOpcode(), VecTy, Op1VK, Op2VK);
       Instruction *I1 = cast<Instruction>(VL[1]);
       VecCost +=
           TTI->getArithmeticInstrCost(I1->getOpcode(), VecTy, Op1VK, Op2VK);
-      VecCost +=
-          TTI->getShuffleCost(TargetTransformInfo::SK_Alternate, VecTy, 0);
+      VecCost += TTI->getShuffleCost(TargetTransformInfo::SK_Select, VecTy, 0);
       return ReuseShuffleCost + VecCost - ScalarCost;
     }
     default:
       llvm_unreachable("Unknown instruction");
   }
 }
 
 bool BoUpSLP::isFullyVectorizableTinyTree() {

test/Analysis/CostModel/X86/alternate-shuffle-cost.ll

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
 ;
   %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
   ret <4 x i32> %1
 }
 
 define <4 x i32> @test_v4i32_3(<4 x i32> %a, <4 x i32> %b) {
-; CHECK-LABEL: 'test_v4i32_3'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
+; SSE2-LABEL: 'test_v4i32_3'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
+;
+; SSSE3-LABEL: 'test_v4i32_3'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
+;
+; SSE42-LABEL: 'test_v4i32_3'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
+;
+; AVX-LABEL: 'test_v4i32_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
 ;
 ; BTVER2-LABEL: 'test_v4i32_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %1
 ;
   %1 = shufflevector <4 x i32> %a, <4 x i32> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 3>
   ret <4 x i32> %1
 }
 
 define <4 x float> @test_v4f32(<4 x float> %a, <4 x float> %b) {
 ; SSE2-LABEL: 'test_v4f32'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
 ;
   %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
   ret <4 x float> %1
 }
 
 define <4 x float> @test_v4f32_3(<4 x float> %a, <4 x float> %b) {
-; CHECK-LABEL: 'test_v4f32_3'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
+; SSE2-LABEL: 'test_v4f32_3'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
+;
+; SSSE3-LABEL: 'test_v4f32_3'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
+;
+; SSE42-LABEL: 'test_v4f32_3'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
+;
+; AVX-LABEL: 'test_v4f32_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
 ;
 ; BTVER2-LABEL: 'test_v4f32_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %1
 ;
   %1 = shufflevector <4 x float> %a, <4 x float> %b, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
   ret <4 x float> %1
 }
 
 define <4 x i64> @test_v4i64(<4 x i64> %a, <4 x i64> %b) {
 ; SSE-LABEL: 'test_v4i64'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
 ;
   %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
   ret <4 x i64> %1
 }
 
 define <4 x i64> @test_v4i64_3(<4 x i64> %a, <4 x i64> %b) {
 ; SSE-LABEL: 'test_v4i64_3'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
 ;
-; AVX1-LABEL: 'test_v4i64_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
-;
-; AVX2-LABEL: 'test_v4i64_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
-;
-; XOPAVX1-LABEL: 'test_v4i64_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
-;
-; XOPAVX2-LABEL: 'test_v4i64_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
+; AVX-LABEL: 'test_v4i64_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
 ;
 ; BTVER2-LABEL: 'test_v4i64_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1
 ;
   %1 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
   ret <4 x i64> %1
 }
 
 define <4 x double> @test_v4f64(<4 x double> %a, <4 x double> %b) {
 ; SSE-LABEL: 'test_v4f64'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
 ;
   %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
   ret <4 x double> %1
 }
 
 define <4 x double> @test_v4f64_3(<4 x double> %a, <4 x double> %b) {
 ; SSE-LABEL: 'test_v4f64_3'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
 ;
-; AVX1-LABEL: 'test_v4f64_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
-;
-; AVX2-LABEL: 'test_v4f64_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
-;
-; XOPAVX1-LABEL: 'test_v4f64_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
-;
-; XOPAVX2-LABEL: 'test_v4f64_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
+; AVX-LABEL: 'test_v4f64_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
 ;
 ; BTVER2-LABEL: 'test_v4f64_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
 ;
   %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
   ret <4 x double> %1
 }
 
 ; Test shuffles on packed vectors of eight elements.
 

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
   %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 8, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 7>
   ret <8 x i16> %1
 }
 
 define <8 x i16> @test_v8i16_3(<8 x i16> %a, <8 x i16> %b) {
 ; SSE2-LABEL: 'test_v8i16_3'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
 ; SSSE3-LABEL: 'test_v8i16_3'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
 ; SSE42-LABEL: 'test_v8i16_3'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
-; AVX1-LABEL: 'test_v8i16_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
-;
-; AVX2-LABEL: 'test_v8i16_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
-;
-; XOPAVX1-LABEL: 'test_v8i16_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
-;
-; XOPAVX2-LABEL: 'test_v8i16_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
+; AVX-LABEL: 'test_v8i16_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
 ; BTVER2-LABEL: 'test_v8i16_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i16> %1
 ;
   %1 = shufflevector <8 x i16> %a, <8 x i16> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
   ret <8 x i16> %1
 }
 
 define <8 x i32> @test_v8i32(<8 x i32> %a, <8 x i32> %b) {
 ; SSE2-LABEL: 'test_v8i32'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = 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>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
 ;
   %1 = 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> %1
 }
 
 define <8 x i32> @test_v8i32_3(<8 x i32> %a, <8 x i32> %b) {
-; SSE-LABEL: 'test_v8i32_3'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
-;
-; AVX1-LABEL: 'test_v8i32_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
-;
-; AVX2-LABEL: 'test_v8i32_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
-;
-; XOPAVX1-LABEL: 'test_v8i32_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
-;
-; XOPAVX2-LABEL: 'test_v8i32_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
+; SSE2-LABEL: 'test_v8i32_3'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
+;
+; SSSE3-LABEL: 'test_v8i32_3'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
+;
+; SSE42-LABEL: 'test_v8i32_3'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
+;
+; AVX-LABEL: 'test_v8i32_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
 ;
 ; BTVER2-LABEL: 'test_v8i32_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %1
 ;
   %1 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
   ret <8 x i32> %1
 }
 
 define <8 x float> @test_v8f32(<8 x float> %a, <8 x float> %b) {
 ; SSE2-LABEL: 'test_v8f32'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = 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>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
 ;
   %1 = 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> %1
 }
 
 define <8 x float> @test_v8f32_3(<8 x float> %a, <8 x float> %b) {
-; SSE-LABEL: 'test_v8f32_3'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
-;
-; AVX1-LABEL: 'test_v8f32_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
-;
-; AVX2-LABEL: 'test_v8f32_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
-;
-; XOPAVX1-LABEL: 'test_v8f32_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
-;
-; XOPAVX2-LABEL: 'test_v8f32_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
+; SSE2-LABEL: 'test_v8f32_3'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
+;
+; SSSE3-LABEL: 'test_v8f32_3'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
+;
+; SSE42-LABEL: 'test_v8f32_3'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
+;
+; AVX-LABEL: 'test_v8f32_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
 ;
 ; BTVER2-LABEL: 'test_v8f32_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %1
 ;
   %1 = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 1, i32 10, i32 3, i32 12, i32 5, i32 14, i32 15>
   ret <8 x float> %1
 }
 
 ; Test shuffles on packed vectors of sixteen elements.
 

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
   %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 13, i32 30, i32 15>
   ret <16 x i8> %1
 }
 
 define <16 x i8> @test_v16i8_3(<16 x i8> %a, <16 x i8> %b) {
 ; SSE2-LABEL: 'test_v16i8_3'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
 ; SSSE3-LABEL: 'test_v16i8_3'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
 ; SSE42-LABEL: 'test_v16i8_3'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
-; AVX1-LABEL: 'test_v16i8_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
-;
-; AVX2-LABEL: 'test_v16i8_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
-;
-; XOPAVX1-LABEL: 'test_v16i8_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
-;
-; XOPAVX2-LABEL: 'test_v16i8_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
+; AVX-LABEL: 'test_v16i8_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
 ; BTVER2-LABEL: 'test_v16i8_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %1
 ;
   %1 = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
   ret <16 x i8> %1
 }
 
 define <16 x i16> @test_v16i16(<16 x i16> %a, <16 x i16> %b) {
 ; SSE2-LABEL: 'test_v16i16'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
   %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 16, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 13, i32 30, i32 15>
   ret <16 x i16> %1
 }
 
 define <16 x i16> @test_v16i16_3(<16 x i16> %a, <16 x i16> %b) {
 ; SSE2-LABEL: 'test_v16i16_3'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; SSSE3-LABEL: 'test_v16i16_3'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; SSE42-LABEL: 'test_v16i16_3'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; AVX1-LABEL: 'test_v16i16_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; AVX2-LABEL: 'test_v16i16_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; XOPAVX1-LABEL: 'test_v16i16_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; XOPAVX2-LABEL: 'test_v16i16_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
 ; BTVER2-LABEL: 'test_v16i16_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %1
 ;
   %1 = shufflevector <16 x i16> %a, <16 x i16> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 3, i32 20, i32 5, i32 22, i32 7, i32 24, i32 9, i32 26, i32 11, i32 28, i32 29, i32 30, i32 31>
   ret <16 x i16> %1
 }
 
 define <32 x i8> @test_v32i8(<32 x i8> %a, <32 x i8> %b) {
 ; SSE2-LABEL: 'test_v32i8'

 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
   %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 32, i32 1, i32 34, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 29, i32 62, i32 31>
   ret <32 x i8> %1
 }
 
 define <32 x i8> @test_v32i8_3(<32 x i8> %a, <32 x i8> %b) {
 ; SSE2-LABEL: 'test_v32i8_3'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; SSSE3-LABEL: 'test_v32i8_3'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; SSE42-LABEL: 'test_v32i8_3'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; AVX1-LABEL: 'test_v32i8_3'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; AVX2-LABEL: 'test_v32i8_3'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; XOPAVX1-LABEL: 'test_v32i8_3'
-; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; XOPAVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; XOPAVX2-LABEL: 'test_v32i8_3'
-; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; XOPAVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
 ; BTVER2-LABEL: 'test_v32i8_3'
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %1
 ;
   %1 = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 36, i32 5, i32 38, i32 7, i32 40, i32 9, i32 42, i32 11, i32 44, i32 13, i32 46, i32 15, i32 48, i32 17, i32 50, i32 19, i32 52, i32 21, i32 54, i32 23, i32 56, i32 25, i32 58, i32 27, i32 60, i32 61, i32 62, i32 63>
   ret <32 x i8> %1
 }