[X86] Improve cost model for subvector extraction of less than 128-bit vectors

Now that we're using widening legalization. We need to improve our extract_subvector cost model for these types. This patch begins by modeling these as a subvector extract followed by a permute. I've left FIXMEs in the code for future improvements.

Differential Revision: https://reviews.llvm.org/D65892

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@369022 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86TargetTransformInfo.cpp b/lib/Target/X86/X86TargetTransformInfo.cpp
index 92f0b907d5029a571ac8b24bc2f084f8c091964a..0ff08776c07102c815d142c2b8e1600b90fcc893 100644 (file)
--- a/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -911,6 +911,39 @@ int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
       int NumSubElts = SubLT.second.getVectorNumElements();
       if ((Index % NumSubElts) == 0 && (NumElts % NumSubElts) == 0)
         return SubLT.first;
+      // Handle some cases for widening legalization. For now we only handle
+      // cases where the original subvector was naturally aligned and evenly
+      // fit in its legalized subvector type.
+      // FIXME: Remove some of the alignment restrictions.
+      // FIXME: We can use permq for 64-bit or larger extracts from 256-bit
+      // vectors.
+      int OrigSubElts = SubTp->getVectorNumElements();
+      if (NumSubElts > OrigSubElts &&
+          (Index % OrigSubElts) == 0 && (NumSubElts % OrigSubElts) == 0 &&
+          LT.second.getVectorElementType() ==
+            SubLT.second.getVectorElementType() &&
+          LT.second.getVectorElementType().getSizeInBits() ==
+            Tp->getVectorElementType()->getPrimitiveSizeInBits()) {
+        assert(NumElts >= NumSubElts && NumElts > OrigSubElts &&
+               "Unexpected number of elements!");
+        Type *VecTy = VectorType::get(Tp->getVectorElementType(),
+                                      LT.second.getVectorNumElements());
+        Type *SubTy = VectorType::get(Tp->getVectorElementType(),
+                                      SubLT.second.getVectorNumElements());
+        int ExtractIndex = alignDown((Index % NumElts), NumSubElts);
+        int ExtractCost = getShuffleCost(TTI::SK_ExtractSubvector, VecTy,
+                                         ExtractIndex, SubTy);
+
+        // If the original size is 32-bits or more, we can use pshufd. Otherwise
+        // if we have SSSE3 we can use pshufb.
+        if (SubTp->getPrimitiveSizeInBits() >= 32 || ST->hasSSSE3())
+          return ExtractCost + 1; // pshufd or pshufb
+
+        assert(SubTp->getPrimitiveSizeInBits() == 16 &&
+               "Unexpected vector size");
+
+        return ExtractCost + 2; // worst case pshufhw + pshufd
+      }
     }
   }
 

diff --git a/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll b/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
index e96e1ad0280a940d97758011a5912296aa456ea2..b35111e02afbed8522617ebd64818845304fcd38 100644 (file)
--- a/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
+++ b/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
@@ -141,21 +141,21 @@ define void @test_vXi64(<4 x i64> %src256, <8 x i64> %src512) {
 define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
@@ -166,21 +166,21 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
 ;
 ; AVX-LABEL: 'test_vXi32'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
@@ -191,21 +191,21 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 4, i32 5>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 0, i32 1>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 4, i32 5>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 8, i32 9>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 12, i32 13>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
@@ -216,21 +216,21 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
 ;
 ; BTVER2-LABEL: 'test_vXi32'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 10, i32 11>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
@@ -267,55 +267,55 @@ define void @test_vXi32(<4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512
 define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src512) {
 ; SSE-LABEL: 'test_vXi16'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_89 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_2345 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_6789 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 16, i32 17>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_18_19 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 24, i32 25>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_02_03_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_06_07_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_08_09_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_18_19_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
@@ -326,55 +326,55 @@ define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256,
 ;
 ; AVX-LABEL: 'test_vXi16'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_2345 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_6789 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 16, i32 17>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 24, i32 25>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_02_03_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_06_07_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
@@ -385,55 +385,55 @@ define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256,
 ;
 ; AVX512F-LABEL: 'test_vXi16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_2345 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_6789 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 16, i32 17>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 24, i32 25>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_02_03_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_06_07_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
@@ -444,55 +444,55 @@ define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256,
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_2345 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_6789 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 16, i32 17>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 24, i32 25>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_02_03_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_06_07_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
@@ -503,55 +503,55 @@ define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256,
 ;
 ; BTVER2-LABEL: 'test_vXi16'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 12, i32 13>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_2345 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_6789 = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 12, i32 13>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 16, i32 17>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 18, i32 19>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 20, i32 21>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 22, i32 23>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 24, i32 25>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 26, i32 27>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 28, i32 29>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <2 x i32> <i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_02_03_04_05 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_06_07_08_09 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <32 x i16> %src512, <32 x i16> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <32 x i16> %src512, <32 x i16> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
@@ -620,236 +620,474 @@ define void @test_vXi16(<4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256,
 }
 
 define void @test_vXi8(<8 x i8> %src64, <16 x i8> %src128, <32 x i8> %src256, <64 x i8> %src512) {
-; SSE-LABEL: 'test_vXi8'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 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 31>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SSE2-LABEL: 'test_vXi8'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 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 31>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXi8'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 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 31>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXi8'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 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 31>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F_30_31_32_33_34_35_36_37_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXi8'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
@@ -860,115 +1098,115 @@ define void @test_vXi8(<8 x i8> %src64, <16 x i8> %src128, <32 x i8> %src256, <6
 ;
 ; AVX512F-LABEL: 'test_vXi8'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
@@ -979,115 +1217,115 @@ define void @test_vXi8(<8 x i8> %src64, <16 x i8> %src128, <32 x i8> %src256, <6
 ;
 ; AVX512BW-LABEL: 'test_vXi8'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
@@ -1098,115 +1336,115 @@ define void @test_vXi8(<8 x i8> %src64, <16 x i8> %src128, <32 x i8> %src256, <6
 ;
 ; BTVER2-LABEL: 'test_vXi8'
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> undef, <2 x i32> <i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89 = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 8, i32 9>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 12, i32 13>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_2345 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_6789 = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 8, i32 9>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 12, i32 13>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 16, i32 17>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 18, i32 19>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 20, i32 21>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19 = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 24, i32 25>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 26, i32 27>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 28, i32 29>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_02_03_04_05 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_06_07_08_09 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 6, i32 7, i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17 = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 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>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <32 x i8> %src256, <32 x i8> undef, <16 x i32> <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 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 0, i32 1>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 2, i32 3>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 4, i32 5>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 8, i32 9>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 12, i32 13>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 16, i32 17>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 18, i32 19>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 20, i32 21>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 24, i32 25>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 26, i32 27>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 28, i32 29>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 32, i32 33>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 34, i32 35>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 36, i32 37>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 38, i32 39>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 40, i32 41>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 42, i32 43>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 44, i32 45>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 46, i32 47>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 48, i32 49>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 50, i32 51>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 52, i32 53>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 54, i32 55>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39 = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 56, i32 57>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 58, i32 59>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 60, i32 61>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <2 x i32> <i32 62, i32 63>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 16, i32 17, i32 18, i32 19>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 20, i32 21, i32 22, i32 23>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 24, i32 25, i32 26, i32 27>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 32, i32 33, i32 34, i32 35>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 36, i32 37, i32 38, i32 39>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 40, i32 41, i32 42, i32 43>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 44, i32 45, i32 46, i32 47>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 48, i32 49, i32 50, i32 51>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 52, i32 53, i32 54, i32 55>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 56, i32 57, i32 58, i32 59>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <4 x i32> <i32 60, i32 61, i32 62, i32 63>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_30_31_32_33_34_35_36_37 = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55>
-; BTVER2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; BTVER2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_38_39_3A_3B_3C_3D_3E_3F = shufflevector <64 x i8> %src512, <64 x i8> undef, <8 x i32> <i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_00_01_02_03_04_05_06_07_08_09_0A_0B_0C_0D_0E_0F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 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>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_10_11_12_13_14_15_16_17_18_19_1A_1B_1C_1D_1E_1F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <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 31>
 ; BTVER2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_20_21_22_23_24_25_26_27_28_29_2A_2B_2C_2D_2E_2F = shufflevector <64 x i8> %src512, <64 x i8> undef, <16 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47>