[llvm] [CostModel] getInstructionCost - match SK_InsertSubvector shuffle patterns before SK_Select (PR #145920)

Thu Jun 26 10:59:14 PDT 2025

https://github.com/RKSimon updated https://github.com/llvm/llvm-project/pull/145920

>From 8548a40488ae06fb7300f85edee40593b9a70e9e Mon Sep 17 00:00:00 2001
From: Simon Pilgrim <llvm-dev at redking.me.uk>
Date: Thu, 26 Jun 2025 16:48:59 +0100
Subject: [PATCH] [CostModel] getInstructionCost - match SK_InsertSubvector
 shuffle patterns before SK_Select

More closely match improveShuffleKindFromMask's shuffle ordering by trying to match a SK_InsertSubvector shuffles patterns before SK_Select - both can match many of the same patterns, but its much easier to recognise when a SK_InsertSubvector can be converted to SK_Select then vice-versa.

Another step towards #145335 - which I'm hoping will allow us to generalise improveShuffleKindFromMask and remove getInstructionCost's shuffle matching entirely.
---
 .../llvm/Analysis/TargetTransformInfoImpl.h   |  14 +-
 .../CostModel/AArch64/shuffle-other.ll        |  10 +-
 .../CostModel/AMDGPU/shufflevector.ll         |  48 ++---
 llvm/test/Analysis/CostModel/ARM/shuffle.ll   |   4 +-
 .../CostModel/RISCV/shuffle-exact-vlen.ll     |  24 +--
 .../RISCV/shuffle-insert_subvector.ll         | 120 ++++++-------
 .../CostModel/RISCV/shuffle-select.ll         |   8 +-
 .../CostModel/X86/alternate-shuffle-cost.ll   |  46 ++---
 .../CostModel/X86/shuffle-insert_subvector.ll | 164 ++++++------------
 .../Analysis/CostModel/X86/shuffle-select.ll  | 113 +++++-------
 .../test/Transforms/PhaseOrdering/X86/hadd.ll |   3 +-
 .../test/Transforms/PhaseOrdering/X86/hsub.ll |   3 +-
 12 files changed, 223 insertions(+), 334 deletions(-)

diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
index c22928c9bcd94..d513e9472a152 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
@@ -1538,7 +1538,7 @@ class TargetTransformInfoImplCRTPBase : public TargetTransformInfoImplBase {
       if (Shuffle->changesLength()) {
         // Treat a 'subvector widening' as a free shuffle.
         if (Shuffle->increasesLength() && Shuffle->isIdentityWithPadding())
-          return 0;
+          return TTI::TCC_Free;
 
         if (Shuffle->isExtractSubvectorMask(SubIndex))
           return TargetTTI->getShuffleCost(TTI::SK_ExtractSubvector, VecTy,
@@ -1599,18 +1599,13 @@ class TargetTransformInfoImplCRTPBase : public TargetTransformInfoImplBase {
       }
 
       if (Shuffle->isIdentity())
-        return 0;
+        return TTI::TCC_Free;
 
       if (Shuffle->isReverse())
         return TargetTTI->getShuffleCost(TTI::SK_Reverse, VecTy, VecSrcTy, Mask,
                                          CostKind, 0, nullptr, Operands,
                                          Shuffle);
 
-      if (Shuffle->isSelect())
-        return TargetTTI->getShuffleCost(TTI::SK_Select, VecTy, VecSrcTy, Mask,
-                                         CostKind, 0, nullptr, Operands,
-                                         Shuffle);
-
       if (Shuffle->isTranspose())
         return TargetTTI->getShuffleCost(TTI::SK_Transpose, VecTy, VecSrcTy,
                                          Mask, CostKind, 0, nullptr, Operands,
@@ -1632,6 +1627,11 @@ class TargetTransformInfoImplCRTPBase : public TargetTransformInfoImplBase {
             FixedVectorType::get(VecTy->getScalarType(), NumSubElts), Operands,
             Shuffle);
 
+      if (Shuffle->isSelect())
+        return TargetTTI->getShuffleCost(TTI::SK_Select, VecTy, VecSrcTy, Mask,
+                                         CostKind, 0, nullptr, Operands,
+                                         Shuffle);
+
       if (Shuffle->isSplice(SubIndex))
         return TargetTTI->getShuffleCost(TTI::SK_Splice, VecTy, VecSrcTy, Mask,
                                          CostKind, SubIndex, nullptr, Operands,
diff --git a/llvm/test/Analysis/CostModel/AArch64/shuffle-other.ll b/llvm/test/Analysis/CostModel/AArch64/shuffle-other.ll
index 8d68781d0c1b7..41c272291d7ca 100644
--- a/llvm/test/Analysis/CostModel/AArch64/shuffle-other.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/shuffle-other.ll
@@ -87,21 +87,21 @@ define void @concat() {
 
 define void @insert_subvec() {
 ; CHECK-LABEL: 'insert_subvec'
-; CHECK-NEXT:  Cost Model: Found costs of RThru:12 CodeSize:6 Lat:12 SizeLat:12 for: %v4i8_2_0 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v4i8_2_0 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v4i8_2_1 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-NEXT:  Cost Model: Found costs of RThru:28 CodeSize:14 Lat:28 SizeLat:28 for: %v8i8_2_0 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i8_2_0 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i8_2_1 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i8_2_2 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i8_2_3 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; CHECK-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:3 Lat:6 SizeLat:6 for: %v8i8_2_05 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 0, i32 8, i32 9, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found costs of RThru:60 CodeSize:30 Lat:60 SizeLat:60 for: %v16i8_4_0 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i8_4_0 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i8_4_1 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i8_4_2 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i8_4_3 = shufflevector <16 x i8> undef, <16 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 16, i32 17, i32 18, i32 19>
 ; CHECK-NEXT:  Cost Model: Found costs of RThru:14 CodeSize:7 Lat:14 SizeLat:14 for: %v16i8_4_05 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 18, i32 19, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v4i16_2_0 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v4i16_2_1 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-NEXT:  Cost Model: Found costs of RThru:28 CodeSize:14 Lat:28 SizeLat:28 for: %v8i16_2_0 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i16_2_0 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i16_2_1 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i16_2_2 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i16_2_3 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
@@ -118,7 +118,7 @@ define void @insert_subvec() {
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i32_2_2 = shufflevector <8 x i32> undef, <8 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v8i32_2_3 = shufflevector <8 x i32> undef, <8 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; CHECK-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:3 Lat:6 SizeLat:6 for: %v8i32_2_05 = shufflevector <8 x i32> undef, <8 x i32> undef, <8 x i32> <i32 0, i32 8, i32 9, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found costs of 0 for: %v16i32_4_0 = shufflevector <16 x i32> undef, <16 x i32> undef, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i32_4_0 = shufflevector <16 x i32> undef, <16 x i32> undef, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i32_4_1 = shufflevector <16 x i32> undef, <16 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i32_4_2 = shufflevector <16 x i32> undef, <16 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found costs of 1 for: %v16i32_4_3 = shufflevector <16 x i32> undef, <16 x i32> 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 16, i32 17, i32 18, i32 19>
diff --git a/llvm/test/Analysis/CostModel/AMDGPU/shufflevector.ll b/llvm/test/Analysis/CostModel/AMDGPU/shufflevector.ll
index 1705cd2a2f091..4fc8c44e12668 100644
--- a/llvm/test/Analysis/CostModel/AMDGPU/shufflevector.ll
+++ b/llvm/test/Analysis/CostModel/AMDGPU/shufflevector.ll
@@ -16,11 +16,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 2>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 0>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf22 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 2>
-; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
+; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 0>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf33 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 2>
-; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
+; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 1>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 3>
@@ -53,11 +53,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 2>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 0>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf22_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 2>
-; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
+; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 0>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf33_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 2>
-; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
+; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 3>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 1>
 ; GFX9-10-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 3>
@@ -93,11 +93,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 2>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 0>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf22 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 2>
-; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
+; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 0>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf33 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 2>
-; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
+; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 1>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 3>
@@ -130,11 +130,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 2>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 0>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf22_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 2>
-; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
+; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 0>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf33_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 2>
-; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
+; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 3>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 1>
 ; VI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 3>
@@ -170,11 +170,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 2>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 0>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf22 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 2>
-; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
+; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 0>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf33 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 2>
-; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
+; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 1>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 3>
@@ -207,11 +207,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 2>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 0>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf22_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 2>
-; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
+; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 0>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf33_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 2>
-; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
+; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 3>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 1>
 ; GFX9-10-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 3>
@@ -247,11 +247,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 2>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 0>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf22 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 2>
-; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
+; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 0, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 0>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf33 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 2>
-; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
+; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 1>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 1, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 3, i32 1>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i16> %vec1, <2 x i16> %vec1, <2 x i32> <i32 2, i32 3>
@@ -284,11 +284,11 @@ define amdgpu_kernel void @shufflevector_i16(<2 x i16> %vec1, <2 x i16> %vec2) {
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf02_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 2>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %shuf20_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 0>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf22_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 2>
-; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
+; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf03_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 0, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf30_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 0>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf33_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf12_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 2>
-; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
+; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf21_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 1>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf13_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 1, i32 3>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf31_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 3, i32 1>
 ; VI-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i16> %vec1, <2 x i16> %vec2, <2 x i32> <i32 2, i32 3>
@@ -403,11 +403,11 @@ define amdgpu_kernel void @shufflevector_i8(<2 x i8> %vec1, <2 x i8> %vec2) {
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf02 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 2>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf20 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 0>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf22 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 2>
-; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf03 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 3>
+; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf30 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 0>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf33 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf12 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 1, i32 2>
-; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf21 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 1>
+; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 1>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf13 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 1, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf31 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 1>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 3>
@@ -440,11 +440,11 @@ define amdgpu_kernel void @shufflevector_i8(<2 x i8> %vec1, <2 x i8> %vec2) {
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf02_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 2>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf20_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 0>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf22_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 2>
-; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf03_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 3>
+; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf30_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 0>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf33_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf12_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 1, i32 2>
-; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf21_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 1>
+; ALL-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 1>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf13_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 1, i32 3>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf31_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 1>
 ; ALL-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 3>
@@ -480,11 +480,11 @@ define amdgpu_kernel void @shufflevector_i8(<2 x i8> %vec1, <2 x i8> %vec2) {
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf02 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 2>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf20 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 0>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf22 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 2>
-; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf03 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 3>
+; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 0, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf30 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 0>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf33 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf12 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 1, i32 2>
-; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf21 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 1>
+; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 1>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf13 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 1, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf31 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 3, i32 1>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23 = shufflevector <2 x i8> %vec1, <2 x i8> %vec1, <2 x i32> <i32 2, i32 3>
@@ -517,11 +517,11 @@ define amdgpu_kernel void @shufflevector_i8(<2 x i8> %vec1, <2 x i8> %vec2) {
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf02_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 2>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf20_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 0>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf22_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 2>
-; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf03_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 3>
+; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf03_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 0, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf30_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 0>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %shuf33_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf12_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 1, i32 2>
-; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf21_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 1>
+; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %shuf21_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 1>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf13_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 1, i32 3>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %shuf31_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 3, i32 1>
 ; ALL-SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %shuf23_2 = shufflevector <2 x i8> %vec1, <2 x i8> %vec2, <2 x i32> <i32 2, i32 3>
diff --git a/llvm/test/Analysis/CostModel/ARM/shuffle.ll b/llvm/test/Analysis/CostModel/ARM/shuffle.ll
index 14f16252581aa..a17bbaba9d9af 100644
--- a/llvm/test/Analysis/CostModel/ARM/shuffle.ll
+++ b/llvm/test/Analysis/CostModel/ARM/shuffle.ll
@@ -241,7 +241,7 @@ define void @concat() {
 
 define void @select() {
 ; CHECK-MVE-LABEL: 'select'
-; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %v2i8 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> <i32 0, i32 3>
+; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v2i8 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> <i32 0, i32 3>
 ; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %v4i8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
 ; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 128 for instruction: %v8i8 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 8, i32 1, i32 10, i32 11, i32 4, i32 5, i32 6, i32 15>
 ; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 256 for instruction: %v16i8 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 0, i32 17, 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>
@@ -266,7 +266,7 @@ define void @select() {
 ; CHECK-MVE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; CHECK-NEON-LABEL: 'select'
-; CHECK-NEON-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v2i8 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> <i32 0, i32 3>
+; CHECK-NEON-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %v2i8 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> <i32 0, i32 3>
 ; CHECK-NEON-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %v4i8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
 ; CHECK-NEON-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %v8i8 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> <i32 8, i32 1, i32 10, i32 11, i32 4, i32 5, i32 6, i32 15>
 ; CHECK-NEON-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %v16i8 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> <i32 0, i32 17, 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>
diff --git a/llvm/test/Analysis/CostModel/RISCV/shuffle-exact-vlen.ll b/llvm/test/Analysis/CostModel/RISCV/shuffle-exact-vlen.ll
index b73abbfeba084..6640ff9793a38 100644
--- a/llvm/test/Analysis/CostModel/RISCV/shuffle-exact-vlen.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/shuffle-exact-vlen.ll
@@ -151,21 +151,21 @@ define void @concat() vscale_range(2,2) {
 
 define void @insert_subvec() vscale_range(2,2) {
 ; CHECK-LABEL: 'insert_subvec'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i8_2_0 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i8_2_0 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i8_2_1 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v8i8_2_0 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_0 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_1 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_2 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_3 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_05 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 8, i32 9, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v16i8_4_0 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_0 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_1 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_2 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_3 = shufflevector <16 x i8> poison, <16 x i8> poison, <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 16, i32 17, i32 18, i32 19>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_05 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 18, i32 19, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i16_2_0 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i16_2_0 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i16_2_1 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v8i16_2_0 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_0 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_1 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_2 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_3 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
@@ -175,7 +175,7 @@ define void @insert_subvec() vscale_range(2,2) {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_2 = shufflevector <16 x i16> poison, <16 x i16> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_3 = shufflevector <16 x i16> poison, <16 x i16> poison, <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 16, i32 17, i32 18, i32 19>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_05 = shufflevector <16 x i16> poison, <16 x i16> poison, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 18, i32 19, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i32_2_0 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i32_2_0 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i32_2_1 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i32_2_0 = shufflevector <8 x i32> poison, <8 x i32> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i32_2_1 = shufflevector <8 x i32> poison, <8 x i32> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
@@ -190,21 +190,21 @@ define void @insert_subvec() vscale_range(2,2) {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; CHECK-SIZE-LABEL: 'insert_subvec'
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i8_2_0 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i8_2_0 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i8_2_1 = shufflevector <4 x i8> poison, <4 x i8> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v8i8_2_0 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_0 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_1 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_2 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_3 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i8_2_05 = shufflevector <8 x i8> poison, <8 x i8> poison, <8 x i32> <i32 0, i32 8, i32 9, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v16i8_4_0 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_0 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_1 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_2 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_3 = shufflevector <16 x i8> poison, <16 x i8> poison, <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 16, i32 17, i32 18, i32 19>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i8_4_05 = shufflevector <16 x i8> poison, <16 x i8> poison, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 18, i32 19, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i16_2_0 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i16_2_0 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i16_2_1 = shufflevector <4 x i16> poison, <4 x i16> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v8i16_2_0 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_0 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_1 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_2 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i16_2_3 = shufflevector <8 x i16> poison, <8 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
@@ -214,7 +214,7 @@ define void @insert_subvec() vscale_range(2,2) {
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_2 = shufflevector <16 x i16> poison, <16 x i16> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_3 = shufflevector <16 x i16> poison, <16 x i16> poison, <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 16, i32 17, i32 18, i32 19>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v16i16_4_05 = shufflevector <16 x i16> poison, <16 x i16> poison, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 18, i32 19, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %v4i32_2_0 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i32_2_0 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v4i32_2_1 = shufflevector <4 x i32> poison, <4 x i32> poison, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i32_2_0 = shufflevector <8 x i32> poison, <8 x i32> poison, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v8i32_2_1 = shufflevector <8 x i32> poison, <8 x i32> poison, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
diff --git a/llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll b/llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
index ccfb5a9d42e7b..f4e29e15111ed 100644
--- a/llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
@@ -9,13 +9,13 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <2 x double> %src128, <2 x double> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <2 x double> %src128, <2 x double> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <4 x double> %src256, <4 x double> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_23 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_45 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_0123 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0123 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_4567 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -23,13 +23,13 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <2 x double> %src128, <2 x double> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <2 x double> %src128, <2 x double> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <4 x double> %src256, <4 x double> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <4 x double> %src256, <4 x double> %src128_256, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_67 = shufflevector <8 x double> %src512, <8 x double> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_0123 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0123 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <8 x double> %src512, <8 x double> %src256_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
@@ -53,13 +53,13 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512)
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <2 x i64> %src128, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <2 x i64> %src128, <2 x i64> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <4 x i64> %src256, <4 x i64> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_23 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_45 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_0123 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0123 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_4567 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -67,13 +67,13 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <2 x i64> %src128, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <2 x i64> %src128, <2 x i64> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <4 x i64> %src256, <4 x i64> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <4 x i64> %src256, <4 x i64> %src128_256, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_67 = shufflevector <8 x i64> %src512, <8 x i64> %src128_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_0123 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0123 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <8 x i64> %src512, <8 x i64> %src256_512, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
@@ -100,15 +100,15 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <4 x float> %src128, <4 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <4 x float> %src128, <4 x float> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <8 x float> %src256, <8 x float> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_23 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_45 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0123 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0123 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_4567 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 16, i32 17, 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>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 16, i32 17, 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>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_45 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -116,11 +116,11 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_CD = shufflevector <16 x float> %src512, <16 x float> %src64_512, <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 16, i32 17, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x float> %src512, <16 x float> %src64_512, <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 16, i32 17>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_0123 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0123 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_4567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_89AB = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_CDEF = shufflevector <16 x float> %src512, <16 x float> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01234567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01234567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_89ABCDEF = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -131,15 +131,15 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <4 x float> %src128, <4 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <4 x float> %src128, <4 x float> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <8 x float> %src256, <8 x float> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x float> %src128, <4 x float> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <8 x float> %src256, <8 x float> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_0123 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0123 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <8 x float> %src256, <8 x float> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 16, i32 17, 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>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 16, i32 17, 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>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_67 = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -147,11 +147,11 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_AB = shufflevector <16 x float> %src512, <16 x float> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CD = shufflevector <16 x float> %src512, <16 x float> %src64_512, <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 16, i32 17, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_EF = shufflevector <16 x float> %src512, <16 x float> %src64_512, <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 16, i32 17>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_0123 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0123 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89AB = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CDEF = shufflevector <16 x float> %src512, <16 x float> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01234567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01234567 = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89ABCDEF = shufflevector <16 x float> %src512, <16 x float> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
@@ -197,15 +197,15 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -213,11 +213,11 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -228,15 +228,15 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -244,11 +244,11 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
@@ -298,15 +298,15 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <8 x i16> %src128, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <8 x i16> %src128, <8 x i16> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <16 x i16> %src256, <16 x i16> 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -314,11 +314,11 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <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 16, i32 17, i32 18, i32 19>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -333,15 +333,15 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <8 x i16> %src128, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <8 x i16> %src128, <8 x i16> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <16 x i16> %src256, <16 x i16> 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -349,11 +349,11 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <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 16, i32 17, i32 18, i32 19>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
@@ -413,15 +413,15 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <16 x i8> %src128, <16 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <16 x i8> %src128, <16 x i8> undef, <64 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <32 x i8> %src256, <32 x i8> undef, <64 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, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32_23 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -429,11 +429,11 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <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 16, i32 17, i32 18, i32 19>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
@@ -453,15 +453,15 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_256 = shufflevector <16 x i8> %src128, <16 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src128_512 = shufflevector <16 x i8> %src128, <16 x i8> undef, <64 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %src256_512 = shufflevector <32 x i8> %src256, <32 x i8> undef, <64 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, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32_23 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -469,11 +469,11 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <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 16, i32 17, i32 18, i32 19>
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
diff --git a/llvm/test/Analysis/CostModel/RISCV/shuffle-select.ll b/llvm/test/Analysis/CostModel/RISCV/shuffle-select.ll
index 0fd1cfc908d97..2e17818c10d05 100644
--- a/llvm/test/Analysis/CostModel/RISCV/shuffle-select.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/shuffle-select.ll
@@ -6,11 +6,11 @@
 
 define <8 x i8> @select_start_v8i8(<8 x i8> %v, <8 x i8> %w) {
 ; CHECK-LABEL: 'select_start_v8i8'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res = shufflevector <8 x i8> %v, <8 x i8> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res = shufflevector <8 x i8> %v, <8 x i8> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i8> %res
 ;
 ; CHECK-SIZE-LABEL: 'select_start_v8i8'
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res = shufflevector <8 x i8> %v, <8 x i8> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res = shufflevector <8 x i8> %v, <8 x i8> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i8> %res
 ;
   %res = shufflevector <8 x i8> %v, <8 x i8> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
@@ -32,11 +32,11 @@ define <8 x i8> @select_non_contiguous_v8i8(<8 x i8> %v, <8 x i8> %w) {
 
 define <8 x i64> @select_start_v8i64(<8 x i64> %v, <8 x i64> %w) {
 ; CHECK-LABEL: 'select_start_v8i64'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %res = shufflevector <8 x i64> %v, <8 x i64> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res = shufflevector <8 x i64> %v, <8 x i64> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %res
 ;
 ; CHECK-SIZE-LABEL: 'select_start_v8i64'
-; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %res = shufflevector <8 x i64> %v, <8 x i64> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %res = shufflevector <8 x i64> %v, <8 x i64> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret <8 x i64> %res
 ;
   %res = shufflevector <8 x i64> %v, <8 x i64> %w, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
diff --git a/llvm/test/Analysis/CostModel/X86/alternate-shuffle-cost.ll b/llvm/test/Analysis/CostModel/X86/alternate-shuffle-cost.ll
index 033537d93bf77..dc770f4ff74c8 100644
--- a/llvm/test/Analysis/CostModel/X86/alternate-shuffle-cost.ll
+++ b/llvm/test/Analysis/CostModel/X86/alternate-shuffle-cost.ll
@@ -39,21 +39,9 @@ define <2 x i32> @test_v2i32(<2 x i32> %a, <2 x i32> %b) {
 }
 
 define <2 x float> @test_v2f32(<2 x float> %a, <2 x float> %b) {
-; SSE2-LABEL: 'test_v2f32'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; SSSE3-LABEL: 'test_v2f32'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; SSE42-LABEL: 'test_v2f32'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; AVX-LABEL: 'test_v2f32'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
+; CHECK-LABEL: 'test_v2f32'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
 ;
   %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 0, i32 3>
   ret <2 x float> %1
@@ -81,21 +69,9 @@ define <2 x i32> @test_v2i32_2(<2 x i32> %a, <2 x i32> %b) {
 }
 
 define <2 x float> @test_v2f32_2(<2 x float> %a, <2 x float> %b) {
-; SSE2-LABEL: 'test_v2f32_2'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; SSSE3-LABEL: 'test_v2f32_2'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; SSE42-LABEL: 'test_v2f32_2'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
-;
-; AVX-LABEL: 'test_v2f32_2'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
+; CHECK-LABEL: 'test_v2f32_2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <2 x float> %1
 ;
   %1 = shufflevector <2 x float> %a, <2 x float> %b, <2 x i32> <i32 2, i32 1>
   ret <2 x float> %1
@@ -317,9 +293,13 @@ define <4 x double> @test_v4f64_2(<4 x double> %a, <4 x double> %b) {
 }
 
 define <4 x double> @test_v4f64_3(<4 x double> %a, <4 x double> %b) {
-; CHECK-LABEL: 'test_v4f64_3'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
+; SSE-LABEL: 'test_v4f64_3'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
+;
+; AVX-LABEL: 'test_v4f64_3'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %1
 ;
   %1 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 6, i32 3>
   ret <4 x double> %1
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
index 62e8e7c3e79c9..40cbf26157e5d 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
@@ -292,98 +292,36 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 }
 
 define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512) {
-; SSE2-LABEL: 'test_vXi32'
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src64_128 = shufflevector <2 x i32> %src64, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src64_256 = shufflevector <2 x i32> %src64, <2 x i32> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src64_512 = shufflevector <2 x i32> %src64, <2 x i32> undef, <16 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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 costs of 2 for: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of 1 for: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of 2 for: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 0 for: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 0 for: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; SSE2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; SSSE3-LABEL: 'test_vXi32'
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src64_128 = shufflevector <2 x i32> %src64, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src64_256 = shufflevector <2 x i32> %src64, <2 x i32> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src64_512 = shufflevector <2 x i32> %src64, <2 x i32> undef, <16 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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 costs of 2 for: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of 1 for: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of 2 for: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 0 for: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; SSSE3-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; SSE42-LABEL: 'test_vXi32'
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src64_128 = shufflevector <2 x i32> %src64, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src64_256 = shufflevector <2 x i32> %src64, <2 x i32> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src64_512 = shufflevector <2 x i32> %src64, <2 x i32> undef, <16 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
-; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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 costs of 2 for: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of 1 for: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of 2 for: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 0 for: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 0 for: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
-; SSE42-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+; SSE-LABEL: 'test_vXi32'
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src64_128 = shufflevector <2 x i32> %src64, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src64_256 = shufflevector <2 x i32> %src64, <2 x i32> undef, <8 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src64_512 = shufflevector <2 x i32> %src64, <2 x i32> undef, <16 x i32> <i32 0, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <4 x i32> %src128, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <4 x i32> %src128, <4 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <8 x i32> %src256, <8 x i32> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V128_01 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V128_23 = shufflevector <4 x i32> %src128, <4 x i32> %src64_128, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V256_23 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V256_45 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V256_67 = shufflevector <8 x i32> %src256, <8 x i32> %src64_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V256_0123 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V256_4567 = shufflevector <8 x i32> %src256, <8 x i32> %src128_256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V512_01 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 16, i32 17, 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 costs of 2 for: %V512_23 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of 1 for: %V512_45 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of 2 for: %V512_67 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V512_89 = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V512_AB = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V512_CD = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V512_EF = shufflevector <16 x i32> %src512, <16 x i32> %src64_512, <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 16, i32 17>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V512_0123 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 0 for: %V512_4567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V512_89AB = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V512_CDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <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 16, i32 17, i32 18, i32 19>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V512_01234567 = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 0 for: %V512_89ABCDEF = shufflevector <16 x i32> %src512, <16 x i32> %src128_512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; SSE-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found costs of 0 for: %src64_128 = shufflevector <2 x i32> %src64, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
@@ -524,13 +462,13 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <8 x i16> %src128, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <8 x i16> %src128, <8 x i16> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <16 x i16> %src256, <16 x i16> 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:8 Lat:8 SizeLat:8 for: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:8 Lat:8 SizeLat:8 for: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:8 Lat:8 SizeLat:8 for: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:6 CodeSize:8 Lat:8 SizeLat:8 for: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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 costs of RThru:6 CodeSize:8 Lat:8 SizeLat:8 for: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -559,13 +497,13 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <8 x i16> %src128, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <8 x i16> %src128, <8 x i16> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <16 x i16> %src256, <16 x i16> 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V64_01 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_23 = shufflevector <4 x i16> %src64, <4 x i16> %src32_64, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V128_01 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_23 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_45 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
 ; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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 costs of 3 for: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -637,7 +575,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V128_67 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128_0123 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V128_4567 = shufflevector <8 x i16> %src128, <8 x i16> %src32_128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
+; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V256_01 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 16, i32 17, 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>
 ; AVX1-NEXT:  Cost Model: Found costs of 9 for: %V256_23 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 9 for: %V256_45 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 9 for: %V256_67 = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -645,11 +583,11 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found costs of 15 for: %V256_AB = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 15 for: %V256_CD = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 15 for: %V256_EF = shufflevector <16 x i16> %src256, <16 x i16> %src64_256, <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 16, i32 17>
-; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V256_0123 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 9 for: %V256_4567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V256_89AB = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 15 for: %V256_CDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <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 16, i32 17, i32 18, i32 19>
-; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V256_01234567 = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V256_89ABCDEF = shufflevector <16 x i16> %src256, <16 x i16> %src128_256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
@@ -849,15 +787,15 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <16 x i8> %src128, <16 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <16 x i8> %src128, <16 x i8> undef, <64 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <32 x i8> %src256, <32 x i8> undef, <64 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, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SSE2-NEXT:  Cost Model: Found costs of 4 for: %V32_23 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of 7 for: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of 7 for: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of 7 for: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SSE2-NEXT:  Cost Model: Found costs of 7 for: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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 costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -865,11 +803,11 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 1 for: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <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 16, i32 17, i32 18, i32 19>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:11 CodeSize:13 Lat:13 SizeLat:13 for: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SSE2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
@@ -889,15 +827,15 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_256 = shufflevector <16 x i8> %src128, <16 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src128_512 = shufflevector <16 x i8> %src128, <16 x i8> undef, <64 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 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
 ; SSSE3-NEXT:  Cost Model: Found costs of 0 for: %src256_512 = shufflevector <32 x i8> %src256, <32 x i8> undef, <64 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, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V32_01 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V32_23 = shufflevector <4 x i8> %src32, <4 x i8> %src16_32, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V64_01 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_23 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_45 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_67 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 8, i32 9>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V64_0123 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 4, i32 5, i32 6, i32 7>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64_4567 = shufflevector <8 x i8> %src64, <8 x i8> %src32_64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V128_01 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 16, i32 17, 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 costs of 3 for: %V128_23 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 16, i32 17, 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 costs of 3 for: %V128_45 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, 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 costs of 3 for: %V128_67 = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 16, i32 17, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -905,11 +843,11 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16x i
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_AB = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 16, i32 17, i32 12, i32 13, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_CD = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_EF = shufflevector <16 x i8> %src128, <16 x i8> %src16_128, <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 16, i32 17>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 1 for: %V128_0123 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, 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 costs of 3 for: %V128_4567 = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_89AB = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_CDEF = shufflevector <16 x i8> %src128, <16 x i8> %src32_128, <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 16, i32 17, i32 18, i32 19>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V128_01234567 = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128_89ABCDEF = shufflevector <16 x i8> %src128, <16 x i8> %src64_128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
 ; SSSE3-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-select.ll b/llvm/test/Analysis/CostModel/X86/shuffle-select.ll
index ba126f9028c7a..0d20860d3b659 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-select.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-select.ll
@@ -28,14 +28,14 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 ; XOP-LABEL: 'test_vXf64'
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 0, i32 3>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; XOP-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; XOP-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 16, i32 17, 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>
 ; XOP-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX-LABEL: 'test_vXf64'
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 0, i32 3>
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 16, i32 17, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
@@ -64,14 +64,14 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512,
 ; XOP-LABEL: 'test_vXi64'
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 0, i32 3>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; XOP-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; XOP-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 16, i32 17, 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>
 ; XOP-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX-LABEL: 'test_vXi64'
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 0, i32 3>
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found costs of 2 for: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 16, i32 17, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
@@ -90,29 +90,13 @@ define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512,
 }
 
 define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %src256, <16 x float> %src512, <32 x float> %src1024, <2 x float> %src64_1, <4 x float> %src128_1, <8 x float> %src256_1, <16 x float> %src512_1, <32 x float> %src1024_1) {
-; SSE2-LABEL: 'test_vXf32'
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 0, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; SSSE3-LABEL: 'test_vXf32'
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 0, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; SSE42-LABEL: 'test_vXf32'
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 0, i32 3>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+; SSE-LABEL: 'test_vXf32'
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 0, i32 3>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; XOP-LABEL: 'test_vXf32'
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 0, i32 3>
@@ -150,7 +134,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE2-LABEL: 'test_vXi32'
 ; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 0, i32 3>
 ; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
@@ -158,7 +142,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSSE3-LABEL: 'test_vXi32'
 ; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 0, i32 3>
 ; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
@@ -166,7 +150,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE42-LABEL: 'test_vXi32'
 ; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 0, i32 3>
 ; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSE42-NEXT:  Cost Model: Found costs of 2 for: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; SSE42-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
@@ -207,7 +191,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-LABEL: 'test_vXi16'
 ; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
 ; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
@@ -216,7 +200,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-LABEL: 'test_vXi16'
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
@@ -235,28 +219,19 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; XOP-NEXT:  Cost Model: Found costs of 3 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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>
+; XOP-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
 ; XOP-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
-; AVX1-LABEL: 'test_vXi16'
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
-; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; AVX2-LABEL: 'test_vXi16'
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
-; AVX2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+; AVX-LABEL: 'test_vXi16'
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
+; AVX-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX512-LABEL: 'test_vXi16'
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 0, i32 3>
@@ -280,8 +255,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE2-LABEL: 'test_vXi8'
 ; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
 ; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found costs of 3 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSE2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
@@ -289,8 +264,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSSE3-LABEL: 'test_vXi8'
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
 ; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found costs of 3 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; SSSE3-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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 costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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 costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
@@ -309,27 +284,18 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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>
-; XOP-NEXT:  Cost Model: Found costs of 3 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
+; XOP-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
 ; XOP-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
 ; XOP-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
-; AVX1-LABEL: 'test_vXi8'
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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>
-; AVX1-NEXT:  Cost Model: Found costs of 3 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
-; AVX1-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
-; AVX1-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-;
-; AVX2-LABEL: 'test_vXi8'
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
-; AVX2-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
-; AVX2-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+; AVX-LABEL: 'test_vXi8'
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 16, i32 17, 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 costs of 1 for: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 32, i32 33, i32 34, 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>
+; AVX-NEXT:  Cost Model: Found costs of 1 for: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
+; AVX-NEXT:  Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
 ;
 ; AVX512-LABEL: 'test_vXi8'
 ; AVX512-NEXT:  Cost Model: Found costs of 1 for: %V16 = shufflevector <2 x i8> %src16, <2 x i8> %src16_1, <2 x i32> <i32 0, i32 3>
@@ -348,3 +314,6 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
   %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 64, i32 65, i32 66, i32 67, 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, 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>
   ret void
 }
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; AVX1: {{.*}}
+; AVX2: {{.*}}
diff --git a/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll b/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
index b645288f96028..798df4cd4ff54 100644
--- a/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
+++ b/llvm/test/Transforms/PhaseOrdering/X86/hadd.ll
@@ -132,7 +132,8 @@ define <8 x i16> @add_v8i16_76u43210(<8 x i16> %a, <8 x i16> %b) {
 ; SSE2-NEXT:    [[TMP6:%.*]] = add <8 x i16> [[TMP4]], [[TMP5]]
 ; SSE2-NEXT:    [[HADD41:%.*]] = shufflevector <8 x i16> [[TMP1]], <8 x i16> [[TMP6]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:    [[HADD6:%.*]] = shufflevector <8 x i16> [[HADD41]], <8 x i16> [[TMP2]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 poison, i32 12, i32 poison>
-; SSE2-NEXT:    [[RESULT:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[HADD6]], <8 x i32> <i32 7, i32 14, i32 poison, i32 12, i32 11, i32 10, i32 9, i32 8>
+; SSE2-NEXT:    [[HADD7:%.*]] = shufflevector <8 x i16> [[HADD6]], <8 x i16> [[TMP3]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 poison, i32 6, i32 15>
+; SSE2-NEXT:    [[RESULT:%.*]] = shufflevector <8 x i16> [[HADD7]], <8 x i16> poison, <8 x i32> <i32 7, i32 6, i32 poison, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:    ret <8 x i16> [[RESULT]]
 ;
 ; SSE4-LABEL: @add_v8i16_76u43210(
diff --git a/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll b/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
index c7ebb73e8e35c..fd160b7c57024 100644
--- a/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
+++ b/llvm/test/Transforms/PhaseOrdering/X86/hsub.ll
@@ -132,7 +132,8 @@ define <8 x i16> @sub_v8i16_76u43210(<8 x i16> %a, <8 x i16> %b) {
 ; SSE2-NEXT:    [[TMP6:%.*]] = sub <8 x i16> [[TMP4]], [[TMP5]]
 ; SSE2-NEXT:    [[HSUB41:%.*]] = shufflevector <8 x i16> [[TMP1]], <8 x i16> [[TMP6]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 poison, i32 poison, i32 poison>
 ; SSE2-NEXT:    [[HSUB6:%.*]] = shufflevector <8 x i16> [[HSUB41]], <8 x i16> [[TMP2]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 poison, i32 12, i32 poison>
-; SSE2-NEXT:    [[RESULT:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[HSUB6]], <8 x i32> <i32 7, i32 14, i32 poison, i32 12, i32 11, i32 10, i32 9, i32 8>
+; SSE2-NEXT:    [[HSUB7:%.*]] = shufflevector <8 x i16> [[HSUB6]], <8 x i16> [[TMP3]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 poison, i32 6, i32 15>
+; SSE2-NEXT:    [[RESULT:%.*]] = shufflevector <8 x i16> [[HSUB7]], <8 x i16> poison, <8 x i32> <i32 7, i32 6, i32 poison, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:    ret <8 x i16> [[RESULT]]
 ;
 ; SSE4-LABEL: @sub_v8i16_76u43210(

