[llvm] [CostModel][X86] getShuffleCost - use processShuffleMasks to split SK_PermuteTwoSrc shuffles to legal types (PR #120599)

Thu Dec 19 08:26:20 PST 2024

https://github.com/RKSimon created https://github.com/llvm/llvm-project/pull/120599

processShuffleMasks can now correctly handle 2 src shuffles, so we can use the existing SK_PermuteSingleSrc splitting cost logic to handle SK_PermuteTwoSrc as well and correctly recognise the number of active subvectors per legalised shuffle.

>From bd011d6be1d6c6068415bbe8b8e0feb9d97cfbdc Mon Sep 17 00:00:00 2001
From: Simon Pilgrim <llvm-dev at redking.me.uk>
Date: Thu, 19 Dec 2024 16:24:46 +0000
Subject: [PATCH] [CostModel][X86] getShuffleCost - use processShuffleMasks to
 split SK_PermuteTwoSrc shuffles to legal types

processShuffleMasks can now correctly handle 2 src shuffles, so we can use the existing SK_PermuteSingleSrc splitting cost logic to handle SK_PermuteTwoSrc as well and correctly recognise the number of active subvectors per legalised shuffle.
---
 .../lib/Target/X86/X86TargetTransformInfo.cpp |  11 +-
 .../X86/shuffle-concat_subvector-codesize.ll  | 108 ++++-----
 .../X86/shuffle-concat_subvector-latency.ll   | 108 ++++-----
 .../shuffle-concat_subvector-sizelatency.ll   | 108 ++++-----
 .../CostModel/X86/shuffle-concat_subvector.ll | 108 ++++-----
 .../X86/shuffle-insert_subvector-codesize.ll  | 152 ++++++-------
 .../X86/shuffle-insert_subvector-latency.ll   | 152 ++++++-------
 .../shuffle-insert_subvector-sizelatency.ll   | 152 ++++++-------
 .../CostModel/X86/shuffle-insert_subvector.ll | 152 ++++++-------
 .../CostModel/X86/shuffle-two-src-codesize.ll | 210 +++++++++++-------
 .../X86/shuffle-two-src-fp16-codesize.ll      |   2 +-
 .../X86/shuffle-two-src-fp16-latency.ll       |   2 +-
 .../X86/shuffle-two-src-fp16-sizelatency.ll   |   2 +-
 .../CostModel/X86/shuffle-two-src-fp16.ll     |   2 +-
 .../CostModel/X86/shuffle-two-src-latency.ll  | 210 +++++++++++-------
 .../X86/shuffle-two-src-sizelatency.ll        | 210 +++++++++++-------
 .../Analysis/CostModel/X86/shuffle-two-src.ll | 210 +++++++++++-------
 .../Transforms/PhaseOrdering/X86/pr94546.ll   |  39 ++--
 .../SLPVectorizer/X86/horizontal-minmax.ll    |  19 +-
 .../X86/minbitwidth-transformed-operand.ll    |   7 +-
 .../VectorCombine/X86/shuffle-of-casts.ll     |  16 +-
 .../VectorCombine/X86/shuffle-of-shuffles.ll  |  27 ++-
 22 files changed, 1073 insertions(+), 934 deletions(-)

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 54c9998c0ead23..808f48eb92a61e 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -1698,7 +1698,8 @@ InstructionCost X86TTIImpl::getShuffleCost(
   // We are going to permute multiple sources and the result will be in multiple
   // destinations. Providing an accurate cost only for splits where the element
   // type remains the same.
-  if (Kind == TTI::SK_PermuteSingleSrc && LT.first != 1) {
+  if ((Kind == TTI::SK_PermuteSingleSrc || Kind == TTI::SK_PermuteTwoSrc) &&
+      LT.first != 1) {
     MVT LegalVT = LT.second;
     if (LegalVT.isVector() &&
         LegalVT.getVectorElementType().getSizeInBits() ==
@@ -1784,14 +1785,6 @@ InstructionCost X86TTIImpl::getShuffleCost(
     return BaseT::getShuffleCost(Kind, BaseTp, Mask, CostKind, Index, SubTp);
   }
 
-  // For 2-input shuffles, we must account for splitting the 2 inputs into many.
-  if (Kind == TTI::SK_PermuteTwoSrc && !IsInLaneShuffle && LT.first != 1) {
-    // We assume that source and destination have the same vector type.
-    InstructionCost NumOfDests = LT.first;
-    InstructionCost NumOfShufflesPerDest = LT.first * 2 - 1;
-    LT.first = NumOfDests * NumOfShufflesPerDest;
-  }
-
   static const CostTblEntry AVX512VBMIShuffleTbl[] = {
       {TTI::SK_Reverse, MVT::v64i8, 1}, // vpermb
       {TTI::SK_Reverse, MVT::v32i8, 1}, // vpermb
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-codesize.ll b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-codesize.ll
index c78023e24572ca..176a794ea666f2 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-codesize.ll
@@ -2,15 +2,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse2 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s -check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s -check-prefixes=AVX
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s --check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 ;
 ; Verify the cost model for concat_subvector style shuffles.
@@ -19,14 +19,14 @@
 define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a512, <2 x double> %b128, <4 x double> %b256, <8 x double> %b512) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXf64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -48,14 +48,14 @@ define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a5
 define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXi64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -77,24 +77,17 @@ define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x
 define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512, <4 x float> %b128, <8 x float> %b256, <16 x float> %b512) {
 ; SSE-LABEL: 'test_vXf32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXf32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXf32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXf32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -113,24 +106,17 @@ define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512
 define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXi32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXi32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -147,19 +133,19 @@ define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x
 }
 
 define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
-; AVX1-LABEL: 'test_vXi16'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi16'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi16'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -190,19 +176,19 @@ define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8
 }
 
 define void @test_vXi8(<16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
-; AVX1-LABEL: 'test_vXi8'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi8'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi8'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi8'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-latency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-latency.ll
index 78aaf6e15cd2d5..a99b1f0d7dbe87 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-latency.ll
@@ -2,15 +2,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse2 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s -check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s --check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 ;
 ; Verify the cost model for concat_subvector style shuffles.
@@ -19,14 +19,14 @@
 define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a512, <2 x double> %b128, <4 x double> %b256, <8 x double> %b512) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXf64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -48,14 +48,14 @@ define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a5
 define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXi64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -77,24 +77,17 @@ define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x
 define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512, <4 x float> %b128, <8 x float> %b256, <16 x float> %b512) {
 ; SSE-LABEL: 'test_vXf32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXf32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXf32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXf32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -113,24 +106,17 @@ define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512
 define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXi32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXi32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -147,19 +133,19 @@ define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x
 }
 
 define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
-; AVX1-LABEL: 'test_vXi16'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi16'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi16'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -190,19 +176,19 @@ define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8
 }
 
 define void @test_vXi8(<16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
-; AVX1-LABEL: 'test_vXi8'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi8'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi8'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi8'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-sizelatency.ll
index 4edfa8c1384d4d..85996551710ee8 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector-sizelatency.ll
@@ -2,15 +2,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse2 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s -check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s --check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 ;
 ; Verify the cost model for concat_subvector style shuffles.
@@ -19,14 +19,14 @@
 define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a512, <2 x double> %b128, <4 x double> %b256, <8 x double> %b512) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXf64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -48,14 +48,14 @@ define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a5
 define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXi64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -77,24 +77,17 @@ define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x
 define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512, <4 x float> %b128, <8 x float> %b256, <16 x float> %b512) {
 ; SSE-LABEL: 'test_vXf32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXf32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXf32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXf32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -113,24 +106,17 @@ define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512
 define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXi32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX-LABEL: 'test_vXi32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -147,19 +133,19 @@ define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x
 }
 
 define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
-; AVX1-LABEL: 'test_vXi16'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi16'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi16'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -190,19 +176,19 @@ define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8
 }
 
 define void @test_vXi8(<16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
-; AVX1-LABEL: 'test_vXi8'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi8'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE-LABEL: 'test_vXi8'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi8'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector.ll b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector.ll
index 3d5c2cfb2143d6..d530d11432d967 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-concat_subvector.ll
@@ -2,15 +2,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse2 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s -check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s -check-prefixes=AVX
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=slm | FileCheck %s --check-prefixes=SSE
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 ;
 ; Verify the cost model for concat_subvector style shuffles.
@@ -19,14 +19,14 @@
 define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a512, <2 x double> %b128, <4 x double> %b256, <8 x double> %b512) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXf64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x double> %a128, <2 x double> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x double> %a256, <4 x double> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x double> %a512, <8 x double> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
@@ -48,14 +48,14 @@ define void @test_vXf64(<2 x double> %a128, <4 x double> %a256, <8 x double> %a5
 define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX-LABEL: 'test_vXi64'
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <2 x i64> %a128, <2 x i64> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <4 x i64> %a256, <4 x i64> %b256, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <8 x i64> %a512, <8 x i64> %b512, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
@@ -77,24 +77,17 @@ define void @test_vXi64(<2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, <2 x
 define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512, <4 x float> %b128, <8 x float> %b256, <16 x float> %b512) {
 ; SSE-LABEL: 'test_vXf32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXf32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXf32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX-LABEL: 'test_vXf32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x float> %a256, <8 x float> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x float> %a512, <16 x float> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x float> %a128, <4 x float> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -113,24 +106,17 @@ define void @test_vXf32(<4 x float> %a128, <8 x float> %a256, <16 x float> %a512
 define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX1-LABEL: 'test_vXi32'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi32'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX-LABEL: 'test_vXi32'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <8 x i32> %a256, <8 x i32> %b256, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <16 x i32> %a512, <16 x i32> %b512, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <4 x i32> %a128, <4 x i32> %b128, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -147,19 +133,19 @@ define void @test_vXi32(<4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, <4 x
 }
 
 define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
-; AVX1-LABEL: 'test_vXi16'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 0 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi16'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 0 for instruction: ret void
+; SSE-LABEL: 'test_vXi16'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi16'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <16 x i16> %a256, <16 x i16> %b256, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <32 x i16> %a512, <32 x i16> %b512, <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 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 an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <8 x i16> %a128, <8 x i16> %b128, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
@@ -190,19 +176,19 @@ define void @test_vXi16(<8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, <8
 }
 
 define void @test_vXi8(<16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
-; AVX1-LABEL: 'test_vXi8'
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; AVX2-LABEL: 'test_vXi8'
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SSE-LABEL: 'test_vXi8'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX-LABEL: 'test_vXi8'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <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 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>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %V512_256 = shufflevector <32 x i8> %a256, <32 x i8> %b256, <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 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 an estimated cost of 0 for instruction: %V1024_512 = shufflevector <64 x i8> %a512, <64 x i8> %b512, <128 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 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, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
+; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_128 = shufflevector <16 x i8> %a128, <16 x i8> %b128, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
index 57f72056f9c8d9..4e4235198b45ed 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
@@ -143,19 +143,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -174,19 +174,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -205,19 +205,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -242,13 +242,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -273,13 +273,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -364,19 +364,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -395,19 +395,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -426,19 +426,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -463,13 +463,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -494,13 +494,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -595,17 +595,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -630,17 +630,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -665,17 +665,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
index ae91a70f67aee1..61978badb34d41 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
@@ -143,19 +143,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -174,19 +174,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -205,19 +205,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -242,13 +242,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -273,13 +273,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -364,19 +364,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -395,19 +395,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -426,19 +426,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -463,13 +463,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -494,13 +494,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -595,17 +595,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -630,17 +630,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -665,17 +665,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
index c84b2847cf3e97..94e3bc3a610bd7 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
@@ -143,19 +143,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -174,19 +174,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -205,19 +205,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -242,13 +242,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -273,13 +273,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -364,19 +364,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -395,19 +395,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -426,19 +426,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -463,13 +463,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -494,13 +494,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -595,17 +595,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -630,17 +630,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
@@ -665,17 +665,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
index c4cbd6141d1f0b..fe3e61d23397d3 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
@@ -143,19 +143,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -174,19 +174,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -205,19 +205,19 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -242,13 +242,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -273,13 +273,13 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -364,19 +364,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE2-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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE2-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>
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -395,19 +395,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSSE3-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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSSE3-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>
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -426,19 +426,19 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 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>
+; SSE42-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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 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>
+; SSE42-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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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 an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 56 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
@@ -463,13 +463,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX1-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>
 ; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX1-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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 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>
+; AVX1-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>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX1-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>
 ; AVX1-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>
@@ -494,13 +494,13 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX2-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>
 ; AVX2-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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
-; AVX2-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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 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>
 ; AVX2-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>
 ; AVX2-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>
@@ -595,17 +595,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-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>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 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>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
@@ -630,17 +630,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSSE3-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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-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>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
@@ -665,17 +665,17 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256_89 = 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 16, i32 17, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 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>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 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>
 ; SSE42-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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 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>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
index 6b2029e71a6e17..027af628ea3228 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-codesize.ll
@@ -3,15 +3,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE,SSSE3
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+xop | FileCheck %s -check-prefixes=XOP
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s -check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s -check-prefixes=AVX2
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s --check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 ;
 ; Verify the cost model for 2 src shuffles
@@ -20,30 +20,37 @@
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024, <2 x double> %src128_1, <4 x double> %src256_1, <8 x double> %src512_1, <16 x double> %src1024_1) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXf64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXf64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXf64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -56,30 +63,37 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512, <16 x i64> %src1024, <2 x i64> %src128_1, <4 x i64> %src256_1, <8 x i64> %src512_1, <16 x i64> %src1024_1) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXi64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXi64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXi64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -90,36 +104,52 @@ 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) {
-; SSE-LABEL: 'test_vXf32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXf32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXf32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXf32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXf32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXf32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
@@ -127,7 +157,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -139,36 +169,52 @@ 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, <32 x i32> %src1024, <2 x i32> %src64_1, <4 x i32> %src128_1, <8 x i32> %src256_1, <16 x i32> %src512_1, <32 x i32> %src1024_1) {
-; SSE-LABEL: 'test_vXi32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXi32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXi32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXi32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
@@ -176,7 +222,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -192,27 +238,27 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 224 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 960 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi16'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi16'
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi16'
@@ -220,8 +266,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi16'
@@ -229,8 +275,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 420 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi16'
@@ -238,8 +284,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 196 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
@@ -248,7 +294,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
@@ -257,7 +303,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512VBMI-LABEL: 'test_vXi16'
@@ -266,7 +312,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
@@ -284,8 +330,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 364 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi8'
@@ -293,8 +339,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi8'
@@ -302,8 +348,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi8'
@@ -312,7 +358,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi8'
@@ -321,7 +367,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi8'
@@ -330,7 +376,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-codesize.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-codesize.ll
index ae05855337b6cb..dc7e8292c6eef8 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-codesize.ll
@@ -6,7 +6,7 @@ define void @test_vXf16(<8 x half> %src128, <16 x half> %src256, <32 x half> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x half> %src256, <16 x half> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x half> %src512, <32 x half> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-latency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-latency.ll
index ca13511c32707b..8ad1fa46f59e66 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-latency.ll
@@ -6,7 +6,7 @@ define void @test_vXf16(<8 x half> %src128, <16 x half> %src256, <32 x half> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x half> %src256, <16 x half> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x half> %src512, <32 x half> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-sizelatency.ll
index 5312e8a87930e4..dc7f4c04d0f35a 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16-sizelatency.ll
@@ -6,7 +6,7 @@ define void @test_vXf16(<8 x half> %src128, <16 x half> %src256, <32 x half> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x half> %src256, <16 x half> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x half> %src512, <32 x half> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16.ll
index 9e444b6888a491..8ed785453d4e01 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-fp16.ll
@@ -6,7 +6,7 @@ define void @test_vXf16(<8 x half> %src128, <16 x half> %src256, <32 x half> %sr
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x half> %src256, <16 x half> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x half> %src512, <32 x half> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x half> %src1024, <64 x half> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V128 = shufflevector <8 x half> %src128, <8 x half> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
index 1c072479142632..f9f045f3a172bc 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-latency.ll
@@ -3,15 +3,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE,SSSE3
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+xop | FileCheck %s -check-prefixes=XOP
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s -check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX2
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s --check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 ;
 ; Verify the cost model for 2 src shuffles
@@ -20,30 +20,37 @@
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024, <2 x double> %src128_1, <4 x double> %src256_1, <8 x double> %src512_1, <16 x double> %src1024_1) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXf64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXf64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXf64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -56,30 +63,37 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512, <16 x i64> %src1024, <2 x i64> %src128_1, <4 x i64> %src256_1, <8 x i64> %src512_1, <16 x i64> %src1024_1) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXi64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXi64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXi64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -90,36 +104,52 @@ 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) {
-; SSE-LABEL: 'test_vXf32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXf32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXf32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXf32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXf32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXf32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
@@ -127,7 +157,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -139,36 +169,52 @@ 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, <32 x i32> %src1024, <2 x i32> %src64_1, <4 x i32> %src128_1, <8 x i32> %src256_1, <16 x i32> %src512_1, <32 x i32> %src1024_1) {
-; SSE-LABEL: 'test_vXi32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXi32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXi32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXi32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
@@ -176,7 +222,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -192,27 +238,27 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 224 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 960 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi16'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi16'
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi16'
@@ -220,8 +266,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi16'
@@ -229,8 +275,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 420 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi16'
@@ -238,8 +284,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 196 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
@@ -248,7 +294,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
@@ -257,7 +303,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512VBMI-LABEL: 'test_vXi16'
@@ -266,7 +312,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
@@ -284,8 +330,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 364 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi8'
@@ -293,8 +339,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi8'
@@ -302,8 +348,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi8'
@@ -312,7 +358,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi8'
@@ -321,7 +367,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi8'
@@ -330,7 +376,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
index 8f40881800687d..76690afecabdd4 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src-sizelatency.ll
@@ -3,15 +3,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE,SSSE3
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+xop | FileCheck %s -check-prefixes=XOP
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s -check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s -check-prefixes=AVX2
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s --check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 ;
 ; Verify the cost model for 2 src shuffles
@@ -20,30 +20,37 @@
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024, <2 x double> %src128_1, <4 x double> %src256_1, <8 x double> %src512_1, <16 x double> %src1024_1) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXf64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXf64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXf64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -56,30 +63,37 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512, <16 x i64> %src1024, <2 x i64> %src128_1, <4 x i64> %src256_1, <8 x i64> %src512_1, <16 x i64> %src1024_1) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXi64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; AVX1-LABEL: 'test_vXi64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXi64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -90,36 +104,52 @@ 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) {
-; SSE-LABEL: 'test_vXf32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXf32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXf32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXf32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXf32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXf32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
@@ -127,7 +157,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -139,36 +169,52 @@ 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, <32 x i32> %src1024, <2 x i32> %src64_1, <4 x i32> %src128_1, <8 x i32> %src256_1, <16 x i32> %src512_1, <32 x i32> %src1024_1) {
-; SSE-LABEL: 'test_vXi32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SSE2-LABEL: 'test_vXi32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXi32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXi32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
@@ -176,7 +222,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -192,27 +238,27 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 224 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 960 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi16'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi16'
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi16'
@@ -220,8 +266,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi16'
@@ -229,8 +275,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 420 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi16'
@@ -238,8 +284,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 196 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
@@ -248,7 +294,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
@@ -257,7 +303,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512VBMI-LABEL: 'test_vXi16'
@@ -266,7 +312,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
@@ -284,8 +330,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 364 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi8'
@@ -293,8 +339,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi8'
@@ -302,8 +348,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi8'
@@ -312,7 +358,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi8'
@@ -321,7 +367,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi8'
@@ -330,7 +376,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
diff --git a/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll b/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
index 32c06d4c44cecb..034ec0acf79d9f 100644
--- a/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
+++ b/llvm/test/Analysis/CostModel/X86/shuffle-two-src.ll
@@ -3,15 +3,15 @@
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+ssse3 | FileCheck %s -check-prefixes=SSE,SSSE3
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse4.2 | FileCheck %s -check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+xop | FileCheck %s -check-prefixes=XOP
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s -check-prefixes=AVX,AVX1
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s -check-prefixes=AVX,AVX2
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s -check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s -check-prefixes=AVX2
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512,AVX512F
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512,AVX512BW
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f,+avx512bw,+avx512vbmi | FileCheck %s --check-prefixes=AVX512,AVX512VBMI
 ;
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=slm | FileCheck %s --check-prefixes=SSE,SSE42
 ; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=goldmont | FileCheck %s --check-prefixes=SSE,SSE42
-; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX,AVX1
+; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 ;
 ; Verify the cost model for 2 src shuffles
@@ -20,30 +20,37 @@
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024, <2 x double> %src128_1, <4 x double> %src256_1, <8 x double> %src512_1, <16 x double> %src1024_1) {
 ; SSE-LABEL: 'test_vXf64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXf64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX1-LABEL: 'test_vXf64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXf64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V128 = shufflevector <2 x double> %src128, <2 x double> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -56,30 +63,37 @@ define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double>
 define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512, <16 x i64> %src1024, <2 x i64> %src128_1, <4 x i64> %src256_1, <8 x i64> %src512_1, <16 x i64> %src1024_1) {
 ; SSE-LABEL: 'test_vXi64'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi64'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
-; AVX-LABEL: 'test_vXi64'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; AVX1-LABEL: 'test_vXi64'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; AVX2-LABEL: 'test_vXi64'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi64'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> %src256_1, <4 x i32> <i32 3, i32 3, i32 7, i32 6>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> %src512_1, <8 x i32> <i32 7, i32 6, i32 12, i32 4, i32 3, i32 2, i32 1, i32 15>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <16 x i64> %src1024, <16 x i64> %src1024_1, <16 x i32> <i32 30, i32 14, i32 13, i32 12, i32 13, i32 10, i32 18, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V128 = shufflevector <2 x i64> %src128, <2 x i64> %src128_1, <2 x i32> <i32 3, i32 0>
@@ -90,36 +104,52 @@ 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) {
-; SSE-LABEL: 'test_vXf32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SSE2-LABEL: 'test_vXf32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXf32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXf32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXf32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXf32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXf32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXf32'
@@ -127,7 +157,7 @@ define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %sr
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x float> %src1024, <32 x float> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V64 = shufflevector <2 x float> %src64, <2 x float> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -139,36 +169,52 @@ 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, <32 x i32> %src1024, <2 x i32> %src64_1, <4 x i32> %src128_1, <8 x i32> %src256_1, <16 x i32> %src512_1, <32 x i32> %src1024_1) {
-; SSE-LABEL: 'test_vXi32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 240 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SSE2-LABEL: 'test_vXi32'
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSSE3-LABEL: 'test_vXi32'
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SSE42-LABEL: 'test_vXi32'
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
@@ -176,7 +222,7 @@ define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256,
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> %src128_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> %src256_1, <8 x i32> <i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 12, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> %src512_1, <16 x i32> <i32 15, i32 17, i32 13, i32 20, i32 11, i32 10, i32 8, i32 8, i32 7, i32 22, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> %src1024_1, <32 x i32> <i32 31, i32 33, i32 20, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 48, i32 13, i32 12, i32 11, i32 11, i32 9, i32 45, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V64 = shufflevector <2 x i32> %src64, <2 x i32> %src64_1, <2 x i32> <i32 3, i32 0>
@@ -192,27 +238,27 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 224 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 960 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 27 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi16'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi16'
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 360 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi16'
@@ -220,8 +266,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 26 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi16'
@@ -229,8 +275,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 420 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi16'
@@ -238,8 +284,8 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> %src64_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 196 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi16'
@@ -248,7 +294,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512F-NEXT:  Cost Model: Found an estimated cost of 252 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
@@ -257,7 +303,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512VBMI-LABEL: 'test_vXi16'
@@ -266,7 +312,7 @@ define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> %src128_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> %src256_1, <16 x i32> <i32 15, i32 14, i32 13, i32 20, i32 21, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> %src512_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 38, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
+; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> %src1024_1, <64 x i32> <i32 63, i32 62, i32 71, i32 60, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 66, i32 2, i32 1, i32 0>
 ; AVX512VBMI-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
   %V32 = shufflevector <2 x i16> %src32, <2 x i16> %src32_1, <2 x i32> <i32 3, i32 0>
@@ -284,8 +330,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 364 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SSSE3-LABEL: 'test_vXi8'
@@ -293,8 +339,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SSE42-LABEL: 'test_vXi8'
@@ -302,8 +348,8 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> %src32_1, <4 x i32> <i32 3, i32 6, i32 1, i32 5>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 84 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi8'
@@ -312,7 +358,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi8'
@@ -321,7 +367,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 90 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 23 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi8'
@@ -330,7 +376,7 @@ define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> %src64_1, <8 x i32> <i32 7, i32 6, i32 6, i32 8, i32 9, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> %src128_1, <16 x i32> <i32 29, i32 14, i32 28, i32 12, i32 11, i32 10, i32 11, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> %src256_1, <32 x i32> <i32 31, i32 30, i32 45, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 22, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 8, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 42 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> %src512_1, <64 x i32> <i32 63, i32 100, i32 61, i32 96, i32 59, i32 58, i32 57, i32 56, i32 55, i32 54, i32 53, i32 52, i32 51, i32 50, i32 49, i32 48, i32 47, i32 46, i32 45, i32 44, i32 43, i32 42, i32 41, i32 40, i32 39, i32 38, i32 37, i32 36, i32 35, i32 34, i32 33, i32 32, i32 31, i32 30, i32 29, i32 28, i32 27, i32 26, i32 25, i32 24, i32 23, i32 20, i32 21, i32 20, i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512F-LABEL: 'test_vXi8'
diff --git a/llvm/test/Transforms/PhaseOrdering/X86/pr94546.ll b/llvm/test/Transforms/PhaseOrdering/X86/pr94546.ll
index cbdafde5e421bd..354a988416c3ee 100644
--- a/llvm/test/Transforms/PhaseOrdering/X86/pr94546.ll
+++ b/llvm/test/Transforms/PhaseOrdering/X86/pr94546.ll
@@ -1,19 +1,32 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -O3                   -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -O3                   -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -O3                   -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -O3                   -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -passes="default<O3>" -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -passes="default<O3>" -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -passes="default<O3>" -S < %s  | FileCheck %s
-; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -passes="default<O3>" -S < %s  | FileCheck %s
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -O3                   -S < %s  | FileCheck %s --check-prefixes=SSE2
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -O3                   -S < %s  | FileCheck %s --check-prefixes=SSE4
+; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -O3                   -S < %s  | FileCheck %s --check-prefixes=AVX
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -O3                   -S < %s  | FileCheck %s --check-prefixes=AVX
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=SSE2
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=SSE4
+; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=AVX
+; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=AVX
 
 define <4 x double> @PR94546(<4 x double> %a, <4 x double> %b) {
-; CHECK-LABEL: @PR94546(
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <4 x i32> <i32 0, i32 poison, i32 poison, i32 6>
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <4 x i32> <i32 1, i32 poison, i32 poison, i32 7>
-; CHECK-NEXT:    [[TMP3:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
-; CHECK-NEXT:    ret <4 x double> [[TMP3]]
+; SSE2-LABEL: @PR94546(
+; SSE2-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <2 x i32> <i32 0, i32 6>
+; SSE2-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <2 x i32> <i32 1, i32 7>
+; SSE2-NEXT:    [[TMP3:%.*]] = fadd <2 x double> [[TMP1]], [[TMP2]]
+; SSE2-NEXT:    [[TMP4:%.*]] = shufflevector <2 x double> [[TMP3]], <2 x double> poison, <4 x i32> <i32 0, i32 poison, i32 poison, i32 1>
+; SSE2-NEXT:    ret <4 x double> [[TMP4]]
+;
+; SSE4-LABEL: @PR94546(
+; SSE4-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <4 x i32> <i32 0, i32 poison, i32 poison, i32 6>
+; SSE4-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <4 x i32> <i32 1, i32 poison, i32 poison, i32 7>
+; SSE4-NEXT:    [[TMP3:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
+; SSE4-NEXT:    ret <4 x double> [[TMP3]]
+;
+; AVX-LABEL: @PR94546(
+; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <4 x i32> <i32 0, i32 poison, i32 poison, i32 6>
+; AVX-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <4 x i32> <i32 1, i32 poison, i32 poison, i32 7>
+; AVX-NEXT:    [[TMP3:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
+; AVX-NEXT:    ret <4 x double> [[TMP3]]
 ;
   %vecext = extractelement <4 x double> %a, i32 0
   %vecext1 = extractelement <4 x double> %a, i32 1
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/horizontal-minmax.ll b/llvm/test/Transforms/SLPVectorizer/X86/horizontal-minmax.ll
index 0bc91d42b0f132..7626eea85f219e 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/horizontal-minmax.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/horizontal-minmax.ll
@@ -972,22 +972,15 @@ define i32 @maxi8_wrong_parent(i32) {
 ; SSE2-NEXT:    ret i32 [[TMP23]]
 ;
 ; SSE4-LABEL: @maxi8_wrong_parent(
-; SSE4-NEXT:    [[TMP2:%.*]] = load i32, ptr @arr, align 16
-; SSE4-NEXT:    [[TMP3:%.*]] = load i32, ptr getelementptr inbounds ([32 x i32], ptr @arr, i64 0, i64 1), align 4
+; SSE4-NEXT:    [[TMP2:%.*]] = load <2 x i32>, ptr @arr, align 16
 ; SSE4-NEXT:    br label [[PP:%.*]]
 ; SSE4:       pp:
 ; SSE4-NEXT:    [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([32 x i32], ptr @arr, i64 0, i64 2), align 8
-; SSE4-NEXT:    [[TMP5:%.*]] = load i32, ptr getelementptr inbounds ([32 x i32], ptr @arr, i64 0, i64 6), align 8
-; SSE4-NEXT:    [[TMP6:%.*]] = load i32, ptr getelementptr inbounds ([32 x i32], ptr @arr, i64 0, i64 7), align 4
-; SSE4-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> [[TMP4]])
-; SSE4-NEXT:    [[OP_RDX:%.*]] = icmp sgt i32 [[TMP7]], [[TMP5]]
-; SSE4-NEXT:    [[OP_RDX1:%.*]] = select i1 [[OP_RDX]], i32 [[TMP7]], i32 [[TMP5]]
-; SSE4-NEXT:    [[OP_RDX2:%.*]] = icmp sgt i32 [[TMP6]], [[TMP2]]
-; SSE4-NEXT:    [[OP_RDX3:%.*]] = select i1 [[OP_RDX2]], i32 [[TMP6]], i32 [[TMP2]]
-; SSE4-NEXT:    [[OP_RDX4:%.*]] = icmp sgt i32 [[OP_RDX1]], [[OP_RDX3]]
-; SSE4-NEXT:    [[OP_RDX5:%.*]] = select i1 [[OP_RDX4]], i32 [[OP_RDX1]], i32 [[OP_RDX3]]
-; SSE4-NEXT:    [[OP_RDX6:%.*]] = icmp sgt i32 [[OP_RDX5]], [[TMP3]]
-; SSE4-NEXT:    [[OP_RDX7:%.*]] = select i1 [[OP_RDX6]], i32 [[OP_RDX5]], i32 [[TMP3]]
+; SSE4-NEXT:    [[TMP8:%.*]] = load <2 x i32>, ptr getelementptr inbounds ([32 x i32], ptr @arr, i64 0, i64 6), align 8
+; SSE4-NEXT:    [[TMP5:%.*]] = call <8 x i32> @llvm.vector.insert.v8i32.v4i32(<8 x i32> poison, <4 x i32> [[TMP4]], i64 0)
+; SSE4-NEXT:    [[TMP6:%.*]] = call <8 x i32> @llvm.vector.insert.v8i32.v2i32(<8 x i32> [[TMP5]], <2 x i32> [[TMP8]], i64 4)
+; SSE4-NEXT:    [[TMP7:%.*]] = call <8 x i32> @llvm.vector.insert.v8i32.v2i32(<8 x i32> [[TMP6]], <2 x i32> [[TMP2]], i64 6)
+; SSE4-NEXT:    [[OP_RDX7:%.*]] = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> [[TMP7]])
 ; SSE4-NEXT:    ret i32 [[OP_RDX7]]
 ;
 ; AVX-LABEL: @maxi8_wrong_parent(
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/minbitwidth-transformed-operand.ll b/llvm/test/Transforms/SLPVectorizer/X86/minbitwidth-transformed-operand.ll
index 57b5d2af48ee60..76104efc1bb781 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/minbitwidth-transformed-operand.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/minbitwidth-transformed-operand.ll
@@ -5,12 +5,11 @@ define void @test(i64 %d.promoted.i) {
 ; CHECK-LABEL: define void @test(
 ; CHECK-SAME: i64 [[D_PROMOTED_I:%.*]]) {
 ; CHECK-NEXT:  entry:
-; CHECK-NEXT:    [[TMP0:%.*]] = insertelement <2 x i64> <i64 poison, i64 0>, i64 0, i32 0
 ; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> <i64 poison, i64 0>, i64 [[D_PROMOTED_I]], i32 0
-; CHECK-NEXT:    [[TMP2:%.*]] = and <2 x i64> [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP2:%.*]] = and <2 x i64> zeroinitializer, [[TMP1]]
 ; CHECK-NEXT:    [[TMP3:%.*]] = trunc <2 x i64> [[TMP2]] to <2 x i1>
-; CHECK-NEXT:    [[TMP4:%.*]] = shufflevector <2 x i1> [[TMP3]], <2 x i1> poison, <16 x i32> <i32 0, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 1, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <16 x i1> <i1 poison, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 poison, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <16 x i1> [[TMP4]], <16 x i32> <i32 16, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT:    [[TMP4:%.*]] = call <16 x i1> @llvm.vector.insert.v16i1.v2i1(<16 x i1> poison, <2 x i1> [[TMP3]], i64 0)
+; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <16 x i1> [[TMP4]], <16 x i1> <i1 poison, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 poison, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <16 x i32> <i32 0, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 1, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
 ; CHECK-NEXT:    [[TMP6:%.*]] = mul <16 x i1> [[TMP5]], zeroinitializer
 ; CHECK-NEXT:    [[TMP8:%.*]] = call i1 @llvm.vector.reduce.or.v16i1(<16 x i1> [[TMP6]])
 ; CHECK-NEXT:    [[TMP9:%.*]] = zext i1 [[TMP8]] to i32
diff --git a/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll
index 179e11136c8836..fba4b60ef417be 100644
--- a/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll
+++ b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-casts.ll
@@ -165,11 +165,17 @@ define <16 x float> @concat_fptrunc_v8f64_v16f32(<8 x double> %a0, <8 x double>
 ; commuted vector concatenation
 
 define <16 x i32> @rconcat_sext_v8i16_v16i32(<8 x i16> %a0, <8 x i16> %a1) {
-; CHECK-LABEL: define <16 x i32> @rconcat_sext_v8i16_v16i32(
-; CHECK-SAME: <8 x i16> [[A0:%.*]], <8 x i16> [[A1:%.*]]) #[[ATTR0]] {
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i16> [[A0]], <8 x i16> [[A1]], <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[R:%.*]] = sext <16 x i16> [[TMP1]] to <16 x i32>
-; CHECK-NEXT:    ret <16 x i32> [[R]]
+; SSE-LABEL: define <16 x i32> @rconcat_sext_v8i16_v16i32(
+; SSE-SAME: <8 x i16> [[A0:%.*]], <8 x i16> [[A1:%.*]]) #[[ATTR0]] {
+; SSE-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i16> [[A1]], <8 x i16> [[A0]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; SSE-NEXT:    [[R:%.*]] = sext <16 x i16> [[TMP1]] to <16 x i32>
+; SSE-NEXT:    ret <16 x i32> [[R]]
+;
+; AVX-LABEL: define <16 x i32> @rconcat_sext_v8i16_v16i32(
+; AVX-SAME: <8 x i16> [[A0:%.*]], <8 x i16> [[A1:%.*]]) #[[ATTR0]] {
+; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i16> [[A0]], <8 x i16> [[A1]], <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; AVX-NEXT:    [[R:%.*]] = sext <16 x i16> [[TMP1]] to <16 x i32>
+; AVX-NEXT:    ret <16 x i32> [[R]]
 ;
   %x0 = sext <8 x i16> %a0 to <8 x i32>
   %x1 = sext <8 x i16> %a1 to <8 x i32>
diff --git a/llvm/test/Transforms/VectorCombine/X86/shuffle-of-shuffles.ll b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-shuffles.ll
index 7dc70cd8b3d771..b30dc9ffdc5969 100644
--- a/llvm/test/Transforms/VectorCombine/X86/shuffle-of-shuffles.ll
+++ b/llvm/test/Transforms/VectorCombine/X86/shuffle-of-shuffles.ll
@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
-; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=sse2 | FileCheck %s
-; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=avx2 | FileCheck %s
+; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=sse2 | FileCheck %s --check-prefixes=CHECK,SSE
+; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=avx2 | FileCheck %s --check-prefixes=CHECK,AVX
 
 ; fold to identity
 
@@ -47,14 +47,21 @@ define <8 x i32> @concat_extract_subvectors_poison(<8 x i32> %x) {
 ; broadcast loads are free on AVX (and blends are much cheap than general 2-operand shuffles)
 
 define  <4 x double> @blend_broadcasts_v4f64(ptr %p0, ptr %p1)  {
-; CHECK-LABEL: define <4 x double> @blend_broadcasts_v4f64(
-; CHECK-SAME: ptr [[P0:%.*]], ptr [[P1:%.*]]) #[[ATTR0]] {
-; CHECK-NEXT:    [[LD0:%.*]] = load <4 x double>, ptr [[P0]], align 32
-; CHECK-NEXT:    [[LD1:%.*]] = load <4 x double>, ptr [[P1]], align 32
-; CHECK-NEXT:    [[BCST0:%.*]] = shufflevector <4 x double> [[LD0]], <4 x double> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[BCST1:%.*]] = shufflevector <4 x double> [[LD1]], <4 x double> undef, <4 x i32> zeroinitializer
-; CHECK-NEXT:    [[BLEND:%.*]] = shufflevector <4 x double> [[BCST0]], <4 x double> [[BCST1]], <4 x i32> <i32 0, i32 5, i32 6, i32 3>
-; CHECK-NEXT:    ret <4 x double> [[BLEND]]
+; SSE-LABEL: define <4 x double> @blend_broadcasts_v4f64(
+; SSE-SAME: ptr [[P0:%.*]], ptr [[P1:%.*]]) #[[ATTR0]] {
+; SSE-NEXT:    [[LD0:%.*]] = load <4 x double>, ptr [[P0]], align 32
+; SSE-NEXT:    [[LD1:%.*]] = load <4 x double>, ptr [[P1]], align 32
+; SSE-NEXT:    [[BLEND:%.*]] = shufflevector <4 x double> [[LD0]], <4 x double> [[LD1]], <4 x i32> <i32 0, i32 4, i32 4, i32 0>
+; SSE-NEXT:    ret <4 x double> [[BLEND]]
+;
+; AVX-LABEL: define <4 x double> @blend_broadcasts_v4f64(
+; AVX-SAME: ptr [[P0:%.*]], ptr [[P1:%.*]]) #[[ATTR0]] {
+; AVX-NEXT:    [[LD0:%.*]] = load <4 x double>, ptr [[P0]], align 32
+; AVX-NEXT:    [[LD1:%.*]] = load <4 x double>, ptr [[P1]], align 32
+; AVX-NEXT:    [[BCST0:%.*]] = shufflevector <4 x double> [[LD0]], <4 x double> undef, <4 x i32> zeroinitializer
+; AVX-NEXT:    [[BCST1:%.*]] = shufflevector <4 x double> [[LD1]], <4 x double> undef, <4 x i32> zeroinitializer
+; AVX-NEXT:    [[BLEND:%.*]] = shufflevector <4 x double> [[BCST0]], <4 x double> [[BCST1]], <4 x i32> <i32 0, i32 5, i32 6, i32 3>
+; AVX-NEXT:    ret <4 x double> [[BLEND]]
 ;
   %ld0 = load <4 x double>, ptr %p0, align 32
   %ld1 = load <4 x double>, ptr %p1, align 32

