[COST]Improve cost model for shuffles in SLP.

[ABataev]

Summary

Introduced masks where they are not added and improved target dependent
cost models to avoid returning of the incorrect cost results after
adding masks.

[RKSimon]

Adding some AArch64 reviewers

[fhahn]

Can you split off the target specific cost model changes? This makes it easier to track down potential regressions.

llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp

1408 ↗

(On Diff #337463)

This should be moved up I think, before the scalable vector handling. It would also be good to have cost-model tests for those shuffles.

[sdesmalen]

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4663–4664

Can the finding of a more specific ShuffleKind be done by getShuffleCost when a Mask is given?
It seems a bit inconvenient to have to do that manually before calling this function.

[RKSimon]

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4663–4664

I made a similar comment on D100495 - we could do with a generic 'ShuffleKind' decoder helper function (e.g. in Analysis\VectorUtils.h) that everybody can use.

[ABataev]

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4663–4664

In this case, we need to update getShuffleCost function for all the targets to translate the mask. Is it ok if I'll do it in this patch?

[sdesmalen]

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4663–4664

There's already a few different changes in this patch (AArch64 cost-model, X86 cost-model and changes to the SLPVectorizer), so I think it makes more sense to do this in a separate patch.

[ABataev]

llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp

1408 ↗

(On Diff #337463)

There is getIntrinsicInstrCost-vector-reverse.ll already affected by the cost changes.

[ABataev]

Rebase

[ABataev]

Rebase + fixes

[ABataev]

Rebase and fixes

[ABataev]

Rebase

[lebedev.ri]

Just because we've legalized to a certain vector type, is that sufficient guarantee that we have an appropriate shuffle operations on that legal vector type?

[ABataev]

In D100486#2773883, @lebedev.ri wrote:

Just because we've legalized to a certain vector type, is that sufficient guarantee that we have an appropriate shuffle operations on that legal vector type?

I think so, at least the existing code relies on this. This code just improves the existing situation. Currently, we conservatively calculate the cost of the shuffles as the permutation of all source subvectors. But in presence of mask we can estimate this cost more precisely, choosing only subvectors actually used in the permutations.

[ABataev]

Ping

[ABataev]

Rebase

[ABataev]

Rebase + fixed gathering cost calculation.

[RKSimon]

Adding @lebedev.ri who's worked on improving x86 shuffle/buildvector costmodels recently.

[lebedev.ri]

No i didn't.
I think while this may be somewhat correct,
is not really correct. For example, before AVX,
there is no non-32-bit shuffles, only unpacks.

[ABataev]

In D100486#2821347, @lebedev.ri wrote:

No i didn't.
I think while this may be somewhat correct,
is not really correct. For example, before AVX,
there is no non-32-bit shuffles, only unpacks.

Checked the cost model calculations and looks like it is correct. E.g. the cost of :

; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>

We have a permutation of 2 registers here. We have a general SK_PermuteSingleSrc for the first register <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 9, i32 8>:

SSE2
{TTI::SK_PermuteSingleSrc, MVT::v8i16, 5}, // 2*pshuflw + 2*pshufhw + pshufd/unpck

and we have SK_Reverse for the second register `<i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>:

{TTI::SK_Reverse, MVT::v8i16, 3}, // pshuflw + pshufhw + pshufd

. The sum result is 8.

[ABataev]

Rebase

[ABataev]

Added correct calculation of the cost for shuffle reuses.

[ABataev]

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

4376–4380

We incorrectly calculate the cost here (using the wrong vector type), with this fix and without this patch at least 4 X86 tests are not vectorized anymore.

llvm/test/Transforms/SLPVectorizer/AArch64/PR38339.ll

6–16 ↗

(On Diff #354335)

These regressions are caused by the incomplete cost model for AArch64, no cost for PermuteSingleSrc shuffle kind for VFxi16

[ABataev]

Rebase

[ABataev]

Rebase

[RKSimon]

Thanks for looking at this - this is well overdue!

llvm/lib/Target/X86/X86TargetTransformInfo.cpp

1223

We already have something similar in DAGTypeLegalizer::SplitVecRes_VECTOR_SHUFFLE - do you think we could have a single version of the code some place?

[ABataev]

llvm/lib/Target/X86/X86TargetTransformInfo.cpp

1223

Will check if can merge it somehow.

[ABataev]

Rebase + address comments.

Currently in hangs the compiler in tests test/CodeGen/Thumb2/mve-vst4.ll and llvm/test/CodeGen/Thumb2/mve-vst3.ll, need to investigate.

[RKSimon]

I hadn't been expecting you'd have to change the legalization implemention, at least not as an initial patch - can we not keep closer to the original SplitVecRes_VECTOR_SHUFFLE implementation?

[ABataev]

In D100486#2910065, @RKSimon wrote:

I hadn't been expecting you'd have to change the legalization implemention, at least not as an initial patch - can we not keep closer to the original SplitVecRes_VECTOR_SHUFFLE implementation?

Do you mean, use extract/inserts in case if we have more than 2 input vectors?

[ABataev]

Rebase + address comments.

[ABataev]

Rebase

[ABataev]

Rebase

[ABataev]

Rebase

[ABataev]

Ping

[ABataev]

Rebase

[ABataev]

Rebase

[ABataev]

Anything else here?

[ABataev]

Rebase

[ABataev]

Ping!

[ABataev]

Rebase

[ABataev]

Ping!

[ABataev]

Rebase

[RKSimon]

Sorry for the slow response on this one - please can you split the DAG legalization change out as the first patch and the TTI change as a followup?

llvm/include/llvm/Analysis/VectorUtils.h

401

Description comment?

402

Should this be splitShuffleMasks?

llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp

2141

I'm rather surprised this didn't reduce further (same for the TTI) - maybe there's more functionality that can be moved into the computeShuffleMasks helper?

[ABataev]

Rebase

[ABataev]

Rebase

[RKSimon]

llvm/test/Analysis/CostModel/X86/reduction.ll

64

Something might be still going wrong here - SSE max legal type is v4i32 so this cost should be 0 as its referencing a single existing vector

[ABataev]

llvm/test/Analysis/CostModel/X86/reduction.ll

64

I'm kind of pessimistic here and if the dest reg is not the same as src, I consider it as a reg copy and add cost TCC_Basic.

[RKSimon]

LGTM

[fhahn]

Just a heads up that with this change building llvm-test-suite/SingleSource/UnitTests/matrix-types-spec.cpp crashes on X86. I'll see if I can extract a reproducer.

[ABataev]

In D100486#3479989, @fhahn wrote:

Just a heads up that with this change building llvm-test-suite/SingleSource/UnitTests/matrix-types-spec.cpp crashes on X86. I'll see if I can extract a reproducer.

Hi, most probably the crash is caused by the opaque pointers. This patch does not change the vectorizer behavior, just adjusts the cost of the vector shuffles ops.

[fhahn]

In D100486#3479994, @ABataev wrote:

In D100486#3479989, @fhahn wrote:

Just a heads up that with this change building llvm-test-suite/SingleSource/UnitTests/matrix-types-spec.cpp crashes on X86. I'll see if I can extract a reproducer.

Hi, most probably the crash is caused by the opaque pointers. This patch does not change the vectorizer behavior, just adjusts the cost of the vector shuffles ops.

It crashes when calculating the cost with

Assertion failed: (idx < size()), function operator[], file SmallVector.h, line 273.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0.	Program arguments:  clang++ -DNDEBUG -O3 -O3 -DNDEBUG -isysroot ...
1.	<eof> parser at end of file
2.	Optimizer
Stack dump without symbol names (ensure you have llvm-symbolizer in your PATH or set the environment var `LLVM_SYMBOLIZER_PATH` to point to it):
0  clang-15                 0x000000010a03f8c7 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) + 39
1  clang-15                 0x000000010a03e6f8 llvm::sys::RunSignalHandlers() + 248
2  clang-15                 0x000000010a03ed20 llvm::sys::CleanupOnSignal(unsigned long) + 208
3  clang-15                 0x0000000109f6417a (anonymous namespace)::CrashRecoveryContextImpl::HandleCrash(int, unsigned long) + 106
4  clang-15                 0x0000000109f6435e CrashRecoverySignalHandler(int) + 110
5  libsystem_platform.dylib 0x00007ff808cfddfd _sigtramp + 29
6  libsystem_platform.dylib 000000000000000000 _sigtramp + 18446603370433094176
7  libsystem_c.dylib        0x00007ff808c33d24 abort + 123
8  libsystem_c.dylib        0x00007ff808c330cb err + 0
9  clang-15                 0x000000010d263f33 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) (.cold.9) + 35
10 clang-15                 0x00000001091758c4 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) + 2244
11 clang-15                 0x000000010886d07b llvm::X86TTIImpl::getShuffleCost(llvm::TargetTransformInfo::ShuffleKind, llvm::VectorType*, llvm::ArrayRef<int>, int, llvm::VectorType*, llvm::ArrayRef<llvm::Value const*>) + 1755
12 clang-15                 0x00000001088666a7 llvm::TargetTransformInfoImplCRTPBase<llvm::X86TTIImpl>::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) + 4551
13 clang-15                 0x000000010913f4b2 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) const + 18
14 clang-15                 0x0000000109d40056 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::TargetTransformInfo::TargetCostKind) const + 214
15 clang-15                 0x0000000108f97ed5 (anonymous namespace)::CallAnalyzer::visitInstruction(llvm::Instruction&) + 37

[ABataev]

In D100486#3480011, @fhahn wrote:

In D100486#3479994, @ABataev wrote:

In D100486#3479989, @fhahn wrote:

Just a heads up that with this change building llvm-test-suite/SingleSource/UnitTests/matrix-types-spec.cpp crashes on X86. I'll see if I can extract a reproducer.

Hi, most probably the crash is caused by the opaque pointers. This patch does not change the vectorizer behavior, just adjusts the cost of the vector shuffles ops.

It crashes when calculating the cost with

Assertion failed: (idx < size()), function operator[], file SmallVector.h, line 273.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0.	Program arguments:  clang++ -DNDEBUG -O3 -O3 -DNDEBUG -isysroot ...
1.	<eof> parser at end of file
2.	Optimizer
Stack dump without symbol names (ensure you have llvm-symbolizer in your PATH or set the environment var `LLVM_SYMBOLIZER_PATH` to point to it):
0  clang-15                 0x000000010a03f8c7 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) + 39
1  clang-15                 0x000000010a03e6f8 llvm::sys::RunSignalHandlers() + 248
2  clang-15                 0x000000010a03ed20 llvm::sys::CleanupOnSignal(unsigned long) + 208
3  clang-15                 0x0000000109f6417a (anonymous namespace)::CrashRecoveryContextImpl::HandleCrash(int, unsigned long) + 106
4  clang-15                 0x0000000109f6435e CrashRecoverySignalHandler(int) + 110
5  libsystem_platform.dylib 0x00007ff808cfddfd _sigtramp + 29
6  libsystem_platform.dylib 000000000000000000 _sigtramp + 18446603370433094176
7  libsystem_c.dylib        0x00007ff808c33d24 abort + 123
8  libsystem_c.dylib        0x00007ff808c330cb err + 0
9  clang-15                 0x000000010d263f33 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) (.cold.9) + 35
10 clang-15                 0x00000001091758c4 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) + 2244
11 clang-15                 0x000000010886d07b llvm::X86TTIImpl::getShuffleCost(llvm::TargetTransformInfo::ShuffleKind, llvm::VectorType*, llvm::ArrayRef<int>, int, llvm::VectorType*, llvm::ArrayRef<llvm::Value const*>) + 1755
12 clang-15                 0x00000001088666a7 llvm::TargetTransformInfoImplCRTPBase<llvm::X86TTIImpl>::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) + 4551
13 clang-15                 0x000000010913f4b2 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) const + 18
14 clang-15                 0x0000000109d40056 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::TargetTransformInfo::TargetCostKind) const + 214
15 clang-15                 0x0000000108f97ed5 (anonymous namespace)::CallAnalyzer::visitInstruction(llvm::Instruction&) + 37

Ok, the reproducer should help.

[ABataev]

In D100486#3480011, @fhahn wrote:

In D100486#3479994, @ABataev wrote:

In D100486#3479989, @fhahn wrote:

Just a heads up that with this change building llvm-test-suite/SingleSource/UnitTests/matrix-types-spec.cpp crashes on X86. I'll see if I can extract a reproducer.

Hi, most probably the crash is caused by the opaque pointers. This patch does not change the vectorizer behavior, just adjusts the cost of the vector shuffles ops.

It crashes when calculating the cost with

Assertion failed: (idx < size()), function operator[], file SmallVector.h, line 273.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0.	Program arguments:  clang++ -DNDEBUG -O3 -O3 -DNDEBUG -isysroot ...
1.	<eof> parser at end of file
2.	Optimizer
Stack dump without symbol names (ensure you have llvm-symbolizer in your PATH or set the environment var `LLVM_SYMBOLIZER_PATH` to point to it):
0  clang-15                 0x000000010a03f8c7 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) + 39
1  clang-15                 0x000000010a03e6f8 llvm::sys::RunSignalHandlers() + 248
2  clang-15                 0x000000010a03ed20 llvm::sys::CleanupOnSignal(unsigned long) + 208
3  clang-15                 0x0000000109f6417a (anonymous namespace)::CrashRecoveryContextImpl::HandleCrash(int, unsigned long) + 106
4  clang-15                 0x0000000109f6435e CrashRecoverySignalHandler(int) + 110
5  libsystem_platform.dylib 0x00007ff808cfddfd _sigtramp + 29
6  libsystem_platform.dylib 000000000000000000 _sigtramp + 18446603370433094176
7  libsystem_c.dylib        0x00007ff808c33d24 abort + 123
8  libsystem_c.dylib        0x00007ff808c330cb err + 0
9  clang-15                 0x000000010d263f33 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) (.cold.9) + 35
10 clang-15                 0x00000001091758c4 llvm::processShuffleMasks(llvm::ArrayRef<int>, unsigned int, unsigned int, unsigned int, llvm::function_ref<void ()>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>, llvm::function_ref<void (llvm::ArrayRef<int>, unsigned int, unsigned int)>) + 2244
11 clang-15                 0x000000010886d07b llvm::X86TTIImpl::getShuffleCost(llvm::TargetTransformInfo::ShuffleKind, llvm::VectorType*, llvm::ArrayRef<int>, int, llvm::VectorType*, llvm::ArrayRef<llvm::Value const*>) + 1755
12 clang-15                 0x00000001088666a7 llvm::TargetTransformInfoImplCRTPBase<llvm::X86TTIImpl>::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) + 4551
13 clang-15                 0x000000010913f4b2 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::ArrayRef<llvm::Value const*>, llvm::TargetTransformInfo::TargetCostKind) const + 18
14 clang-15                 0x0000000109d40056 llvm::TargetTransformInfo::getUserCost(llvm::User const*, llvm::TargetTransformInfo::TargetCostKind) const + 214
15 clang-15                 0x0000000108f97ed5 (anonymous namespace)::CallAnalyzer::visitInstruction(llvm::Instruction&) + 37

Could you just share your compile options for the source file compilation?

[fhahn]

Reproducer below crashes with bin/opt -passes="print<cost-model>"

target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-apple-darwin"

define void @test() {
entry:
  %matins.2.2 = shufflevector <9 x double> undef, <9 x double> undef, <9 x i32> <i32 0, i32 3, i32 6, i32 1, i32 4, i32 7, i32 2, i32 5, i32 8>
  ret void
}

[ABataev]

In D100486#3480073, @fhahn wrote:

Reproducer below crashes with bin/opt -passes="print<cost-model>"

target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-apple-darwin"

define void @test() {
entry:
  %matins.2.2 = shufflevector <9 x double> undef, <9 x double> undef, <9 x i32> <i32 0, i32 3, i32 6, i32 1, i32 4, i32 7, i32 2, i32 5, i32 8>
  ret void
}

Thanks, will revert the patch to fix the bug.

[vdmitrie]

Hi Alexey. Here is another crash reproducer:

; bin/opt -mcpu=corei7-avx -passes="print<cost-model>" -S test.ll

target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

define dso_local <12 x i64> @foo(<12 x i64> noundef %src) {
entry:

%shuffle = shufflevector <12 x i64> %src, <12 x i64> poison, <12 x i32> <i32 0, i32 3, i32 6, i32 9, i32 1, i32 4, i32 7, i32 10, i32 2, i32 5, i32 8, i32 11>
ret <12 x i64> %shuffle

}

[ABataev]

In D100486#3481545, @vdmitrie wrote:

Hi Alexey. Here is another crash reproducer:

; bin/opt -mcpu=corei7-avx -passes="print<cost-model>" -S test.ll

target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

define dso_local <12 x i64> @foo(<12 x i64> noundef %src) {
entry:

%shuffle = shufflevector <12 x i64> %src, <12 x i64> poison, <12 x i32> <i32 0, i32 3, i32 6, i32 9, i32 1, i32 4, i32 7, i32 10, i32 2, i32 5, i32 8, i32 11>
ret <12 x i64> %shuffle

}

Fixed in 371412e065a63107d5d79330da6757ff693d91cc

[asbirlea]

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

[ABataev]

In D100486#3494581, @asbirlea wrote:

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

Most probably, some overoptimistic cost model estimations turned on some extra vectorization. The reproducer will help, if you can provide it.

[asbirlea]

In D100486#3494746, @ABataev wrote:

In D100486#3494581, @asbirlea wrote:

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

Most probably, some overoptimistic cost model estimations turned on some extra vectorization. The reproducer will help, if you can provide it.

The reproducer is open source: https://gitlab.com/libeigen/eigen
Regressions are on haswell platform, xfdo configuration; the compiler used is bootstrapped against itself; revision tested includes this patch and all follow-up changes to the SLPVectorizer including D114171 and excluding D115750.
A few number samples:
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_16x64_ 16%
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_64x64_ 17%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_64x64_ 15%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_16x64_ 14%

[ABataev]

In D100486#3495157, @asbirlea wrote:

In D100486#3494746, @ABataev wrote:

In D100486#3494581, @asbirlea wrote:

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

Most probably, some overoptimistic cost model estimations turned on some extra vectorization. The reproducer will help, if you can provide it.

The reproducer is open source: https://gitlab.com/libeigen/eigen
Regressions are on haswell platform, xfdo configuration; the compiler used is bootstrapped against itself; revision tested includes this patch and all follow-up changes to the SLPVectorizer including D114171 and excluding D115750.
A few number samples:
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_16x64_ 16%
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_64x64_ 17%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_64x64_ 15%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_16x64_ 14%

Could you at least extract the llvm ir code for one of these functions before the changes and after, so I coukd at least compare them?

[asbirlea]

In D100486#3495174, @ABataev wrote:

In D100486#3495157, @asbirlea wrote:

In D100486#3494746, @ABataev wrote:

In D100486#3494581, @asbirlea wrote:

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

Most probably, some overoptimistic cost model estimations turned on some extra vectorization. The reproducer will help, if you can provide it.

The reproducer is open source: https://gitlab.com/libeigen/eigen
Regressions are on haswell platform, xfdo configuration; the compiler used is bootstrapped against itself; revision tested includes this patch and all follow-up changes to the SLPVectorizer including D114171 and excluding D115750.
A few number samples:
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_16x64_ 16%
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_64x64_ 17%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_64x64_ 15%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_16x64_ 14%

Could you at least extract the llvm ir code for one of these functions before the changes and after, so I coukd at least compare them?

Here are the only differences I see for dynamic, 64x64; base is before the patch, experiment includes this - revert+recommit -, https://reviews.llvm.org/rG371412e065a63107d5d79330da6757ff693d91cc, https://reviews.llvm.org/D114171 and the NFCs.
Looks like the change is in the presence of poison:

dumpslp_base_eigencomplex13 KBDownload

dumpslp_exp_eigencomplex13 KBDownload

Another 30-40% regression on a non-public benchmark was root-caused to https://reviews.llvm.org/D114171, so investigating that.

[ABataev]

In D100486#3504277, @asbirlea wrote:

In D100486#3495174, @ABataev wrote:

In D100486#3495157, @asbirlea wrote:

In D100486#3494746, @ABataev wrote:

In D100486#3494581, @asbirlea wrote:

We currently root-caused a 20% regression in eigen complex to this patch and it is blocking our compiler release.
I need some time to do additional testing, given there are at least a few patches on top of this. So far I see a crash fix, two NFCs and two additional changes that may also affect performance (https://reviews.llvm.org/D114171 and https://reviews.llvm.org/D115750).
Can you please hold off on any additional changes while we test further, or include them here for review so I can check if they resolve the regressions seen so far?

Most probably, some overoptimistic cost model estimations turned on some extra vectorization. The reproducer will help, if you can provide it.

The reproducer is open source: https://gitlab.com/libeigen/eigen
Regressions are on haswell platform, xfdo configuration; the compiler used is bootstrapped against itself; revision tested includes this patch and all follow-up changes to the SLPVectorizer including D114171 and excluding D115750.
A few number samples:
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_16x64_ 16%
BM_MatrixVectorMultiply_Complex_EigenRowMajorDynamic_float_64x64_ 17%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_64x64_ 15%
BM_MatrixVectorMultiply_Complex_EigenRowMajorFixed_float_16x64_ 14%

Could you at least extract the llvm ir code for one of these functions before the changes and after, so I coukd at least compare them?

Here are the only differences I see for dynamic, 64x64; base is before the patch, experiment includes this - revert+recommit -, https://reviews.llvm.org/rG371412e065a63107d5d79330da6757ff693d91cc, https://reviews.llvm.org/D114171 and the NFCs.
Looks like the change is in the presence of poison:

dumpslp_base_eigencomplex13 KBDownload

dumpslp_exp_eigencomplex13 KBDownload

Another 30-40% regression on a non-public benchmark was root-caused to https://reviews.llvm.org/D114171, so investigating that.

Hmm, I don't see the difference, actually, just changed the order of evaluation and lanes switched, no actual difference in IR. Most probably, the regression is caused by something else, maybe some lowering passes were tweaked for the previous order but not for the current.

[asbirlea]

Following up on the regression I was seeing with https://reviews.llvm.org/D114171, the only meaningful IR change seems to be:
Before the patch (after slp):

%65 = shufflevector <2 x float> %9, <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
%66 = insertelement <4 x float> %65, float %16, i32 2  
%67 = insertelement <4 x float> %66, float %15, i32 3
%68 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %67)
%69 = fcmp oeq <4 x float> %68, <float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000>
%70 = freeze <4 x i1> %69 
%71 = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> %70)

After the patch (after slp)

%65 = shufflevector <2 x float> %9, <2 x float> poison, <4 x i32> <i32 1, i32 0, i32 undef, i32 undef>  
%66 = shufflevector <4 x float> poison, <4 x float> %65, <4 x i32> <i32 4, i32 5, i32 2, i32 3> 
%67 = insertelement <4 x float> %66, float %16, i32 2  
%68 = insertelement <4 x float> %67, float %15, i32 3
%69 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %68)
%70 = fcmp oeq <4 x float> %69, <float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000>
%71 = freeze <4 x i1> %70  
%72 = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> %71)

I'm not clear why the order of 0,1 in the shufflevector changed or the purpose of the second shufflevector adding poison on positions that will be overwritten, but the operations that follow seem to not affected by these changes.
Just providing an update, and not digging further into it at this time.

[ABataev]

In D100486#3517448, @asbirlea wrote:

Following up on the regression I was seeing with https://reviews.llvm.org/D114171, the only meaningful IR change seems to be:
Before the patch (after slp):

%65 = shufflevector <2 x float> %9, <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
%66 = insertelement <4 x float> %65, float %16, i32 2  
%67 = insertelement <4 x float> %66, float %15, i32 3
%68 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %67)
%69 = fcmp oeq <4 x float> %68, <float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000>
%70 = freeze <4 x i1> %69 
%71 = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> %70)

After the patch (after slp)

%65 = shufflevector <2 x float> %9, <2 x float> poison, <4 x i32> <i32 1, i32 0, i32 undef, i32 undef>  
%66 = shufflevector <4 x float> poison, <4 x float> %65, <4 x i32> <i32 4, i32 5, i32 2, i32 3> 
%67 = insertelement <4 x float> %66, float %16, i32 2  
%68 = insertelement <4 x float> %67, float %15, i32 3
%69 = call <4 x float> @llvm.fabs.v4f32(<4 x float> %68)
%70 = fcmp oeq <4 x float> %69, <float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000, float 0x7FF0000000000000>
%71 = freeze <4 x i1> %70  
%72 = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> %71)

I'm not clear why the order of 0,1 in the shufflevector changed or the purpose of the second shufflevector adding poison on positions that will be overwritten, but the operations that follow seem to not affected by these changes.
Just providing an update, and not digging further into it at this time.

I'm still working on the reductions improvements. Hope to fix all these regressions.

llvm/include/llvm/Analysis/VectorUtils.h

 ///   <4 x i32> <3, 2, 0, -1>
 ///
 /// This is the reverse process of narrowing shuffle mask elements if it
 /// succeeds. This transform is not always possible because indexes may not
 /// divide evenly (scale down) to map to wider vector elements.
 bool widenShuffleMaskElts(int Scale, ArrayRef<int> Mask,
                           SmallVectorImpl<int> &ScaledMask);
 
+SmallVector<SmallVector<SmallVector<int>>>
+computeShuffleMasks(ArrayRef<int> Mask, unsigned NumOfSrcRegs,
+                    unsigned NumOfDestRegs);
+
 /// Compute a map of integer instructions to their minimum legal type
 /// size.
 ///
 /// C semantics force sub-int-sized values (e.g. i8, i16) to be promoted to int
 /// type (e.g. i32) whenever arithmetic is performed on them.
 ///
 /// For targets with native i8 or i16 operations, usually InstCombine can shrink
 /// the arithmetic type down again. However InstCombine refuses to create

llvm/lib/Analysis/VectorUtils.cpp

bool llvm::widenShuffleMaskElts(int Scale, ArrayRef<int> Mask,
 
   assert((int)ScaledMask.size() * Scale == NumElts && "Unexpected scaled mask");
 
   // All elements of the original mask can be scaled down to map to the elements
   // of a mask with wider elements.
   return true;
 }
 
+SmallVector<SmallVector<SmallVector<int>>>
+llvm::computeShuffleMasks(ArrayRef<int> Mask, unsigned NumOfSrcRegs,
+                          unsigned NumOfDestRegs) {
+  SmallVector<SmallVector<SmallVector<int>>> Res(NumOfDestRegs);
+  // Try to perform better estimation of the permutation.
+  // 1. Split the source/destination vectors into real registers.
+  // 2. Do the mask analysis to identify which real registers are
+  // permuted.
+  int Sz = Mask.size();
+  unsigned SzDest = Sz / NumOfDestRegs;
+  unsigned SzSrc = Sz / NumOfSrcRegs;
+  for (unsigned I = 0; I < NumOfDestRegs; ++I) {
+    int FirstReg = -1;
+    auto &RegMasks = Res[I];
+    RegMasks.assign(NumOfSrcRegs, {});
+    // Check that the values in dest registers are in the one src
+    // register.
+    for (unsigned K = 0; K < SzDest; ++K) {
+      int Idx = I * SzDest + K;
+      if (Idx == Sz)
+        break;
+      if (Mask[Idx] >= Sz || Mask[Idx] == UndefMaskElem)
+        continue;
+      int SrcRegIdx = Mask[Idx] / SzSrc;
+      // Add a cost of PermuteTwoSrc for each new source register permute,
+      // if we have more than one source registers.
+      if (RegMasks[SrcRegIdx].empty()) {
+        if (FirstReg < 0)
+          FirstReg = SrcRegIdx;
+        RegMasks[SrcRegIdx].assign(SzDest, UndefMaskElem);
+      }
+      RegMasks[SrcRegIdx][K] = Mask[Idx] % SzSrc;
+    }
+  }
+  return Res;
+}
+
 MapVector<Instruction *, uint64_t>
 llvm::computeMinimumValueSizes(ArrayRef<BasicBlock *> Blocks, DemandedBits &DB,
                                const TargetTransformInfo *TTI) {
 
   // DemandedBits will give us every value's live-out bits. But we want
   // to ensure no extra casts would need to be inserted, so every DAG
   // of connected values must have the same minimum bitwidth.
   EquivalenceClasses<Value *> ECs;

llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp

 // Splitting is the act of changing a computation in an invalid vector type to
 // be a computation in two vectors of half the size.  For example, implementing
 // <128 x f32> operations in terms of two <64 x f32> operations.
 //
 //===----------------------------------------------------------------------===//
 
 #include "LegalizeTypes.h"
 #include "llvm/Analysis/MemoryLocation.h"
+#include "llvm/Analysis/VectorUtils.h"
 #include "llvm/IR/DataLayout.h"
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/TypeSize.h"
 #include "llvm/Support/raw_ostream.h"
 using namespace llvm;
 
 #define DEBUG_TYPE "legalize-types"
 
void DAGTypeLegalizer::SplitVecRes_VECTOR_SHUFFLE(ShuffleVectorSDNode *N,
   GetSplitVector(N->getOperand(0), Inputs[0], Inputs[1]);
   GetSplitVector(N->getOperand(1), Inputs[2], Inputs[3]);
   EVT NewVT = Inputs[0].getValueType();
   unsigned NewElts = NewVT.getVectorNumElements();
 
   // If Lo or Hi uses elements from at most two of the four input vectors, then
   // express it as a vector shuffle of those two inputs.  Otherwise extract the
   // input elements by hand and construct the Lo/Hi output using a BUILD_VECTOR.
-  SmallVector<int, 16> Ops;
   for (unsigned High = 0; High < 2; ++High) {
     SDValue &Output = High ? Hi : Lo;
 
     // Build a shuffle mask for the output, discovering on the fly which
     // input vectors to use as shuffle operands (recorded in InputUsed).
-    // If building a suitable shuffle vector proves too hard, then bail
-    // out with useBuildVector set.
-    unsigned InputUsed[2] = { -1U, -1U }; // Not yet discovered.
     unsigned FirstMaskIdx = High * NewElts;
-    bool useBuildVector = false;
-    for (unsigned MaskOffset = 0; MaskOffset < NewElts; ++MaskOffset) {
-      // The mask element.  This indexes into the input.
-      int Idx = N->getMaskElt(FirstMaskIdx + MaskOffset);
-
-      // The input vector this mask element indexes into.
-      unsigned Input = (unsigned)Idx / NewElts;
-
-      if (Input >= array_lengthof(Inputs)) {
-        // The mask element does not index into any input vector.
-        Ops.push_back(-1);
-        continue;
-      }
-
-      // Turn the index into an offset from the start of the input vector.
-      Idx -= Input * NewElts;
-
-      // Find or create a shuffle vector operand to hold this input.
-      unsigned OpNo;
-      for (OpNo = 0; OpNo < array_lengthof(InputUsed); ++OpNo) {
-        if (InputUsed[OpNo] == Input) {
-          // This input vector is already an operand.
-          break;
-        } else if (InputUsed[OpNo] == -1U) {
-          // Create a new operand for this input vector.
-          InputUsed[OpNo] = Input;
-          break;
-        }
-      }
-
-      if (OpNo >= array_lengthof(InputUsed)) {
-        // More than two input vectors used!  Give up on trying to create a
-        // shuffle vector.  Insert all elements into a BUILD_VECTOR instead.
-        useBuildVector = true;
-        break;
-      }
-
-      // Add the mask index for the new shuffle vector.
-      Ops.push_back(Idx + OpNo * NewElts);
-    }
-
-    if (useBuildVector) {
+    SmallVector<int> Mask(NewElts * array_lengthof(Inputs), UndefMaskElem);
+    copy(N->getMask().slice(FirstMaskIdx, NewElts), Mask.begin());
+    SmallVector<SmallVector<SmallVector<int>>> DestVectors =
+        computeShuffleMasks(Mask, array_lengthof(Inputs),
+                            array_lengthof(Inputs));
+    assert(DestVectors.size() == array_lengthof(Inputs) &&
+           "Destination vectors and permutations not synced.");
+    auto &Dest = DestVectors.front();
+    int NumSrcRegs =
+        count_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+    if (NumSrcRegs > 2) {
+      // Use build vector currently.
+      // TODO: investigate if we can switch to shuffles for > 2 inputs.
       EVT EltVT = NewVT.getVectorElementType();
-      SmallVector<SDValue, 16> SVOps;
-
-      // Extract the input elements by hand.
-      for (unsigned MaskOffset = 0; MaskOffset < NewElts; ++MaskOffset) {
-        // The mask element.  This indexes into the input.
-        int Idx = N->getMaskElt(FirstMaskIdx + MaskOffset);
-
-        // The input vector this mask element indexes into.
-        unsigned Input = (unsigned)Idx / NewElts;
-
-        if (Input >= array_lengthof(Inputs)) {
-          // The mask element is "undef" or indexes off the end of the input.
-          SVOps.push_back(DAG.getUNDEF(EltVT));
-          continue;
+      SmallVector<SDValue> SVOps(NewElts);
+      for (unsigned Input = 0, E = Dest.size(); Input < E; ++Input) {
+        ArrayRef<int> Mask = Dest[Input];
+        for (int I = 0, E = Mask.size(); I < E; ++I)
+          if (Mask[I] != UndefMaskElem) {
+            assert(!SVOps[I] && "Expected no ops yet.");
+            SVOps[I] =
+                DAG.getNode(ISD::EXTRACT_VECTOR_ELT, dl, EltVT, Inputs[Input],
+                            DAG.getVectorIdxConstant(Mask[I], dl));
           }
-
-        // Turn the index into an offset from the start of the input vector.
-        Idx -= Input * NewElts;
-
-        // Extract the vector element by hand.
-        SVOps.push_back(DAG.getNode(ISD::EXTRACT_VECTOR_ELT, dl, EltVT,
-                                    Inputs[Input],
-                                    DAG.getVectorIdxConstant(Idx, dl)));
+      }
+      for (unsigned I = 0; I < NewElts; ++I) {
+        if (!SVOps[I])
+          SVOps[I] = DAG.getUNDEF(EltVT);
       }
 
       // Construct the Lo/Hi output using a BUILD_VECTOR.
       Output = DAG.getBuildVector(NewVT, dl, SVOps);
-    } else if (InputUsed[0] == -1U) {
+    } else if (NumSrcRegs == 0) {
       // No input vectors were used!  The result is undefined.
       Output = DAG.getUNDEF(NewVT);
+    } else if (NumSrcRegs == 1) {
+      // Find the only mask with at least single undef mask elem.
+      auto *It =
+          find_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+      Output = DAG.getVectorShuffle(NewVT, dl,
+                                    Inputs[std::distance(Dest.begin(), It)],
+                                    DAG.getUNDEF(NewVT), *It);
     } else {
-      SDValue Op0 = Inputs[InputUsed[0]];
-      // If only one input was used, use an undefined vector for the other.
-      SDValue Op1 = InputUsed[1] == -1U ?
-        DAG.getUNDEF(NewVT) : Inputs[InputUsed[1]];
-      // At least one input vector was used.  Create a new shuffle vector.
-      Output =  DAG.getVectorShuffle(NewVT, dl, Op0, Op1, Ops);
+      assert(NumSrcRegs > 1 && "Expected more than one input register.");
+      // Try to merge any masked elements.
+      auto *It =
+          find_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+      unsigned InitIdx = std::distance(Dest.begin(), It);
+      SDValue Op0 = Inputs[InitIdx];
+      auto *FirstNotInitIt =
+          find_if(drop_begin(Dest, InitIdx + 1),
+                  [](ArrayRef<int> Mask) { return !Mask.empty(); });
+      unsigned SecondIdx = std::distance(Dest.begin(), FirstNotInitIt);
+      SDValue Op1 = Inputs[SecondIdx];
+      auto &MergedMask = *It;
+      for (unsigned Idx = 0; Idx < NewElts; ++Idx) {
+        if ((*FirstNotInitIt)[Idx] != UndefMaskElem) {
+          assert(MergedMask[Idx] == UndefMaskElem &&
+                 "Expected undefined mask element.");
+          MergedMask[Idx] = (*FirstNotInitIt)[Idx] + NewElts;
+        }
+      }
+      Op0 = DAG.getVectorShuffle(NewVT, dl, Op0, Op1, MergedMask);
+      // At least two input vector were used.  Create shuffles.
+      for (unsigned I = SecondIdx + 1, E = array_lengthof(Inputs); I < E; ++I) {
+        MutableArrayRef<int> RegMask = Dest[I];
+        if (RegMask.empty())
+          continue;
+        for (unsigned Idx = 0; Idx < NewElts; ++Idx) {
+          if (RegMask[Idx] == UndefMaskElem) {
+            if (MergedMask[Idx] != UndefMaskElem)
+              RegMask[Idx] = Idx;
+          } else {
+            assert(MergedMask[Idx] == UndefMaskElem &&
+                   "Expected undefined mask element.");
+            RegMask[Idx] += NewElts;
+          }
+          MergedMask[Idx] = RegMask[Idx];
+        }
+        assert(RegMask.size() == NewElts &&
+               "Mask size does not match the vector length.");
+        Op0 = DAG.getVectorShuffle(NewVT, dl, Op0, Inputs[I], RegMask);
+      }
+      Output = Op0;
     }
-
-    Ops.clear();
   }
 }
 
 void DAGTypeLegalizer::SplitVecRes_VAARG(SDNode *N, SDValue &Lo, SDValue &Hi) {
   EVT OVT = N->getValueType(0);
   EVT NVT = OVT.getHalfNumVectorElementsVT(*DAG.getContext());
   SDValue Chain = N->getOperand(0);
   SDValue Ptr = N->getOperand(1);

llvm/lib/Target/X86/X86TargetTransformInfo.cpp

if (LegalVT.isVector() &&
       // Number of source vectors after legalization:
       unsigned NumOfSrcs = (VecTySize + LegalVTSize - 1) / LegalVTSize;
       // Number of destination vectors after legalization:
       InstructionCost NumOfDests = LT.first;
 
       auto *SingleOpTy = FixedVectorType::get(BaseTp->getElementType(),
                                               LegalVT.getVectorNumElements());
 
+      if (!Mask.empty() && NumOfDests.isValid()) {
+        // Try to perform better estimation of the permutation.
+        // 1. Split the source/destination vectors into real registers.
+        // 2. Do the mask analysis to identify which real registers are
+        // permuted. If more than 1 source registers are used for the
+        // destination register building, the cost for this destination register
+        // is (Number_of_source_register - 1) * Cost_PermuteTwoSrc. If only one
+        // source register is used, build mask and calculate the cost as a cost
+        // of PermuteSingleSrc.
+        // Also, for the single register permute we try to identify if the
+        // destination register is just a copy of the source register or the
+        // copy of the previous destination register (the cost is
+        // TTI::TCC_Basic). If the source register is just reused, the cost for
+        // this operation is 0.
+        unsigned E = *NumOfDests.getValue();
+        SmallVector<SmallVector<SmallVector<int>>> DestVectors =
+            computeShuffleMasks(Mask, NumOfSrcs, E);
+        assert(E == DestVectors.size() &&
+               "Destination vectors and permutations not synced.");
+        InstructionCost Cost = 0;
+        unsigned PrevSrcReg = 0;
+        ArrayRef<int> PrevRegMask;
+        for (unsigned I = 0; I < E; ++I) {
+          auto &Dest = DestVectors[I];
+          int NumSrcRegs =
+              count_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+          // NumSrcRegs might be 0 if the mask is just undef.
+          if (NumSrcRegs == 1) {
+            // Analysis of the single source register permutation.
+            auto *It =
+                find_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+            unsigned SrcReg = std::distance(Dest.begin(), It);
+            ArrayRef<int> RegMask = *It;
+            if (!ShuffleVectorInst::isIdentityMask(RegMask)) {
+              // Check if the previous register can be just copied to the next
+              // one.
+              if (PrevRegMask.empty() || PrevSrcReg != SrcReg ||
+                  PrevRegMask != RegMask)
+                Cost += getShuffleCost(TTI::SK_PermuteSingleSrc, SingleOpTy,
+                                       RegMask, 0, nullptr);
+              else
+                // Just a copy of previous destination register.
+                Cost += TTI::TCC_Basic;
+            } else if (SrcReg != I && any_of(RegMask, [](int I) {
+                         return I != UndefMaskElem;
+                       })) {
+              // Just a copy of the source register.
+              Cost += TTI::TCC_Basic;
+            }
+            PrevSrcReg = SrcReg;
+            PrevRegMask = RegMask;
+          } else if (NumSrcRegs > 1) {
+            // The first mask is a permutation of a single register, which can
+            // be transformed into a shuffle of 2 registers instead of 1 reg
+            // shuffle + 2 reg shuffles in case if we have 2+ input vectors.
+            auto *It =
+                find_if(Dest, [](ArrayRef<int> Mask) { return !Mask.empty(); });
+            auto *FirstNotInitIt =
+                find_if(drop_begin(Dest, std::distance(Dest.begin(), It) + 1),
+                        [](ArrayRef<int> Mask) { return !Mask.empty(); });
+            int VF = It->size();
+            auto &MergedMask = *It;
+            for (int Idx = 0; Idx < VF; ++Idx) {
+              if ((*FirstNotInitIt)[Idx] != UndefMaskElem) {
+                assert(MergedMask[Idx] == UndefMaskElem &&
+                       "Expected undefined mask element.");
+                MergedMask[Idx] = (*FirstNotInitIt)[Idx] + VF;
+              }
+            }
+            Cost += getShuffleCost(TTI::SK_PermuteTwoSrc, SingleOpTy,
+                                   MergedMask, 0, nullptr);
+            for (MutableArrayRef<int> RegMask :
+                 make_range(FirstNotInitIt + 1, Dest.end())) {
+              if (RegMask.empty())
+                continue;
+
+              for (int Idx = 0; Idx < VF; ++Idx) {
+                if (RegMask[Idx] == UndefMaskElem) {
+                  if (MergedMask[Idx] != UndefMaskElem)
+                    RegMask[Idx] = Idx;
+                } else {
+                  assert(MergedMask[Idx] == UndefMaskElem &&
+                         "Expected undefined mask element.");
+                  RegMask[Idx] += VF;
+                }
+                MergedMask[Idx] = RegMask[Idx];
+              }
+              Cost += getShuffleCost(TTI::SK_PermuteTwoSrc, SingleOpTy, RegMask,
+                                     0, nullptr);
+            }
+          }
+        }
+        return Cost;
+      }
+
       InstructionCost NumOfShuffles = (NumOfSrcs - 1) * NumOfDests;
       return NumOfShuffles * getShuffleCost(TTI::SK_PermuteTwoSrc, SingleOpTy,
                                             None, 0, nullptr);
     }
 
     return BaseT::getShuffleCost(Kind, BaseTp, Mask, Index, SubTp);
   }
 

llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

computeExtractCost(ArrayRef<Value *> VL, FixedVectorType *VecTy,
   bool AllConsecutive = true;
   unsigned EltsPerVector = VecTy->getNumElements() / NumOfParts;
   unsigned Idx = -1;
   InstructionCost Cost = 0;
 
   // Process extracts in blocks of EltsPerVector to check if the source vector
   // operand can be re-used directly. If not, add the cost of creating a shuffle
   // to extract the values into a vector register.
+  SmallVector<int> RegMask(EltsPerVector, UndefMaskElem);
   for (auto *V : VL) {
     ++Idx;
 
     // Reached the start of a new vector registers.
     if (Idx % EltsPerVector == 0) {
+      RegMask.assign(EltsPerVector, UndefMaskElem);
       AllConsecutive = true;
       continue;
     }
 
     // Check all extracts for a vector register on the target directly
     // extract values in order.
     unsigned CurrentIdx = *getExtractIndex(cast<Instruction>(V));
     unsigned PrevIdx = *getExtractIndex(cast<Instruction>(VL[Idx - 1]));
     AllConsecutive &= PrevIdx + 1 == CurrentIdx &&
                       CurrentIdx % EltsPerVector == Idx % EltsPerVector;
+    RegMask[Idx % EltsPerVector] = CurrentIdx % EltsPerVector;
 
     if (AllConsecutive)
       continue;
 
     // Skip all indices, except for the last index per vector block.
     if ((Idx + 1) % EltsPerVector != 0 && Idx + 1 != VL.size())
       continue;
 
     // If we have a series of extracts which are not consecutive and hence
     // cannot re-use the source vector register directly, compute the shuffle
     // cost to extract the a vector with EltsPerVector elements.
     Cost += TTI.getShuffleCost(
         TargetTransformInfo::SK_PermuteSingleSrc,
-        FixedVectorType::get(VecTy->getElementType(), EltsPerVector));
+        FixedVectorType::get(VecTy->getElementType(), EltsPerVector), RegMask);
   }
   return Cost;
 }
 
 /// Build shuffle mask for shuffle graph entries and lists of main and alternate
 /// operations operands.
 static void
 buildSuffleEntryMask(ArrayRef<Value *> VL, ArrayRef<unsigned> ReorderIndices,
InstructionCost BoUpSLP::getEntryCost(const TreeEntry *E,
   // that the costs will be accurate.
   if (MinBWs.count(VL[0]))
     VecTy = FixedVectorType::get(
         IntegerType::get(F->getContext(), MinBWs[VL[0]].first), VL.size());
   auto *FinalVecTy = VecTy;
 
   unsigned ReuseShuffleNumbers = E->ReuseShuffleIndices.size();
   bool NeedToShuffleReuses = !E->ReuseShuffleIndices.empty();
   if (NeedToShuffleReuses)
     FinalVecTy =
         FixedVectorType::get(VecTy->getElementType(), ReuseShuffleNumbers);
   // FIXME: it tries to fix a problem with MSVC buildbots.
   TargetTransformInfo &TTIRef = *TTI;
   auto &&AdjustExtractsCost = [this, &TTIRef, CostKind, VL, VecTy,
                                VectorizedVals](InstructionCost &Cost,
                                                bool IsGather) {
     DenseMap<Value *, int> ExtractVectorsTys;
     for (auto *V : VL) {
       // If all users of instruction are going to be vectorized and this
       // instruction itself is not going to be vectorized, consider this
       // instruction as dead and remove its cost from the final cost of the
case Instruction::ExtractElement: {
                                                   *getExtractIndex(EE));
           } else {
             --Idx;
             CommonCost += TTI->getVectorInstrCost(Instruction::ExtractElement,
                                                   VecTy, Idx);
           }
         }
       }
       if (ShuffleOrOp == Instruction::ExtractValue) {
         for (unsigned I = 0, E = VL.size(); I < E; ++I) {
           auto *EI = cast<Instruction>(VL[I]);
           // Take credit for instruction that will become dead.
           if (EI->hasOneUse()) {
             Instruction *Ext = EI->user_back();
             if ((isa<SExtInst>(Ext) || isa<ZExtInst>(Ext)) &&
                 all_of(Ext->users(),
                        [](User *U) { return isa<GetElementPtrInst>(U); })) {
               // Use getExtractWithExtendCost() to calculate the cost of
case Instruction::ShuffleVector: {
         auto *Src0Ty = FixedVectorType::get(Src0SclTy, VL.size());
         auto *Src1Ty = FixedVectorType::get(Src1SclTy, VL.size());
         VecCost = TTI->getCastInstrCost(E->getOpcode(), VecTy, Src0Ty,
                                         TTI::CastContextHint::None, CostKind);
         VecCost += TTI->getCastInstrCost(E->getAltOpcode(), VecTy, Src1Ty,
                                          TTI::CastContextHint::None, CostKind);
       }
 
+      if (E->ReuseShuffleIndices.empty()) {
+        CommonCost =
+            TTI->getShuffleCost(TargetTransformInfo::SK_Select, FinalVecTy);
+      } else {
         SmallVector<int> Mask;
         buildSuffleEntryMask(
             E->Scalars, E->ReorderIndices, E->ReuseShuffleIndices,
             [E](Instruction *I) {
               assert(E->isOpcodeOrAlt(I) && "Unexpected main/alternate opcode");
               return I->getOpcode() == E->getAltOpcode();
             },
             Mask);
-      CommonCost =
-          TTI->getShuffleCost(TargetTransformInfo::SK_Select, FinalVecTy, Mask);
+        CommonCost = TTI->getShuffleCost(TargetTransformInfo::SK_PermuteTwoSrc,
+                                         FinalVecTy, Mask);
+      }
       LLVM_DEBUG(dumpTreeCosts(E, CommonCost, VecCost, ScalarCost));
       return CommonCost + VecCost - ScalarCost;
     }
     default:
       llvm_unreachable("Unknown instruction");
   }
 }
 
for (ExternalUser &EU : ExternalUses) {
     } else {
       ExtractCost +=
           TTI->getVectorInstrCost(Instruction::ExtractElement, VecTy, EU.Lane);
     }
   }
 
   InstructionCost SpillCost = getSpillCost();
   Cost += SpillCost + ExtractCost;
-  for (int I = 0, E = FirstUsers.size(); I < E; ++I) {
-    // For the very first element - simple shuffle of the source vector.
-    int Limit = ShuffleMask[I].size() * 2;
-    if (I == 0 &&
-        all_of(ShuffleMask[I], [Limit](int Idx) { return Idx < Limit; }) &&
-        !ShuffleVectorInst::isIdentityMask(ShuffleMask[I])) {
+  if (FirstUsers.size() == 1) {
+    int Limit = ShuffleMask.front().size() * 2;
+    if (!all_of(ShuffleMask.front(),
+                [Limit](int Idx) { return Idx < Limit; }) ||
+        !ShuffleVectorInst::isIdentityMask(ShuffleMask.front())) {
       InstructionCost C = TTI->getShuffleCost(
           TTI::SK_PermuteSingleSrc,
-          cast<FixedVectorType>(FirstUsers[I]->getType()), ShuffleMask[I]);
+          cast<FixedVectorType>(FirstUsers.front()->getType()),
+          ShuffleMask.front());
       LLVM_DEBUG(dbgs() << "SLP: Adding cost " << C
                         << " for final shuffle of insertelement external users "
                         << *VectorizableTree.front()->Scalars.front() << ".\n"
                         << "SLP: Current total cost = " << Cost << "\n");
       Cost += C;
-      continue;
+      InstructionCost InsertCost = TTI->getScalarizationOverhead(
+          cast<FixedVectorType>(FirstUsers.front()->getType()),
+          DemandedElts.front(), /*Insert*/ true, /*Extract*/ false);
+      LLVM_DEBUG(dbgs() << "SLP: Subtracting the cost " << InsertCost
+                        << " for insertelements gather.\n"
+                        << "SLP: Current total cost = " << Cost << "\n");
+      Cost -= InsertCost;
+    }
+  } else if (FirstUsers.size() >= 2) {
+    // Combined masks of the first 2 vectors.
+    auto &CombinedMask = ShuffleMask.front();
+    unsigned VF = CombinedMask.size();
+    auto &CombinedDemandedElts = DemandedElts.front();
+    if (VF < ShuffleMask[1].size())
+      CombinedMask.resize(ShuffleMask[1].size(), UndefMaskElem);
+    for (int I = 0, E = ShuffleMask[1].size(); I < E; ++I) {
+      if (ShuffleMask[1][I] != UndefMaskElem) {
+        assert(CombinedMask[I] == UndefMaskElem &&
+               "Expected undefined mask element.");
+        CombinedMask[I] = ShuffleMask[1][I] + VF;
+        CombinedDemandedElts.setBit(I);
+      }
     }
-    // Other elements - permutation of 2 vectors (the initial one and the next
-    // Ith incoming vector).
+    InstructionCost C = TTI->getShuffleCost(
+        TTI::SK_PermuteTwoSrc,
+        cast<FixedVectorType>(FirstUsers.front()->getType()), CombinedMask);
+    LLVM_DEBUG(dbgs() << "SLP: Adding cost " << C
+                      << " for final shuffle of vector node and external "
+                         "insertelement users "
+                      << *VectorizableTree.front()->Scalars.front() << ".\n"
+                      << "SLP: Current total cost = " << Cost << "\n");
+    Cost += C;
+    InstructionCost InsertCost = TTI->getScalarizationOverhead(
+        cast<FixedVectorType>(FirstUsers.front()->getType()),
+        CombinedDemandedElts, /*Insert*/ true, /*Extract*/ false);
+    LLVM_DEBUG(dbgs() << "SLP: Subtracting the cost " << InsertCost
+                      << " for insertelements gather.\n"
+                      << "SLP: Current total cost = " << Cost << "\n");
+    Cost -= InsertCost;
+    for (int I = 2, E = FirstUsers.size(); I < E; ++I) {
+      if (ShuffleMask[I].empty())
+        continue;
+      // Other elements - permutation of 2 vectors (the initial one and the
+      // next Ith incoming vector).
       unsigned VF = ShuffleMask[I].size();
+      if (CombinedMask.size() < VF)
+        CombinedMask.resize(VF, UndefMaskElem);
       for (unsigned Idx = 0; Idx < VF; ++Idx) {
         int &Mask = ShuffleMask[I][Idx];
-      Mask = Mask == UndefMaskElem ? Idx : VF + Mask;
+        if (Mask == UndefMaskElem) {
+          if (CombinedMask[Idx] != UndefMaskElem)
+            Mask = Idx;
+        } else {
+          assert(CombinedMask[Idx] == UndefMaskElem &&
+                 "Expected undefined mask element.");
+          Mask += VF;
+        }
+        CombinedMask[Idx] = Mask;
       }
       InstructionCost C = TTI->getShuffleCost(
-        TTI::SK_PermuteTwoSrc, cast<FixedVectorType>(FirstUsers[I]->getType()),
-        ShuffleMask[I]);
-    LLVM_DEBUG(
-        dbgs()
-        << "SLP: Adding cost " << C
-        << " for final shuffle of vector node and external insertelement users "
+          TTI::SK_PermuteTwoSrc,
+          cast<FixedVectorType>(FirstUsers[I]->getType()), ShuffleMask[I]);
+      LLVM_DEBUG(dbgs() << "SLP: Adding cost " << C
+                        << " for final shuffle of vector node and external "
+                           "insertelement users "
                         << *VectorizableTree.front()->Scalars.front() << ".\n"
                         << "SLP: Current total cost = " << Cost << "\n");
       Cost += C;
       InstructionCost InsertCost = TTI->getScalarizationOverhead(
           cast<FixedVectorType>(FirstUsers[I]->getType()), DemandedElts[I],
-        /*Insert*/ true,
-        /*Extract*/ false);
-    Cost -= InsertCost;
+          /*Insert*/ true, /*Extract*/ false);
       LLVM_DEBUG(dbgs() << "SLP: subtracting the cost " << InsertCost
                         << " for insertelements gather.\n"
                         << "SLP: Current total cost = " << Cost << "\n");
+      Cost -= InsertCost;
+    }
   }
 
 #ifndef NDEBUG
   SmallString<256> Str;
   {
     raw_svector_ostream OS(Str);
     OS << "SLP: Spill Cost = " << SpillCost << ".\n"
        << "SLP: Extract Cost = " << ExtractCost << ".\n"

llvm/test/Analysis/CostModel/X86/reduction.ll

;
   %bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7
 
   %r = extractelement <4 x float> %bin.rdx8, i32 0
   ret float %r
 }
 
 define fastcc i32 @reduction_cost_int(<8 x i32> %rdx) {
 ; SSE-LABEL: 'reduction_cost_int'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %rdx, %rdx.shuf
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx.2 = add <8 x i32> %bin.rdx, %rdx.shuf.2
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.3 = shufflevector <8 x i32> %bin.rdx.2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.3 = shufflevector <8 x i32> %bin.rdx.2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx.3 = add <8 x i32> %bin.rdx.2, %rdx.shuf.3
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx.3, i32 0
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; AVX1-LABEL: 'reduction_cost_int'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = add <8 x i32> %rdx, %rdx.shuf
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx.2 = add <8 x i32> %bin.rdx, %rdx.shuf.2
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.3 = shufflevector <8 x i32> %bin.rdx.2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx.3 = add <8 x i32> %bin.rdx.2, %rdx.shuf.3
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx.3, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; SLM-LABEL: 'reduction_cost_int'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %rdx, %rdx.shuf
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx.2 = add <8 x i32> %bin.rdx, %rdx.shuf.2
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.3 = shufflevector <8 x i32> %bin.rdx.2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.3 = shufflevector <8 x i32> %bin.rdx.2, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx.3 = add <8 x i32> %bin.rdx.2, %rdx.shuf.3
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx.3, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
   %rdx.shuf = shufflevector <8 x i32> %rdx, <8 x i32> undef,
   <8 x i32> <i32 4    , i32     5, i32     6, i32     7,
   i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx = add <8 x i32> %rdx, %rdx.shuf
;
   %bin.rdx8 = fadd <4 x float> %bin.rdx, %rdx.shuf7
 
   %r = extractelement <4 x float> %bin.rdx8, i32 0
   ret float %r
 }
 
 define fastcc double @no_pairwise_reduction4double(<4 x double> %rdx, double %f1) {
 ; SSE2-LABEL: 'no_pairwise_reduction4double'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SSSE3-LABEL: 'no_pairwise_reduction4double'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SSE42-LABEL: 'no_pairwise_reduction4double'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; AVX1-LABEL: 'no_pairwise_reduction4double'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; AVX2-LABEL: 'no_pairwise_reduction4double'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SLM-LABEL: 'no_pairwise_reduction4double'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
   %rdx.shuf = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
   %bin.rdx = fadd <4 x double> %rdx, %rdx.shuf
   %rdx.shuf7 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
   %bin.rdx8 = fadd <4 x double> %bin.rdx, %rdx.shuf7
 
   %r = extractelement <4 x double> %bin.rdx8, i32 0
   ret double %r
 }
 
 define fastcc float @no_pairwise_reduction8float(<8 x float> %rdx, float %f1) {
 ; SSE2-LABEL: 'no_pairwise_reduction8float'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <8 x float> %bin.rdx, %rdx.shuf7
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx8, i32 0
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SSSE3-LABEL: 'no_pairwise_reduction8float'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <8 x float> %bin.rdx, %rdx.shuf7
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx8, i32 0
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SSE42-LABEL: 'no_pairwise_reduction8float'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <8 x float> %bin.rdx, %rdx.shuf7
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx8, i32 0
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; AVX1-LABEL: 'no_pairwise_reduction8float'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = fadd <8 x float> %bin.rdx, %rdx.shuf7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SLM-LABEL: 'no_pairwise_reduction8float'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <8 x float> %bin.rdx, %rdx.shuf7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
   %rdx.shuf3 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7,i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx4 = fadd <8 x float> %rdx, %rdx.shuf3
   %rdx.shuf = shufflevector <8 x float> %bin.rdx4, <8 x float> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx = fadd <8 x float> %bin.rdx4, %rdx.shuf
;
   %bin.rdx8 = add <4 x i32> %bin.rdx, %rdx.shuf7
 
   %r = extractelement <4 x i32> %bin.rdx8, i32 0
   ret i32 %r
 }
 
 define fastcc i64 @no_pairwise_reduction4i64(<4 x i64> %rdx, i64 %f1) {
 ; SSE-LABEL: 'no_pairwise_reduction4i64'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <4 x i64> %rdx, %rdx.shuf
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <4 x i64> %bin.rdx, %rdx.shuf7
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
 ; AVX1-LABEL: 'no_pairwise_reduction4i64'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = add <4 x i64> %rdx, %rdx.shuf
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = add <4 x i64> %bin.rdx, %rdx.shuf7
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
 ; AVX2-LABEL: 'no_pairwise_reduction4i64'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = add <4 x i64> %rdx, %rdx.shuf
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = add <4 x i64> %bin.rdx, %rdx.shuf7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
 ; SLM-LABEL: 'no_pairwise_reduction4i64'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %bin.rdx = add <4 x i64> %rdx, %rdx.shuf
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %bin.rdx8 = add <4 x i64> %bin.rdx, %rdx.shuf7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
   %rdx.shuf = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
   %bin.rdx = add <4 x i64> %rdx, %rdx.shuf
   %rdx.shuf7 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
   %bin.rdx8 = add <4 x i64> %bin.rdx, %rdx.shuf7
;
   %bin.rdx8 = add <8 x i16> %bin.rdx, %rdx.shuf7
 
   %r = extractelement <8 x i16> %bin.rdx8, i32 0
   ret i16 %r
 }
 
 define fastcc i32 @no_pairwise_reduction8i32(<8 x i32> %rdx, i32 %f1) {
 ; SSE-LABEL: 'no_pairwise_reduction8i32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx4 = add <8 x i32> %rdx, %rdx.shuf3
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %bin.rdx4, %rdx.shuf
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <8 x i32> %bin.rdx, %rdx.shuf7
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx8, i32 0
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; AVX1-LABEL: 'no_pairwise_reduction8i32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx4 = add <8 x i32> %rdx, %rdx.shuf3
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = add <8 x i32> %bin.rdx4, %rdx.shuf
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = add <8 x i32> %bin.rdx, %rdx.shuf7
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; SLM-LABEL: 'no_pairwise_reduction8i32'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx4 = add <8 x i32> %rdx, %rdx.shuf3
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %bin.rdx4, %rdx.shuf
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf7 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf7 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <8 x i32> %bin.rdx, %rdx.shuf7
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
   %rdx.shuf3 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7,i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx4 = add <8 x i32> %rdx, %rdx.shuf3
   %rdx.shuf = shufflevector <8 x i32> %bin.rdx4, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx = add <8 x i32> %bin.rdx4, %rdx.shuf
;
   %bin.rdx8 = fadd <4 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 
   %r = extractelement <4 x float> %bin.rdx8, i32 0
   ret float %r
 }
 
 define fastcc double @pairwise_reduction4double(<4 x double> %rdx, double %f1) {
 ; SSE2-LABEL: 'pairwise_reduction4double'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SSSE3-LABEL: 'pairwise_reduction4double'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SSE42-LABEL: 'pairwise_reduction4double'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; AVX1-LABEL: 'pairwise_reduction4double'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
 ; SLM-LABEL: 'pairwise_reduction4double'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <4 x double> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret double %r
 ;
   %rdx.shuf.0.0 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
   %rdx.shuf.0.1 = shufflevector <4 x double> %rdx, <4 x double> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
   %bin.rdx = fadd <4 x double> %rdx.shuf.0.0, %rdx.shuf.0.1
   %rdx.shuf.1.0 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
   %rdx.shuf.1.1 = shufflevector <4 x double> %bin.rdx, <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
   %bin.rdx8 = fadd <4 x double> %rdx.shuf.1.0, %rdx.shuf.1.1
 
   %r = extractelement <4 x double> %bin.rdx8, i32 0
   ret double %r
 }
 
 define fastcc float @pairwise_reduction8float(<8 x float> %rdx, float %f1) {
 ; SSE2-LABEL: 'pairwise_reduction8float'
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <8 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx9 = fadd <8 x float> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx9, i32 0
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SSSE3-LABEL: 'pairwise_reduction8float'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx8 = fadd <8 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx9 = fadd <8 x float> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx9, i32 0
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SSE42-LABEL: 'pairwise_reduction8float'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <8 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx9 = fadd <8 x float> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx9, i32 0
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; AVX1-LABEL: 'pairwise_reduction8float'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = fadd <8 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx9 = fadd <8 x float> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx9, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
 ; SLM-LABEL: 'pairwise_reduction8float'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x float> %bin.rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = fadd <8 x float> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x float> %bin.rdx8, <8 x float> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx9 = fadd <8 x float> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %r = extractelement <8 x float> %bin.rdx9, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret float %r
 ;
   %rdx.shuf.0.0 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6,i32 undef, i32 undef, i32 undef, i32 undef>
   %rdx.shuf.0.1 = shufflevector <8 x float> %rdx, <8 x float> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7,i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx = fadd <8 x float> %rdx.shuf.0.0, %rdx.shuf.0.1
   %rdx.shuf.1.0 = shufflevector <8 x float> %bin.rdx, <8 x float> undef,<8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
;
   %bin.rdx8 = add <4 x i32> %rdx.shuf.1.0, %rdx.shuf.1.1
 
   %r = extractelement <4 x i32> %bin.rdx8, i32 0
   ret i32 %r
 }
 
 define fastcc i64 @pairwise_reduction4i64(<4 x i64> %rdx, i64 %f1) {
 ; SSE-LABEL: 'pairwise_reduction4i64'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <4 x i64> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <4 x i64> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
 ; AVX1-LABEL: 'pairwise_reduction4i64'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = add <4 x i64> %rdx.shuf.0.0, %rdx.shuf.0.1

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx = add <4 x i64> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = add <4 x i64> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
 ; SLM-LABEL: 'pairwise_reduction4i64'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %bin.rdx = add <4 x i64> %rdx.shuf.0.0, %rdx.shuf.0.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.1.0 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.1.1 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %bin.rdx8 = add <4 x i64> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <4 x i64> %bin.rdx8, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i64 %r
 ;
   %rdx.shuf.0.0 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 undef, i32 undef>
   %rdx.shuf.0.1 = shufflevector <4 x i64> %rdx, <4 x i64> undef, <4 x i32> <i32 1, i32 3, i32 undef, i32 undef>
   %bin.rdx = add <4 x i64> %rdx.shuf.0.0, %rdx.shuf.0.1
   %rdx.shuf.1.0 = shufflevector <4 x i64> %bin.rdx, <4 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
;
   %bin.rdx9 = add <8 x i16> %rdx.shuf.2.0, %rdx.shuf.2.1
 
   %r = extractelement <8 x i16> %bin.rdx9, i32 0
   ret i16 %r
 }
 
 define fastcc i32 @pairwise_reduction8i32(<8 x i32> %rdx, i32 %f1) {
 ; SSE-LABEL: 'pairwise_reduction8i32'
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <8 x i32> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx9 = add <8 x i32> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx9, i32 0
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; AVX1-LABEL: 'pairwise_reduction8i32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %bin.rdx = add <8 x i32> %rdx.shuf.0.0, %rdx.shuf.0.1

 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx8 = add <8 x i32> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bin.rdx9 = add <8 x i32> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx9, i32 0
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
 ; SLM-LABEL: 'pairwise_reduction8i32'
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx = add <8 x i32> %rdx.shuf.0.0, %rdx.shuf.0.1
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.0 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.1.1 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.0 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.1.1 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx8 = add <8 x i32> %rdx.shuf.1.0, %rdx.shuf.1.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %rdx.shuf.2.0 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; SLM-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %rdx.shuf.2.1 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %rdx.shuf.2.1 = shufflevector <8 x i32> %bin.rdx8, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %bin.rdx9 = add <8 x i32> %rdx.shuf.2.0, %rdx.shuf.2.1
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %r = extractelement <8 x i32> %bin.rdx9, i32 0
 ; SLM-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 %r
 ;
   %rdx.shuf.0.0 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6,i32 undef, i32 undef, i32 undef, i32 undef>
   %rdx.shuf.0.1 = shufflevector <8 x i32> %rdx, <8 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7,i32 undef, i32 undef, i32 undef, i32 undef>
   %bin.rdx = add <8 x i32> %rdx.shuf.0.0, %rdx.shuf.0.1
   %rdx.shuf.1.0 = shufflevector <8 x i32> %bin.rdx, <8 x i32> undef,<8 x i32> <i32 0, i32 2, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>

llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll

 ;
 ; Verify the cost model for 1 src shuffles
 ;
 
 define void @test_vXf64(<2 x double> %src128, <4 x double> %src256, <8 x double> %src512, <16 x double> %src1024) {
 ; 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> undef, <2 x i32> <i32 1, i32 1>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, 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> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, 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> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; XOP-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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_vXf64'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 4 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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_vXf64'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x double> %src128, <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 2 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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
 ;
 ; 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> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, 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> undef, <2 x i32> <i32 1, i32 1>
   %V256 = shufflevector <4 x double> %src256, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
   %V512 = shufflevector <8 x double> %src512, <8 x double> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
   %V1024 = shufflevector <16 x double> %src1024, <16 x double> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ret void
 }
 
 define void @test_vXi64(<2 x i64> %src128, <4 x i64> %src256, <8 x i64> %src512) {
 ; 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> undef, <2 x i32> <i32 1, i32 1>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, 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> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, 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_vXi64'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <2 x i64> %src128, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <2 x i32> <i32 1, i32 1>
   %V256 = shufflevector <4 x i64> %src256, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
   %V512 = shufflevector <8 x i64> %src512, <8 x i64> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
   ret void
 }
 
 define void @test_vXf32(<2 x float> %src64, <4 x float> %src128, <8 x float> %src256, <16 x float> %src512) {
 ; SSE-LABEL: 'test_vXf32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, 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 <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, 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_vXf32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 4 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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_vXf32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 6 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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_vXf32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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 2 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, 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
 ;
 ; AVX512-LABEL: 'test_vXf32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, 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
 ;
   %V64  = shufflevector <2 x float> %src64, <2 x float> undef, <2 x i32> <i32 1, i32 1>
   %V128 = shufflevector <4 x float> %src128, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
   %V256 = shufflevector <8 x float> %src256, <8 x float> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
   %V512 = shufflevector <16 x float> %src512, <16 x float> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, i32 10, i32 8, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ret void
 }
 
 define void @test_vXi32(<2 x i32> %src64, <4 x i32> %src128, <8 x i32> %src256, <16 x i32> %src512, <32 x i32> %src1024) {
 ; SSE-LABEL: 'test_vXi32'
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
 ; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, i32 3, i32 2, i32 1, i32 0>
-; SSE-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 112 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, 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 2 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, i32 3, i32 2, i32 1, i32 0>
+; SSE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 10 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, 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
 ;
 ; XOP-LABEL: 'test_vXi32'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 48 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 6 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 12 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi32'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 48 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 14 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 0 for instruction: ret void
 ;
 ; AVX2-LABEL: 'test_vXi32'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 36 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 4 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 8 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 0 for instruction: ret void
 ;
 ; AVX512-LABEL: 'test_vXi32'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <2 x i32> %src64, <2 x i32> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, i32 3, i32 2, i32 1, i32 0>
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, 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 2 for instruction: %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, 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> undef, <2 x i32> <i32 1, i32 1>
   %V128 = shufflevector <4 x i32> %src128, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 0>
   %V256 = shufflevector <8 x i32> %src256, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 5, i32 3, i32 2, i32 1, i32 0>
   %V512 = shufflevector <16 x i32> %src512, <16 x i32> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 12, i32 13, i32 10, i32 9, i32 8, i32 8, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   %V1024 = shufflevector <32 x i32> %src1024, <32 x i32> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
   ret void
 }
 
 define void @test_vXi16(<2 x i16> %src32, <4 x i16> %src64, <8 x i16> %src128, <16 x i16> %src256, <32 x i16> %src512, <64 x i16> %src1024) {
 ; SSE2-LABEL: 'test_vXi16'
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 96 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 448 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 24 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 26 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, 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 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 36 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 168 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 8 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 8 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, 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 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 36 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 168 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 8 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 8 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; XOP-LABEL: 'test_vXi16'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 18 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 108 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 16 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX1-LABEL: 'test_vXi16'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 30 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 180 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 20 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
 ; 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: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, 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 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 10 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
 ; AVX2-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: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <16 x i16> %src256, <16 x i16> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, 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 14 for instruction: %V512 = shufflevector <32 x i16> %src512, <32 x i16> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, 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 84 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; AVX512F-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %V1024 = shufflevector <64 x i16> %src1024, <64 x i16> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; AVX512BW-LABEL: 'test_vXi16'
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <2 x i16> %src32, <2 x i16> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <4 x i16> %src64, <4 x i16> undef, <4 x i32> <i32 2, i32 2, i32 1, i32 1>
 ; AVX512BW-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <8 x i16> %src128, <8 x i16> undef, <8 x i32> <i32 7, i32 6, i32 6, i32 4, i32 3, 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> undef, <16 x i32> <i32 15, i32 14, i32 13, i32 13, i32 11, 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> undef, <32 x i32> <i32 31, i32 30, 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 14, i32 13, i32 12, i32 11, i32 11, i32 9, i32 8, i32 7, i32 11, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>

 }
 
 define void @test_vXi8(<2 x i8> %src16, <4 x i8> %src32, <8 x i8> %src64, <16 x i8> %src128, <32 x i8> %src256, <64 x i8> %src512) {
 ; SSE2-LABEL: 'test_vXi8'
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 26 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 156 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSE2-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 37 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 6 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 36 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 4 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 6 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 36 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 4 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; XOP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 18 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; XOP-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 8 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 30 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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 14 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 3, i32 2, i32 1, i32 0>
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V512 = shufflevector <64 x i8> %src512, <64 x i8> undef, <64 x i32> <i32 63, i32 62, i32 61, 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 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'
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16 = shufflevector <2 x i8> %src16, <2 x i8> undef, <2 x i32> <i32 1, i32 1>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V32 = shufflevector <4 x i8> %src32, <4 x i8> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 1>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V64 = shufflevector <8 x i8> %src64, <8 x i8> undef, <8 x i32> <i32 7, i32 7, i32 5, i32 5, i32 3, i32 3, i32 1, i32 1>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V128 = shufflevector <16 x i8> %src128, <16 x i8> undef, <16 x i32> <i32 15, i32 14, i32 13, 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>
 ; AVX512F-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V256 = shufflevector <32 x i8> %src256, <32 x i8> undef, <32 x i32> <i32 31, i32 30, i32 29, 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>

llvm/test/CodeGen/X86/haddsub-4.ll

; AVX2-NEXT:    retq
   %rhs = shufflevector <8 x double> %shuf0, <8 x double> %shuf1, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
   %fadd = fadd <8 x double> %lhs, %rhs
   ret <8 x double> %fadd
 }
 
 define <16 x float> @hadd_reverse_v16f32(<16 x float> %a0, <16 x float> %a1) nounwind {
 ; SSE-LABEL: hadd_reverse_v16f32:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    movaps %xmm5, %xmm8
-; SSE-NEXT:    movaps %xmm1, %xmm5
-; SSE-NEXT:    haddps %xmm2, %xmm3
-; SSE-NEXT:    shufps {{.*#+}} xmm3 = xmm3[1,0,3,2]
-; SSE-NEXT:    haddps %xmm6, %xmm7
-; SSE-NEXT:    shufps {{.*#+}} xmm7 = xmm7[1,0,3,2]
-; SSE-NEXT:    haddps %xmm0, %xmm5
-; SSE-NEXT:    shufps {{.*#+}} xmm5 = xmm5[1,0,3,2]
-; SSE-NEXT:    haddps %xmm4, %xmm8
-; SSE-NEXT:    shufps {{.*#+}} xmm8 = xmm8[1,0,3,2]
-; SSE-NEXT:    movaps %xmm3, %xmm0
-; SSE-NEXT:    movaps %xmm7, %xmm1
-; SSE-NEXT:    movaps %xmm5, %xmm2
+; SSE-NEXT:    movaps %xmm4, %xmm8
+; SSE-NEXT:    movaps %xmm0, %xmm4
+; SSE-NEXT:    haddps %xmm3, %xmm2
+; SSE-NEXT:    shufps {{.*#+}} xmm2 = xmm2[3,2,1,0]
+; SSE-NEXT:    haddps %xmm7, %xmm6
+; SSE-NEXT:    shufps {{.*#+}} xmm6 = xmm6[3,2,1,0]
+; SSE-NEXT:    haddps %xmm1, %xmm4
+; SSE-NEXT:    shufps {{.*#+}} xmm4 = xmm4[3,2,1,0]
+; SSE-NEXT:    haddps %xmm5, %xmm8
+; SSE-NEXT:    shufps {{.*#+}} xmm8 = xmm8[3,2,1,0]
+; SSE-NEXT:    movaps %xmm2, %xmm0
+; SSE-NEXT:    movaps %xmm6, %xmm1
+; SSE-NEXT:    movaps %xmm4, %xmm2
 ; SSE-NEXT:    movaps %xmm8, %xmm3
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: hadd_reverse_v16f32:
 ; AVX1:       # %bb.0:
 ; AVX1-NEXT:    vperm2f128 {{.*#+}} ymm4 = ymm0[2,3],ymm2[2,3]
 ; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
 ; AVX1-NEXT:    vhaddps %ymm0, %ymm4, %ymm2

llvm/test/CodeGen/X86/pr51615.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx  | FileCheck %s --check-prefixes=ALL,AVX
 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=ALL,AVX2
 
 ; https://bugs.llvm.org/show_bug.cgi?id=51615
 ; We can not replace a wide volatile load with a broadcast-from-memory,
 ; because that would narrow the load, which isn't legal for volatiles.
 
 @g0 = external dso_local global <2 x double>, align 16
 define void @volatile_load_2_elts() {
 ; AVX-LABEL: volatile_load_2_elts:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vmovaps g0(%rip), %xmm0
-; AVX-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
-; AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; AVX-NEXT:    vpermilpd {{.*#+}} ymm1 = ymm1[0,0,3,2]
-; AVX-NEXT:    vxorpd %xmm2, %xmm2, %xmm2
-; AVX-NEXT:    vblendpd {{.*#+}} ymm1 = ymm2[0,1],ymm1[2],ymm2[3]
-; AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; AVX-NEXT:    vblendpd {{.*#+}} ymm0 = ymm2[0],ymm0[1],ymm2[2],ymm0[3]
-; AVX-NEXT:    vmovapd %ymm0, (%rax)
-; AVX-NEXT:    vmovapd %ymm1, (%rax)
+; AVX-NEXT:    vmovddup {{.*#+}} xmm1 = xmm0[0,0]
+; AVX-NEXT:    vinsertf128 $1, %xmm1, %ymm1, %ymm1
+; AVX-NEXT:    vxorps %xmm2, %xmm2, %xmm2
+; AVX-NEXT:    vblendps {{.*#+}} ymm1 = ymm2[0,1],ymm1[2,3],ymm2[4,5],ymm1[6,7]
+; AVX-NEXT:    vperm2f128 {{.*#+}} ymm0 = zero,zero,ymm0[0,1]
+; AVX-NEXT:    vblendps {{.*#+}} ymm0 = ymm2[0,1,2,3],ymm0[4,5],ymm2[6,7]
+; AVX-NEXT:    vmovaps %ymm0, (%rax)
+; AVX-NEXT:    vmovaps %ymm1, (%rax)
 ; AVX-NEXT:    vzeroupper
 ; AVX-NEXT:    retq
 ;
 ; AVX2-LABEL: volatile_load_2_elts:
 ; AVX2:       # %bb.0:
 ; AVX2-NEXT:    vmovaps g0(%rip), %xmm0
 ; AVX2-NEXT:    vbroadcastsd %xmm0, %ymm0
 ; AVX2-NEXT:    vxorps %xmm1, %xmm1, %xmm1

llvm/test/CodeGen/X86/vector-interleaved-store-i16-stride-3.ll

 ; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm0[0,1,2,3,6,5,4,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[0,3,2,1]
 ; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,2,2,1,4,5,6,7]
 ; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,5,5,6,4]
 ; SSE-NEXT:    pand %xmm3, %xmm4
 ; SSE-NEXT:    pandn %xmm2, %xmm3
 ; SSE-NEXT:    por %xmm4, %xmm3
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,1,1]
-; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [65535,0,0,65535,65535,65535,65535,65535]
-; SSE-NEXT:    pand %xmm2, %xmm1
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [0,65535,65535,0,65535,65535,65535,65535]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[3,1,2,3]
 ; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,3,1,3,4,5,6,7]
-; SSE-NEXT:    pandn %xmm0, %xmm2
-; SSE-NEXT:    por %xmm1, %xmm2
+; SSE-NEXT:    pand %xmm2, %xmm0
+; SSE-NEXT:    pandn %xmm1, %xmm2
+; SSE-NEXT:    por %xmm0, %xmm2
 ; SSE-NEXT:    movq %xmm2, 16(%rcx)
 ; SSE-NEXT:    movdqa %xmm3, (%rcx)
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: vf4:
 ; AVX1:       # %bb.0:
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
 ; AVX1-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero

llvm/test/CodeGen/X86/vector-interleaved-store-i32-stride-6.ll

 ; AVX512-NEXT:    movq {{[0-9]+}}(%rsp), %rax
 ; AVX512-NEXT:    vmovdqa (%rdi), %xmm0
 ; AVX512-NEXT:    vmovdqa (%rdx), %xmm1
 ; AVX512-NEXT:    vmovdqa (%r8), %xmm2
 ; AVX512-NEXT:    vinserti128 $1, (%rcx), %ymm1, %ymm1
 ; AVX512-NEXT:    vinserti128 $1, (%rsi), %ymm0, %ymm0
 ; AVX512-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
 ; AVX512-NEXT:    vinserti32x4 $1, (%r9), %zmm2, %zmm1
-; AVX512-NEXT:    vmovdqa {{.*#+}} ymm2 = [2,6,19,23,27,31,3,7]
-; AVX512-NEXT:    vpermi2d %zmm0, %zmm1, %zmm2
+; AVX512-NEXT:    vmovdqa {{.*#+}} ymm2 = [18,22,3,7,11,15,19,23]
+; AVX512-NEXT:    vpermi2d %zmm1, %zmm0, %zmm2
 ; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [0,4,8,12,16,20,1,5,9,13,17,21,2,6,10,14]
 ; AVX512-NEXT:    vpermi2d %zmm1, %zmm0, %zmm3
 ; AVX512-NEXT:    vmovdqu64 %zmm3, (%rax)
 ; AVX512-NEXT:    vmovdqa %ymm2, 64(%rax)
 ; AVX512-NEXT:    vzeroupper
 ; AVX512-NEXT:    retq
   %in.vec0 = load <4 x i32>, <4 x i32>* %in.vecptr0, align 32
   %in.vec1 = load <4 x i32>, <4 x i32>* %in.vecptr1, align 32

llvm/test/CodeGen/X86/vector-interleaved-store-i64-stride-3.ll

 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
 ;
 ; AVX512-LABEL: store_i64_stride3_vf4:
 ; AVX512:       # %bb.0:
 ; AVX512-NEXT:    vmovdqa (%rdi), %ymm0
 ; AVX512-NEXT:    vmovdqa (%rdx), %ymm1
 ; AVX512-NEXT:    vinserti64x4 $1, (%rsi), %zmm0, %zmm0
-; AVX512-NEXT:    vmovdqa {{.*#+}} ymm2 = [2,11,15,3]
-; AVX512-NEXT:    vpermi2q %zmm0, %zmm1, %zmm2
+; AVX512-NEXT:    vmovdqa {{.*#+}} ymm2 = [10,3,7,11]
+; AVX512-NEXT:    vpermi2q %zmm1, %zmm0, %zmm2
 ; AVX512-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [0,4,8,1,5,9,2,6]
 ; AVX512-NEXT:    vpermi2q %zmm1, %zmm0, %zmm3
 ; AVX512-NEXT:    vmovdqu64 %zmm3, (%rcx)
 ; AVX512-NEXT:    vmovdqa %ymm2, 64(%rcx)
 ; AVX512-NEXT:    vzeroupper
 ; AVX512-NEXT:    retq
   %in.vec0 = load <4 x i64>, <4 x i64>* %in.vecptr0, align 32
   %in.vec1 = load <4 x i64>, <4 x i64>* %in.vecptr1, align 32

llvm/test/CodeGen/X86/vector-interleaved-store-i8-stride-6.ll

 ; SSE-NEXT:    pandn %xmm5, %xmm3
 ; SSE-NEXT:    por %xmm7, %xmm3
 ; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[3,1,2,3]
 ; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[3,1,2,3,4,5,6,7]
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[3,1,2,3]
 ; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,1,3,1,4,5,6,7]
 ; SSE-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
 ; SSE-NEXT:    packuswb %xmm1, %xmm1
-; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [65535,0,0,65535,65535,65535,65535,65535]
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [0,65535,65535,0,65535,65535,65535,65535]
+; SSE-NEXT:    pand %xmm2, %xmm1
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[3,1,1,3]
 ; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[2,0,2,3,4,5,6,7]
 ; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,5,7]
 ; SSE-NEXT:    packuswb %xmm0, %xmm0
-; SSE-NEXT:    pand %xmm2, %xmm0
-; SSE-NEXT:    pandn %xmm1, %xmm2
-; SSE-NEXT:    por %xmm0, %xmm2
+; SSE-NEXT:    pandn %xmm0, %xmm2
+; SSE-NEXT:    por %xmm1, %xmm2
 ; SSE-NEXT:    movq %xmm2, 16(%rax)
 ; SSE-NEXT:    movdqa %xmm3, (%rax)
 ; SSE-NEXT:    retq
 ;
 ; AVX1-LABEL: store_i8_stride6_vf4:
 ; AVX1:       # %bb.0:
 ; AVX1-NEXT:    movq {{[0-9]+}}(%rsp), %rax
 ; AVX1-NEXT:    vmovdqa (%rdi), %xmm0

llvm/test/Transforms/SLPVectorizer/X86/PR39774.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-6 | FileCheck %s --check-prefix=CHECK
+; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-4 | FileCheck %s --check-prefix=CHECK
 ; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-8 -slp-min-tree-size=6 | FileCheck %s --check-prefix=FORCE_REDUCTION
 
 define void @Test(i32) {
 ; CHECK-LABEL: @Test(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    br label [[LOOP:%.*]]
 ; CHECK:       loop:
 ; CHECK-NEXT:    [[TMP1:%.*]] = phi <2 x i32> [ [[TMP10:%.*]], [[LOOP]] ], [ zeroinitializer, [[ENTRY:%.*]] ]

llvm/test/Transforms/SLPVectorizer/X86/load-merge-inseltpoison.ll

entry:
   ret i32 %or11
 }
 
 define <4 x float> @PR16739_byref(<4 x float>* nocapture readonly dereferenceable(16) %x) {
 ; CHECK-LABEL: @PR16739_byref(
 ; CHECK-NEXT:    [[GEP0:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X:%.*]], i64 0, i64 0
 ; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X]], i64 0, i64 1
 ; CHECK-NEXT:    [[GEP2:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X]], i64 0, i64 2
-; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[GEP0]] to <2 x float>*
+; CHECK-NEXT:    [[X0:%.*]] = load float, float* [[GEP0]], align 4
+; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[GEP1]] to <2 x float>*
 ; CHECK-NEXT:    [[TMP2:%.*]] = load <2 x float>, <2 x float>* [[TMP1]], align 4
-; CHECK-NEXT:    [[X2:%.*]] = load float, float* [[GEP2]], align 4
+; CHECK-NEXT:    [[I0:%.*]] = insertelement <4 x float> poison, float [[X0]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
-; CHECK-NEXT:    [[I2:%.*]] = insertelement <4 x float> [[TMP3]], float [[X2]], i32 2
-; CHECK-NEXT:    [[I3:%.*]] = insertelement <4 x float> [[I2]], float [[X2]], i32 3
+; CHECK-NEXT:    [[I21:%.*]] = shufflevector <4 x float> [[I0]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 4, i32 5, i32 3>
+; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x float> [[TMP2]], i32 1
+; CHECK-NEXT:    [[I3:%.*]] = insertelement <4 x float> [[I21]], float [[TMP4]], i32 3
 ; CHECK-NEXT:    ret <4 x float> [[I3]]
 ;
   %gep0 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 0
   %gep1 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 1
   %gep2 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 2
   %x0 = load float, float* %gep0
   %x1 = load float, float* %gep1
   %x2 = load float, float* %gep2

llvm/test/Transforms/SLPVectorizer/X86/load-merge.ll

entry:
   ret i32 %or11
 }
 
 define <4 x float> @PR16739_byref(<4 x float>* nocapture readonly dereferenceable(16) %x) {
 ; CHECK-LABEL: @PR16739_byref(
 ; CHECK-NEXT:    [[GEP0:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X:%.*]], i64 0, i64 0
 ; CHECK-NEXT:    [[GEP1:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X]], i64 0, i64 1
 ; CHECK-NEXT:    [[GEP2:%.*]] = getelementptr inbounds <4 x float>, <4 x float>* [[X]], i64 0, i64 2
-; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[GEP0]] to <2 x float>*
+; CHECK-NEXT:    [[X0:%.*]] = load float, float* [[GEP0]], align 4
+; CHECK-NEXT:    [[TMP1:%.*]] = bitcast float* [[GEP1]] to <2 x float>*
 ; CHECK-NEXT:    [[TMP2:%.*]] = load <2 x float>, <2 x float>* [[TMP1]], align 4
-; CHECK-NEXT:    [[X2:%.*]] = load float, float* [[GEP2]], align 4
+; CHECK-NEXT:    [[I0:%.*]] = insertelement <4 x float> undef, float [[X0]], i32 0
 ; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
-; CHECK-NEXT:    [[I2:%.*]] = insertelement <4 x float> [[TMP3]], float [[X2]], i32 2
-; CHECK-NEXT:    [[I3:%.*]] = insertelement <4 x float> [[I2]], float [[X2]], i32 3
+; CHECK-NEXT:    [[I21:%.*]] = shufflevector <4 x float> [[I0]], <4 x float> [[TMP3]], <4 x i32> <i32 0, i32 4, i32 5, i32 3>
+; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x float> [[TMP2]], i32 1
+; CHECK-NEXT:    [[I3:%.*]] = insertelement <4 x float> [[I21]], float [[TMP4]], i32 3
 ; CHECK-NEXT:    ret <4 x float> [[I3]]
 ;
   %gep0 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 0
   %gep1 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 1
   %gep2 = getelementptr inbounds <4 x float>, <4 x float>* %x, i64 0, i64 2
   %x0 = load float, float* %gep0
   %x1 = load float, float* %gep1
   %x2 = load float, float* %gep2

llvm/test/Transforms/SLPVectorizer/X86/vec_list_bias-inseltpoison.ll

 ; CHECK-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T2]], i64 4
 ; CHECK-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4
 ; CHECK-NEXT:    [[T24:%.*]] = add nsw i32 [[T23]], [[T21]]
 ; CHECK-NEXT:    [[T25:%.*]] = sub nsw i32 [[T21]], [[T23]]
 ; CHECK-NEXT:    [[T27:%.*]] = sub nsw i32 [[T3]], [[T24]]
 ; CHECK-NEXT:    [[T29:%.*]] = sub nsw i32 [[T9]], [[T15]]
 ; CHECK-NEXT:    [[T30:%.*]] = add nsw i32 [[T27]], [[T29]]
 ; CHECK-NEXT:    [[T31:%.*]] = mul nsw i32 [[T30]], 4433
-; CHECK-NEXT:    [[T32:%.*]] = mul nsw i32 [[T27]], 6270
 ; CHECK-NEXT:    [[T34:%.*]] = mul nsw i32 [[T29]], -15137
 ; CHECK-NEXT:    [[T37:%.*]] = add nsw i32 [[T25]], [[T11]]
 ; CHECK-NEXT:    [[T38:%.*]] = add nsw i32 [[T17]], [[T5]]
 ; CHECK-NEXT:    [[T39:%.*]] = add nsw i32 [[T37]], [[T38]]
 ; CHECK-NEXT:    [[T40:%.*]] = mul nsw i32 [[T39]], 9633
 ; CHECK-NEXT:    [[T41:%.*]] = mul nsw i32 [[T25]], 2446
 ; CHECK-NEXT:    [[T42:%.*]] = mul nsw i32 [[T17]], 16819
 ; CHECK-NEXT:    [[T47:%.*]] = mul nsw i32 [[T37]], -16069
 ; CHECK-NEXT:    [[T48:%.*]] = mul nsw i32 [[T38]], -3196
-; CHECK-NEXT:    [[T49:%.*]] = add nsw i32 [[T40]], [[T47]]
-; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i32> poison, i32 [[T15]], i32 0
-; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i32> [[TMP1]], i32 [[T40]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i32> poison, i32 [[T9]], i32 0
-; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x i32> [[TMP3]], i32 [[T48]], i32 1
-; CHECK-NEXT:    [[TMP5:%.*]] = add nsw <2 x i32> [[TMP2]], [[TMP4]]
-; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x i32> [[TMP5]], i32 0
-; CHECK-NEXT:    [[T65:%.*]] = insertelement <8 x i32> poison, i32 [[TMP6]], i32 0
-; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x i32> [[TMP5]], i32 1
-; CHECK-NEXT:    [[T66:%.*]] = insertelement <8 x i32> [[T65]], i32 [[TMP7]], i32 1
-; CHECK-NEXT:    [[T67:%.*]] = insertelement <8 x i32> [[T66]], i32 [[T32]], i32 2
-; CHECK-NEXT:    [[T68:%.*]] = insertelement <8 x i32> [[T67]], i32 [[T49]], i32 3
-; CHECK-NEXT:    [[TMP8:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; CHECK-NEXT:    [[T701:%.*]] = shufflevector <8 x i32> [[T68]], <8 x i32> [[TMP8]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
-; CHECK-NEXT:    [[T71:%.*]] = insertelement <8 x i32> [[T701]], i32 [[T34]], i32 6
-; CHECK-NEXT:    [[T72:%.*]] = insertelement <8 x i32> [[T71]], i32 [[T49]], i32 7
+; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> poison, i32 [[T40]], i32 0
+; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[T27]], i32 1
+; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[T40]], i32 2
+; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> [[TMP3]], i32 [[T15]], i32 3
+; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> <i32 poison, i32 6270, i32 poison, i32 poison>, i32 [[T48]], i32 0
+; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[T47]], i32 2
+; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[T9]], i32 3
+; CHECK-NEXT:    [[TMP8:%.*]] = add nsw <4 x i32> [[TMP4]], [[TMP7]]
+; CHECK-NEXT:    [[TMP9:%.*]] = mul nsw <4 x i32> [[TMP4]], [[TMP7]]
+; CHECK-NEXT:    [[TMP10:%.*]] = shufflevector <4 x i32> [[TMP8]], <4 x i32> [[TMP9]], <4 x i32> <i32 0, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP11:%.*]] = extractelement <4 x i32> [[TMP10]], i32 3
+; CHECK-NEXT:    [[T65:%.*]] = insertelement <8 x i32> poison, i32 [[TMP11]], i32 0
+; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <4 x i32> [[TMP10]], <4 x i32> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT:    [[T691:%.*]] = shufflevector <8 x i32> [[T65]], <8 x i32> [[TMP12]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <4 x i32> [[TMP10]], i32 0
+; CHECK-NEXT:    [[T70:%.*]] = insertelement <8 x i32> [[T691]], i32 [[TMP13]], i32 5
+; CHECK-NEXT:    [[T71:%.*]] = insertelement <8 x i32> [[T70]], i32 [[T34]], i32 6
+; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <4 x i32> [[TMP10]], i32 2
+; CHECK-NEXT:    [[T72:%.*]] = insertelement <8 x i32> [[T71]], i32 [[TMP14]], i32 7
 ; CHECK-NEXT:    [[T76:%.*]] = shl <8 x i32> [[T72]], <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
 ; CHECK-NEXT:    [[T79:%.*]] = bitcast i32* [[T2]] to <8 x i32>*
 ; CHECK-NEXT:    store <8 x i32> [[T76]], <8 x i32>* [[T79]], align 4
 ; CHECK-NEXT:    ret void
 ;
   %t3 = load i32, i32* %t2, align 4
   %t4 = getelementptr inbounds i32, i32* %t2, i64 7
   %t5 = load i32, i32* %t4, align 4

llvm/test/Transforms/SLPVectorizer/X86/vec_list_bias.ll

 ; CHECK-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T2]], i64 4
 ; CHECK-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4
 ; CHECK-NEXT:    [[T24:%.*]] = add nsw i32 [[T23]], [[T21]]
 ; CHECK-NEXT:    [[T25:%.*]] = sub nsw i32 [[T21]], [[T23]]
 ; CHECK-NEXT:    [[T27:%.*]] = sub nsw i32 [[T3]], [[T24]]
 ; CHECK-NEXT:    [[T29:%.*]] = sub nsw i32 [[T9]], [[T15]]
 ; CHECK-NEXT:    [[T30:%.*]] = add nsw i32 [[T27]], [[T29]]
 ; CHECK-NEXT:    [[T31:%.*]] = mul nsw i32 [[T30]], 4433
-; CHECK-NEXT:    [[T32:%.*]] = mul nsw i32 [[T27]], 6270
 ; CHECK-NEXT:    [[T34:%.*]] = mul nsw i32 [[T29]], -15137
 ; CHECK-NEXT:    [[T37:%.*]] = add nsw i32 [[T25]], [[T11]]
 ; CHECK-NEXT:    [[T38:%.*]] = add nsw i32 [[T17]], [[T5]]
 ; CHECK-NEXT:    [[T39:%.*]] = add nsw i32 [[T37]], [[T38]]
 ; CHECK-NEXT:    [[T40:%.*]] = mul nsw i32 [[T39]], 9633
 ; CHECK-NEXT:    [[T41:%.*]] = mul nsw i32 [[T25]], 2446
 ; CHECK-NEXT:    [[T42:%.*]] = mul nsw i32 [[T17]], 16819
 ; CHECK-NEXT:    [[T47:%.*]] = mul nsw i32 [[T37]], -16069
 ; CHECK-NEXT:    [[T48:%.*]] = mul nsw i32 [[T38]], -3196
-; CHECK-NEXT:    [[T49:%.*]] = add nsw i32 [[T40]], [[T47]]
-; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <2 x i32> poison, i32 [[T15]], i32 0
-; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <2 x i32> [[TMP1]], i32 [[T40]], i32 1
-; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <2 x i32> poison, i32 [[T9]], i32 0
-; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x i32> [[TMP3]], i32 [[T48]], i32 1
-; CHECK-NEXT:    [[TMP5:%.*]] = add nsw <2 x i32> [[TMP2]], [[TMP4]]
-; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x i32> [[TMP5]], i32 0
-; CHECK-NEXT:    [[T65:%.*]] = insertelement <8 x i32> undef, i32 [[TMP6]], i32 0
-; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x i32> [[TMP5]], i32 1
-; CHECK-NEXT:    [[T66:%.*]] = insertelement <8 x i32> [[T65]], i32 [[TMP7]], i32 1
-; CHECK-NEXT:    [[T67:%.*]] = insertelement <8 x i32> [[T66]], i32 [[T32]], i32 2
-; CHECK-NEXT:    [[T68:%.*]] = insertelement <8 x i32> [[T67]], i32 [[T49]], i32 3
-; CHECK-NEXT:    [[TMP8:%.*]] = shufflevector <2 x i32> [[TMP5]], <2 x i32> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
-; CHECK-NEXT:    [[T701:%.*]] = shufflevector <8 x i32> [[T68]], <8 x i32> [[TMP8]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 6, i32 7>
-; CHECK-NEXT:    [[T71:%.*]] = insertelement <8 x i32> [[T701]], i32 [[T34]], i32 6
-; CHECK-NEXT:    [[T72:%.*]] = insertelement <8 x i32> [[T71]], i32 [[T49]], i32 7
+; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> poison, i32 [[T40]], i32 0
+; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[T27]], i32 1
+; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[T40]], i32 2
+; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x i32> [[TMP3]], i32 [[T15]], i32 3
+; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x i32> <i32 poison, i32 6270, i32 poison, i32 poison>, i32 [[T48]], i32 0
+; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32> [[TMP5]], i32 [[T47]], i32 2
+; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x i32> [[TMP6]], i32 [[T9]], i32 3
+; CHECK-NEXT:    [[TMP8:%.*]] = add nsw <4 x i32> [[TMP4]], [[TMP7]]
+; CHECK-NEXT:    [[TMP9:%.*]] = mul nsw <4 x i32> [[TMP4]], [[TMP7]]
+; CHECK-NEXT:    [[TMP10:%.*]] = shufflevector <4 x i32> [[TMP8]], <4 x i32> [[TMP9]], <4 x i32> <i32 0, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP11:%.*]] = extractelement <4 x i32> [[TMP10]], i32 3
+; CHECK-NEXT:    [[T65:%.*]] = insertelement <8 x i32> undef, i32 [[TMP11]], i32 0
+; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <4 x i32> [[TMP10]], <4 x i32> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT:    [[T691:%.*]] = shufflevector <8 x i32> [[T65]], <8 x i32> [[TMP12]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <4 x i32> [[TMP10]], i32 0
+; CHECK-NEXT:    [[T70:%.*]] = insertelement <8 x i32> [[T691]], i32 [[TMP13]], i32 5
+; CHECK-NEXT:    [[T71:%.*]] = insertelement <8 x i32> [[T70]], i32 [[T34]], i32 6
+; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <4 x i32> [[TMP10]], i32 2
+; CHECK-NEXT:    [[T72:%.*]] = insertelement <8 x i32> [[T71]], i32 [[TMP14]], i32 7
 ; CHECK-NEXT:    [[T76:%.*]] = shl <8 x i32> [[T72]], <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
 ; CHECK-NEXT:    [[T79:%.*]] = bitcast i32* [[T2]] to <8 x i32>*
 ; CHECK-NEXT:    store <8 x i32> [[T76]], <8 x i32>* [[T79]], align 4
 ; CHECK-NEXT:    ret void
 ;
   %t3 = load i32, i32* %t2, align 4
   %t4 = getelementptr inbounds i32, i32* %t2, i64 7
   %t5 = load i32, i32* %t4, align 4

llvm/test/Transforms/SLPVectorizer/X86/vectorize-widest-phis.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: opt -slp-vectorizer -S -mcpu=cascadelake -mtriple=x86_64-unknown-linux-gnu < %s | FileCheck %s
 
 define void @foo() {
 ; CHECK-LABEL: @foo(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[CONV:%.*]] = uitofp i16 undef to float
 ; CHECK-NEXT:    [[SUB:%.*]] = fsub float 6.553500e+04, undef
 ; CHECK-NEXT:    br label [[BB1:%.*]]
 ; CHECK:       bb1:
 ; CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x float> poison, float [[SUB]], i32 0
 ; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x float> [[TMP0]], float [[CONV]], i32 1
 ; CHECK-NEXT:    br label [[BB2:%.*]]
 ; CHECK:       bb2:
-; CHECK-NEXT:    [[TMP2:%.*]] = phi <4 x float> [ [[TMP1]], [[BB1]] ], [ [[TMP18:%.*]], [[BB3:%.*]] ]
+; CHECK-NEXT:    [[TMP2:%.*]] = phi <4 x float> [ [[TMP1]], [[BB1]] ], [ [[TMP10:%.*]], [[BB3:%.*]] ]
 ; CHECK-NEXT:    [[TMP3:%.*]] = load double, double* undef, align 8
 ; CHECK-NEXT:    br i1 undef, label [[BB3]], label [[BB4:%.*]]
 ; CHECK:       bb4:
 ; CHECK-NEXT:    [[CONV2:%.*]] = uitofp i16 undef to double
+; CHECK-NEXT:    [[ADD1:%.*]] = fadd double [[TMP3]], [[CONV2]]
 ; CHECK-NEXT:    [[TMP4:%.*]] = fpext <4 x float> [[TMP2]] to <4 x double>
-; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <2 x double> <double undef, double poison>, double [[TMP3]], i32 1
-; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x double> <double undef, double poison>, double [[CONV2]], i32 1
-; CHECK-NEXT:    [[TMP7:%.*]] = fsub <2 x double> [[TMP5]], [[TMP6]]
-; CHECK-NEXT:    [[TMP8:%.*]] = fadd <2 x double> [[TMP5]], [[TMP6]]
-; CHECK-NEXT:    [[TMP9:%.*]] = shufflevector <2 x double> [[TMP7]], <2 x double> [[TMP8]], <2 x i32> <i32 0, i32 3>
-; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <2 x double> [[TMP9]], i32 0
-; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <4 x double> poison, double [[TMP10]], i32 0
-; CHECK-NEXT:    [[TMP12:%.*]] = extractelement <2 x double> [[TMP9]], i32 1
-; CHECK-NEXT:    [[TMP13:%.*]] = insertelement <4 x double> [[TMP11]], double [[TMP12]], i32 1
-; CHECK-NEXT:    [[TMP14:%.*]] = fcmp ogt <4 x double> [[TMP13]], [[TMP4]]
-; CHECK-NEXT:    [[TMP15:%.*]] = shufflevector <2 x double> [[TMP9]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
-; CHECK-NEXT:    [[TMP16:%.*]] = fptrunc <4 x double> [[TMP15]] to <4 x float>
-; CHECK-NEXT:    [[TMP17:%.*]] = select <4 x i1> [[TMP14]], <4 x float> [[TMP2]], <4 x float> [[TMP16]]
+; CHECK-NEXT:    [[SUB1:%.*]] = fsub double undef, undef
+; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <4 x double> poison, double [[SUB1]], i32 0
+; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <4 x double> [[TMP5]], double [[ADD1]], i32 1
+; CHECK-NEXT:    [[TMP7:%.*]] = fcmp ogt <4 x double> [[TMP6]], [[TMP4]]
+; CHECK-NEXT:    [[TMP8:%.*]] = fptrunc <4 x double> [[TMP6]] to <4 x float>
+; CHECK-NEXT:    [[TMP9:%.*]] = select <4 x i1> [[TMP7]], <4 x float> [[TMP2]], <4 x float> [[TMP8]]
 ; CHECK-NEXT:    br label [[BB3]]
 ; CHECK:       bb3:
-; CHECK-NEXT:    [[TMP18]] = phi <4 x float> [ [[TMP17]], [[BB4]] ], [ [[TMP2]], [[BB2]] ]
+; CHECK-NEXT:    [[TMP10]] = phi <4 x float> [ [[TMP9]], [[BB4]] ], [ [[TMP2]], [[BB2]] ]
 ; CHECK-NEXT:    br label [[BB2]]
 ;
 entry:
   %conv = uitofp i16 undef to float
   %sub = fsub float 6.553500e+04, undef
   br label %bb1
 
 bb1: