Skip to content

[InstCombine] Let shrinkSplatShuffle act on vectors of different lengths #148593

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
+23 −6
Open

[InstCombine] Let shrinkSplatShuffle act on vectors of different lengths #148593

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6fa9358
Expand move trunc through shuffle splat
Jul 14, 2025
fb5b9c0
Answer comments
Jul 15, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 6 additions & 2 deletions llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -708,10 +708,14 @@ static Instruction *shrinkSplatShuffle(TruncInst &Trunc,
auto *Shuf = dyn_cast<ShuffleVectorInst>(Trunc.getOperand(0));
if (Shuf && Shuf->hasOneUse() && match(Shuf->getOperand(1), m_Undef()) &&
all_equal(Shuf->getShuffleMask()) &&
Shuf->getType() == Shuf->getOperand(0)->getType()) {
ElementCount::isKnownGE(Shuf->getType()->getElementCount(),
cast<VectorType>(Shuf->getOperand(0)->getType())
->getElementCount())) {
// trunc (shuf X, Undef, SplatMask) --> shuf (trunc X), Poison, SplatMask
// trunc (shuf X, Poison, SplatMask) --> shuf (trunc X), Poison, SplatMask
Value *NarrowOp = Builder.CreateTrunc(Shuf->getOperand(0), Trunc.getType());
Type *NewTruncTy = Shuf->getOperand(0)->getType()->getWithNewType(
Trunc.getType()->getScalarType());
Value *NarrowOp = Builder.CreateTrunc(Shuf->getOperand(0), NewTruncTy);
return new ShuffleVectorInst(NarrowOp, Shuf->getShuffleMask());
}

Expand Down
4 changes: 2 additions & 2 deletions llvm/test/Transforms/InstCombine/trunc-inseltpoison.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -959,8 +959,8 @@ define <3 x i31> @wide_splat3(<3 x i33> %x) {

define <8 x i8> @wide_lengthening_splat(<4 x i16> %v) {
; CHECK-LABEL: @wide_lengthening_splat(
; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <4 x i16> [[V:%.*]], <4 x i16> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TR:%.*]] = trunc <8 x i16> [[SHUF]] to <8 x i8>
; CHECK-NEXT: [[TMP1:%.*]] = trunc <4 x i16> [[V:%.*]] to <4 x i8>
; CHECK-NEXT: [[TR:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: ret <8 x i8> [[TR]]
;
%shuf = shufflevector <4 x i16> %v, <4 x i16> %v, <8 x i32> zeroinitializer
Expand Down
17 changes: 15 additions & 2 deletions llvm/test/Transforms/InstCombine/trunc.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -960,15 +960,28 @@ define <3 x i31> @wide_splat3(<3 x i33> %x) {

define <8 x i8> @wide_lengthening_splat(<4 x i16> %v) {
; CHECK-LABEL: @wide_lengthening_splat(
; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <4 x i16> [[V:%.*]], <4 x i16> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TR:%.*]] = trunc <8 x i16> [[SHUF]] to <8 x i8>
; CHECK-NEXT: [[TMP1:%.*]] = trunc <4 x i16> [[V:%.*]] to <4 x i8>
; CHECK-NEXT: [[TR:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> poison, <8 x i32> zeroinitializer
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also test shortening splat? I think in that case the profitability is less clear.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah I think we want to restrict it to only when Shuf->getOperand(0) is shorter than Shuf?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't understand why it wouldn't be profitable in that case, could you please elaborate?

Also added test

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We're now performing trunc on a wider type than before, and that can be slower on some targets. E.g. a <8 x i16> trunc may take twice as many uops as a <4 x i16> trunc. There is definitely at least hardware on RISC-V where this is the case.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks! changed

; CHECK-NEXT: ret <8 x i8> [[TR]]
;
%shuf = shufflevector <4 x i16> %v, <4 x i16> %v, <8 x i32> zeroinitializer
%tr = trunc <8 x i16> %shuf to <8 x i8>
ret <8 x i8> %tr
}

; This is a negative test, we expect the trunc to remain after the shuffle as it
; might not be beneficial to preform trunc on a wider type
define <4 x i8> @wide_shortening_splat(<8 x i16> %v) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nit, add a comment above explaining that this is a negative test?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done

; CHECK-LABEL: @wide_shortening_splat(
; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <8 x i16> [[V:%.*]], <8 x i16> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[TR:%.*]] = trunc <4 x i16> [[SHUF]] to <4 x i8>
; CHECK-NEXT: ret <4 x i8> [[TR]]
;
%shuf = shufflevector <8 x i16> %v, <8 x i16> %v, <4 x i32> zeroinitializer
%tr = trunc <4 x i16> %shuf to <4 x i8>
ret <4 x i8> %tr
}

define <2 x i8> @narrow_add_vec_constant(<2 x i32> %x) {
; CHECK-LABEL: @narrow_add_vec_constant(
; CHECK-NEXT: [[TMP1:%.*]] = trunc <2 x i32> [[X:%.*]] to <2 x i8>
Expand Down