Skip to content

[AArch64] Check for exact size when finding 1's for CMLE. #76452

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

Merged
davemgreen merged 1 commit into from
Jan 2, 2024
Merged

[AArch64] Check for exact size when finding 1's for CMLE. #76452

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
9 changes: 4 additions & 5 deletions llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13715,11 +13715,10 @@ static SDValue EmitVectorComparison(SDValue LHS, SDValue RHS,
bool IsCnst = BVN && BVN->isConstantSplat(SplatValue, SplatUndef,
SplatBitSize, HasAnyUndefs);

bool IsSplatUniform =
SrcVT.getVectorElementType().getSizeInBits() >= SplatBitSize;
bool IsZero = IsCnst && SplatValue == 0 && IsSplatUniform;
bool IsOne = IsCnst && SplatValue == 1 && IsSplatUniform;
bool IsMinusOne = IsCnst && SplatValue.isAllOnes() && IsSplatUniform;
bool IsZero = IsCnst && SplatValue == 0;
bool IsOne =
IsCnst && SrcVT.getScalarSizeInBits() == SplatBitSize && SplatValue == 1;
bool IsMinusOne = IsCnst && SplatValue.isAllOnes();

if (SrcVT.getVectorElementType().isFloatingPoint()) {
switch (CC) {
Expand Down
3 changes: 2 additions & 1 deletion llvm/test/CodeGen/AArch64/neon-compare-instructions.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1792,7 +1792,8 @@ define <8 x i1> @not_cmle8xi8(<8 x i8> %0) {
define <4 x i1> @not_cmle16xi8(<4 x i32> %0) {
; CHECK-SD-LABEL: not_cmle16xi8:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: cmle v0.4s, v0.4s, #0
; CHECK-SD-NEXT: movi v1.8h, #1
; CHECK-SD-NEXT: cmgt v0.4s, v1.4s, v0.4s
; CHECK-SD-NEXT: xtn v0.4h, v0.4s
; CHECK-SD-NEXT: ret
;
Expand Down