Skip to content

[ConstantFolding] Fix incorrect nvvm_round folding #151563

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
LewisCrawford merged 2 commits into from
Aug 1, 2025
Merged

[ConstantFolding] Fix incorrect nvvm_round folding #151563

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
40d2346
[ConstantFolding] Fix incorrect nvvm_round folding
LewisCrawford Jul 31, 2025
0b63699
Add comment about cvt.rni
LewisCrawford Aug 1, 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
7 changes: 4 additions & 3 deletions llvm/lib/Analysis/ConstantFolding.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2679,11 +2679,12 @@ static Constant *ConstantFoldScalarCall1(StringRef Name,
case Intrinsic::nvvm_round_ftz_f:
case Intrinsic::nvvm_round_f:
case Intrinsic::nvvm_round_d: {
// Use APFloat implementation instead of native libm call, as some
// implementations (e.g. on PPC) do not preserve the sign of negative 0.
// nvvm_round is lowered to PTX cvt.rni, which will round to nearest
// integer, choosing even integer if source is equidistant between two
// integers, so the semantics are closer to "rint" rather than "round".
bool IsFTZ = nvvm::UnaryMathIntrinsicShouldFTZ(IntrinsicID);
auto V = IsFTZ ? FTZPreserveSign(APF) : APF;
V.roundToIntegral(APFloat::rmNearestTiesToAway);
V.roundToIntegral(APFloat::rmNearestTiesToEven);
return ConstantFP::get(Ty->getContext(), V);
}

Expand Down
48 changes: 48 additions & 0 deletions llvm/test/Transforms/InstSimplify/const-fold-nvvm-unary-arithmetic.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -416,6 +416,54 @@ define float @test_round_ftz_f_neg_1_5() {
ret float %res
}

define double @test_round_d_2_5() {
; CHECK-LABEL: define double @test_round_d_2_5() {
; CHECK-NEXT: ret double 2.000000e+00
;
%res = call double @llvm.nvvm.round.d(double 2.5)
ret double %res
}

define float @test_round_f_2_5() {
; CHECK-LABEL: define float @test_round_f_2_5() {
; CHECK-NEXT: ret float 2.000000e+00
;
%res = call float @llvm.nvvm.round.f(float 2.5)
ret float %res
}

define float @test_round_ftz_f_2_5() {
; CHECK-LABEL: define float @test_round_ftz_f_2_5() {
; CHECK-NEXT: ret float 2.000000e+00
;
%res = call float @llvm.nvvm.round.ftz.f(float 2.5)
ret float %res
}

define double @test_round_d_neg_2_5() {
; CHECK-LABEL: define double @test_round_d_neg_2_5() {
; CHECK-NEXT: ret double -2.000000e+00
;
%res = call double @llvm.nvvm.round.d(double -2.5)
ret double %res
}

define float @test_round_f_neg_2_5() {
; CHECK-LABEL: define float @test_round_f_neg_2_5() {
; CHECK-NEXT: ret float -2.000000e+00
;
%res = call float @llvm.nvvm.round.f(float -2.5)
ret float %res
}

define float @test_round_ftz_f_neg_2_5() {
; CHECK-LABEL: define float @test_round_ftz_f_neg_2_5() {
; CHECK-NEXT: ret float -2.000000e+00
;
%res = call float @llvm.nvvm.round.ftz.f(float -2.5)
ret float %res
}

define double @test_round_d_neg_subnorm() {
; CHECK-LABEL: define double @test_round_d_neg_subnorm() {
; CHECK-NEXT: ret double -0.000000e+00
Expand Down