Skip to content

[SPIRV] Added support for the constrained arithmetic intrinsic #157441

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 4 commits into
base: main
Choose a base branch
from
+71 −0
Open

[SPIRV] Added support for the constrained arithmetic intrinsic #157441

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d512507
Added Support for the Constrained fmuladd
SubashBoopathi Aug 20, 2025
61dad88
Removed the legalization of the Opcode and used lower()
SubashBoopathi Oct 13, 2025
082c8aa
Slight changes in the format of legalizerinfo
SubashBoopathi Oct 13, 2025
b9d3905
Modified the legalization conditions to lower the instruction directly
SubashBoopathi Nov 12, 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
1 change: 1 addition & 0 deletions llvm/include/llvm/Support/TargetOpcodes.def
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -643,6 +643,7 @@ HANDLE_TARGET_OPCODE(G_FMA)

/// Generic FP multiply and add. Behaves as separate fmul and fadd.
HANDLE_TARGET_OPCODE(G_FMAD)
HANDLE_TARGET_OPCODE(G_STRICT_FMAD)

/// Generic FP division.
HANDLE_TARGET_OPCODE(G_FDIV)
Expand Down
1 change: 1 addition & 0 deletions llvm/include/llvm/Target/GenericOpcodes.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1716,6 +1716,7 @@ def G_STRICT_FREM : ConstrainedInstruction<G_FREM>;
def G_STRICT_FMA : ConstrainedInstruction<G_FMA>;
def G_STRICT_FSQRT : ConstrainedInstruction<G_FSQRT>;
def G_STRICT_FLDEXP : ConstrainedInstruction<G_FLDEXP>;
def G_STRICT_FMAD : ConstrainedInstruction<G_FMAD>;

//------------------------------------------------------------------------------
// Memory intrinsics
Expand Down
2 changes: 2 additions & 0 deletions llvm/lib/CodeGen/GlobalISel/IRTranslator.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2080,6 +2080,8 @@ static unsigned getConstrainedOpcode(Intrinsic::ID ID) {
return TargetOpcode::G_STRICT_FSQRT;
case Intrinsic::experimental_constrained_ldexp:
return TargetOpcode::G_STRICT_FLDEXP;
case Intrinsic::experimental_constrained_fmuladd:
return TargetOpcode::G_STRICT_FMAD;
default:
return 0;
}
Expand Down
1 change: 1 addition & 0 deletions llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4523,6 +4523,7 @@ LegalizerHelper::lower(MachineInstr &MI, unsigned TypeIdx, LLT LowerHintTy) {
return Legalized;
}
case TargetOpcode::G_FMAD:
case TargetOpcode::G_STRICT_FMAD:
return lowerFMad(MI);
case TargetOpcode::G_FFLOOR:
return lowerFFloor(MI);
Expand Down
2 changes: 2 additions & 0 deletions llvm/lib/Target/SPIRV/SPIRVLegalizerInfo.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -196,6 +196,8 @@ SPIRVLegalizerInfo::SPIRVLegalizerInfo(const SPIRVSubtarget &ST) {
getActionDefinitionsBuilder({G_FMA, G_STRICT_FMA})
.legalFor(allFloatScalarsAndVectors);

getActionDefinitionsBuilder(G_STRICT_FMAD).lower();

getActionDefinitionsBuilder(G_STRICT_FLDEXP)
.legalForCartesianProduct(allFloatScalarsAndVectors, allIntScalars);

Expand Down
64 changes: 64 additions & 0 deletions llvm/test/CodeGen/SPIRV/llvm-intrinsics/constrained-fmuladd.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,64 @@
; RUN: llc -mtriple=spirv64-unknown-unknown %s -o - | FileCheck %s
; RUN: %if spirv-tools %{ llc -mtriple=spirv64-unknown-unknown %s -o - -filetype=obj | spirv-val %}

; CHECK-DAG: OpDecorate %[[#]] FPRoundingMode RTE
; CHECK-DAG: OpDecorate %[[#]] FPRoundingMode RTZ
; CHECK-DAG: OpDecorate %[[#]] FPRoundingMode RTP
; CHECK-DAG: OpDecorate %[[#]] FPRoundingMode RTN
; CHECK-DAG: OpDecorate %[[#]] FPRoundingMode RTE

; CHECK: OpFMul %[[#]] %[[#]] %[[#]]
; CHECK: OpFAdd %[[#]] %[[#]] %[[#]]
define spir_kernel void @test_f32(float %a) {
entry:
%r = tail call float @llvm.experimental.constrained.fmuladd.f32(
float %a, float %a, float %a,
metadata !"round.tonearest", metadata !"fpexcept.strict")
ret void
}

; CHECK: OpFMul %[[#]] %[[#]] %[[#]]
; CHECK: OpFAdd %[[#]] %[[#]] %[[#]]
define spir_kernel void @test_f64(double %a) {
entry:
%r = tail call double @llvm.experimental.constrained.fmuladd.f64(
double %a, double %a, double %a,
metadata !"round.towardzero", metadata !"fpexcept.strict")
ret void
}

; CHECK: OpFMul %[[#]] %[[#]] %[[#]]
; CHECK: OpFAdd %[[#]] %[[#]] %[[#]]
define spir_kernel void @test_v2f32(<2 x float> %a) {
entry:
%r = tail call <2 x float> @llvm.experimental.constrained.fmuladd.v2f32(
<2 x float> %a, <2 x float> %a, <2 x float> %a,
metadata !"round.upward", metadata !"fpexcept.strict")
ret void
}

; CHECK: OpFMul %[[#]] %[[#]] %[[#]]
; CHECK: OpFAdd %[[#]] %[[#]] %[[#]]
define spir_kernel void @test_v4f32(<4 x float> %a) {
entry:
%r = tail call <4 x float> @llvm.experimental.constrained.fmuladd.v4f32(
<4 x float> %a, <4 x float> %a, <4 x float> %a,
metadata !"round.downward", metadata !"fpexcept.strict")
ret void
}

; CHECK: OpFMul %[[#]] %[[#]] %[[#]]
; CHECK: OpFAdd %[[#]] %[[#]] %[[#]]
define spir_kernel void @test_v2f64(<2 x double> %a) {
entry:
%r = tail call <2 x double> @llvm.experimental.constrained.fmuladd.v2f64(
<2 x double> %a, <2 x double> %a, <2 x double> %a,
metadata !"round.tonearest", metadata !"fpexcept.strict")
ret void
}

declare float @llvm.experimental.constrained.fmuladd.f32(float, float, float, metadata, metadata)
declare double @llvm.experimental.constrained.fmuladd.f64(double, double, double, metadata, metadata)
declare <2 x float> @llvm.experimental.constrained.fmuladd.v2f32(<2 x float>, <2 x float>, <2 x float>, metadata, metadata)
declare <4 x float> @llvm.experimental.constrained.fmuladd.v4f32(<4 x float>, <4 x float>, <4 x float>, metadata, metadata)
declare <2 x double> @llvm.experimental.constrained.fmuladd.v2f64(<2 x double>, <2 x double>, <2 x double>, metadata, metadata)
Loading