Skip to content

AMDGPU: Disable sincos fold for constant inputs #134579

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
merged 1 commit into from
Apr 7, 2025
Merged

AMDGPU: Disable sincos fold for constant inputs #134579

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
5 changes: 5 additions & 0 deletions llvm/lib/Target/AMDGPU/AMDGPULibCalls.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1361,6 +1361,11 @@ bool AMDGPULibCalls::fold_sincos(FPMathOperator *FPOp, IRBuilder<> &B,
bool const isSin = fInfo.getId() == AMDGPULibFunc::EI_SIN;

Value *CArgVal = FPOp->getOperand(0);

// TODO: Constant fold the call
if (isa<ConstantData>(CArgVal))
return false;

CallInst *CI = cast<CallInst>(FPOp);

Function *F = B.GetInsertBlock()->getParent();
Expand Down
8 changes: 3 additions & 5 deletions llvm/test/CodeGen/AMDGPU/amdgpu-simplify-libcall-sincos.defined.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,12 +125,10 @@ define void @sincos_f32_value_is_same_constantfp(ptr addrspace(1) nocapture writ
; CHECK-LABEL: define void @sincos_f32_value_is_same_constantfp
; CHECK-SAME: (ptr addrspace(1) writeonly captures(none) [[SIN_OUT:%.*]], ptr addrspace(1) writeonly captures(none) [[COS_OUT:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[__SINCOS_:%.*]] = alloca float, align 4, addrspace(5)
; CHECK-NEXT: [[TMP0:%.*]] = call contract float @_Z6sincosfPU3AS5f(float 4.200000e+01, ptr addrspace(5) [[__SINCOS_]])
; CHECK-NEXT: [[TMP1:%.*]] = load float, ptr addrspace(5) [[__SINCOS_]], align 4
; CHECK-NEXT: store float [[TMP0]], ptr addrspace(1) [[SIN_OUT]], align 4
; CHECK-NEXT: [[CALL:%.*]] = tail call contract float @_Z3sinf(float 4.200000e+01)
; CHECK-NEXT: store float [[CALL]], ptr addrspace(1) [[SIN_OUT]], align 4
; CHECK-NEXT: [[CALL1:%.*]] = tail call contract float @_Z3cosf(float 4.200000e+01)
; CHECK-NEXT: store float [[TMP1]], ptr addrspace(1) [[COS_OUT]], align 4
; CHECK-NEXT: store float [[CALL1]], ptr addrspace(1) [[COS_OUT]], align 4
; CHECK-NEXT: ret void
;
entry:
Expand Down
11 changes: 5 additions & 6 deletions llvm/test/CodeGen/AMDGPU/amdgpu-simplify-libcall-sincos.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -745,13 +745,12 @@ entry:

define void @sincos_f32_value_is_same_constantfp(ptr addrspace(1) nocapture writeonly %sin_out, ptr addrspace(1) nocapture writeonly %cos_out) {
; CHECK-LABEL: define void @sincos_f32_value_is_same_constantfp
; CHECK-SAME: (ptr addrspace(1) writeonly captures(none) initializes((0, 4)) [[SIN_OUT:%.*]], ptr addrspace(1) writeonly captures(none) initializes((0, 4)) [[COS_OUT:%.*]]) local_unnamed_addr #[[ATTR3]] {
; CHECK-SAME: (ptr addrspace(1) writeonly captures(none) initializes((0, 4)) [[SIN_OUT:%.*]], ptr addrspace(1) writeonly captures(none) initializes((0, 4)) [[COS_OUT:%.*]]) local_unnamed_addr #[[ATTR2]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[__SINCOS_:%.*]] = alloca float, align 4, addrspace(5)
; CHECK-NEXT: [[TMP0:%.*]] = call contract float @_Z6sincosfPU3AS5f(float 4.200000e+01, ptr addrspace(5) [[__SINCOS_]])
; CHECK-NEXT: [[TMP1:%.*]] = load float, ptr addrspace(5) [[__SINCOS_]], align 4
; CHECK-NEXT: store float [[TMP0]], ptr addrspace(1) [[SIN_OUT]], align 4
; CHECK-NEXT: store float [[TMP1]], ptr addrspace(1) [[COS_OUT]], align 4
; CHECK-NEXT: [[CALL:%.*]] = tail call contract float @_Z3sinf(float 4.200000e+01)
; CHECK-NEXT: store float [[CALL]], ptr addrspace(1) [[SIN_OUT]], align 4
; CHECK-NEXT: [[CALL1:%.*]] = tail call contract float @_Z3cosf(float 4.200000e+01)
; CHECK-NEXT: store float [[CALL1]], ptr addrspace(1) [[COS_OUT]], align 4
; CHECK-NEXT: ret void
;
entry:
Expand Down