Skip to content

AMDGPU: Replace unused update.dpp inputs with poison instead of undef #131287

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
Mar 18, 2025
Merged

AMDGPU: Replace unused update.dpp inputs with poison instead of undef #131287

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
4 changes: 2 additions & 2 deletions llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1103,11 +1103,11 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
auto *RM = cast<ConstantInt>(II.getArgOperand(3));
auto *BM = cast<ConstantInt>(II.getArgOperand(4));
if (BC->isZeroValue() || RM->getZExtValue() != 0xF ||
BM->getZExtValue() != 0xF || isa<UndefValue>(Old))
BM->getZExtValue() != 0xF || isa<PoisonValue>(Old))
break;

// If bound_ctrl = 1, row mask = bank mask = 0xf we can omit old value.
return IC.replaceOperand(II, 0, UndefValue::get(Old->getType()));
return IC.replaceOperand(II, 0, PoisonValue::get(Old->getType()));
}
case Intrinsic::amdgcn_permlane16:
case Intrinsic::amdgcn_permlane16_var:
Expand Down
4 changes: 2 additions & 2 deletions llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3232,7 +3232,7 @@ define amdgpu_kernel void @update_dpp_no_combine(ptr addrspace(1) %out, i32 %in1

define amdgpu_kernel void @update_dpp_drop_old(ptr addrspace(1) %out, i32 %in1, i32 %in2) {
; CHECK-LABEL: @update_dpp_drop_old(
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @llvm.amdgcn.update.dpp.i32(i32 undef, i32 [[IN2:%.*]], i32 3, i32 15, i32 15, i1 true)
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @llvm.amdgcn.update.dpp.i32(i32 poison, i32 [[IN2:%.*]], i32 3, i32 15, i32 15, i1 true)
; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[OUT:%.*]], align 4
; CHECK-NEXT: ret void
;
Expand All @@ -3243,7 +3243,7 @@ define amdgpu_kernel void @update_dpp_drop_old(ptr addrspace(1) %out, i32 %in1,

define amdgpu_kernel void @update_dpp_undef_old(ptr addrspace(1) %out, i32 %in1) {
; CHECK-LABEL: @update_dpp_undef_old(
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @llvm.amdgcn.update.dpp.i32(i32 undef, i32 [[IN1:%.*]], i32 4, i32 15, i32 15, i1 true)
; CHECK-NEXT: [[TMP0:%.*]] = call i32 @llvm.amdgcn.update.dpp.i32(i32 poison, i32 [[IN1:%.*]], i32 4, i32 15, i32 15, i1 true)
; CHECK-NEXT: store i32 [[TMP0]], ptr addrspace(1) [[OUT:%.*]], align 4
; CHECK-NEXT: ret void
;
Expand Down
Loading