Skip to content

AMDGPU: Replace unused export inputs with poison instead of undef #131286

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

Jump to bottom
Merged
merged 1 commit into from
Mar 18, 2025
Merged

AMDGPU: Replace unused export inputs with poison instead of undef #131286

merged 1 commit into from
Mar 18, 2025

Conversation

arsenm
Copy link
Contributor

@arsenm arsenm commented Mar 14, 2025

No description provided.

@arsenm arsenm marked this pull request as ready for review March 14, 2025 07:13
@arsenm Graphite App
Copy link
Contributor Author

arsenm commented Mar 14, 2025 •
edited
Loading

This stack of pull requests is managed by Graphite. Learn more about stacking.

This was referenced Mar 14, 2025
Merged
Merged
Merged
@llvmbot llvmbot added llvm:instcombine Covers the InstCombine, InstSimplify and AggressiveInstCombine passes llvm:transforms labels Mar 14, 2025
@llvmbot
Copy link
Member

llvmbot commented Mar 14, 2025 •
edited
Loading

@llvm/pr-subscribers-llvm-transforms

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

Full diff: https://github.com/llvm/llvm-project/pull/131286.diff

2 Files Affected:

  • (modified) llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp (+2-2)
  • (modified) llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll (+21-21)
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp b/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
index 0bbded6a3422a..be1274e181ce2 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
@@ -806,8 +806,8 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
       if ((!IsCompr && (EnBits & (1 << I)) == 0) ||
           (IsCompr && ((EnBits & (0x3 << (2 * I))) == 0))) {
         Value *Src = II.getArgOperand(I + 2);
-        if (!isa<UndefValue>(Src)) {
-          IC.replaceOperand(II, I + 2, UndefValue::get(Src->getType()));
+        if (!isa<PoisonValue>(Src)) {
+          IC.replaceOperand(II, I + 2, PoisonValue::get(Src->getType()));
           Changed = true;
         }
       }
diff --git a/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll b/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
index 124ea57be35e9..9a42fa723dc48 100644
--- a/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
+++ b/llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-intrinsics.ll
@@ -1776,18 +1776,18 @@ declare void @llvm.amdgcn.exp.f32(i32 immarg, i32 immarg, float, float, float, f
 define void @exp_disabled_inputs_to_undef(float %x, float %y, float %z, float %w) {
   ; enable src0..src3 constants
 ; CHECK-LABEL: @exp_disabled_inputs_to_undef(
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float 1.000000e+00, float undef, float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float undef, float 2.000000e+00, float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 4, float undef, float undef, float 5.000000e-01, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 8, float undef, float undef, float undef, float 4.000000e+00, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float [[X:%.*]], float undef, float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float undef, float [[Y:%.*]], float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 4, float undef, float undef, float [[Z:%.*]], float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 8, float undef, float undef, float undef, float [[W:%.*]], i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 0, float undef, float undef, float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 3, float 1.000000e+00, float 2.000000e+00, float undef, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 5, float 1.000000e+00, float undef, float 5.000000e-01, float undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 9, float 1.000000e+00, float undef, float undef, float 4.000000e+00, i1 false, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float 1.000000e+00, float poison, float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float poison, float 2.000000e+00, float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 4, float poison, float poison, float 5.000000e-01, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 8, float poison, float poison, float poison, float 4.000000e+00, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float [[X:%.*]], float poison, float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float poison, float [[Y:%.*]], float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 4, float poison, float poison, float [[Z:%.*]], float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 8, float poison, float poison, float poison, float [[W:%.*]], i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 0, float poison, float poison, float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 3, float 1.000000e+00, float 2.000000e+00, float poison, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 5, float 1.000000e+00, float poison, float 5.000000e-01, float poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 9, float 1.000000e+00, float poison, float poison, float 4.000000e+00, i1 false, i1 false)
 ; CHECK-NEXT:    call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float 1.000000e+00, float 2.000000e+00, float 5.000000e-01, float 4.000000e+00, i1 false, i1 false)
 ; CHECK-NEXT:    ret void
 ;
@@ -1824,15 +1824,15 @@ declare void @llvm.amdgcn.exp.compr.v2f16(i32 immarg, i32 immarg, <2 x half>, <2
 
 define void @exp_compr_disabled_inputs_to_undef(<2 x half> %xy, <2 x half> %zw) {
 ; CHECK-LABEL: @exp_compr_disabled_inputs_to_undef(
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 0, <2 x half> undef, <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 1, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 2, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 3, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 0, <2 x half> undef, <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 1, <2 x half> [[XY:%.*]], <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 2, <2 x half> [[XY]], <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 3, <2 x half> [[XY]], <2 x half> undef, i1 true, i1 false)
-; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 12, <2 x half> undef, <2 x half> [[ZW:%.*]], i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 0, <2 x half> poison, <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 1, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 2, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 3, <2 x half> <half 0xH3C00, half 0xH4000>, <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 0, <2 x half> poison, <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 1, <2 x half> [[XY:%.*]], <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 2, <2 x half> [[XY]], <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 3, <2 x half> [[XY]], <2 x half> poison, i1 true, i1 false)
+; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 12, <2 x half> poison, <2 x half> [[ZW:%.*]], i1 true, i1 false)
 ; CHECK-NEXT:    call void @llvm.amdgcn.exp.compr.v2f16(i32 0, i32 15, <2 x half> [[XY]], <2 x half> [[ZW]], i1 true, i1 false)
 ; CHECK-NEXT:    ret void
 ;

@arsenm Graphite App
Copy link
Contributor Author

arsenm commented Mar 18, 2025 •
edited
Loading

Merge activity

  • Mar 18, 6:23 AM EDT: A user started a stack merge that includes this pull request via Graphite.
  • Mar 18, 6:27 AM EDT: Graphite rebased this pull request as part of a merge.
  • Mar 18, 6:30 AM EDT: A user merged this pull request with Graphite.

@arsenm arsenm force-pushed the users/arsenm/amdgpu/promote-alloca-use-freeze-poison-instead-of-undef branch from bdbf4ed to b8b5105 Compare March 18, 2025 10:24
Base automatically changed from users/arsenm/amdgpu/promote-alloca-use-freeze-poison-instead-of-undef to main March 18, 2025 10:27
@arsenm arsenm force-pushed the users/arsenm/amdgpu/replace-unused-exp-with-poison-instead-of-undef branch from b323f0d to 7893fd4 Compare March 18, 2025 10:27
@arsenm arsenm merged commit 8392573 into main Mar 18, 2025
6 of 9 checks passed
@arsenm arsenm deleted the users/arsenm/amdgpu/replace-unused-exp-with-poison-instead-of-undef branch March 18, 2025 10:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jayfoad jayfoad jayfoad approved these changes

@nunoplopes nunoplopes Awaiting requested review from nunoplopes

@perlfu perlfu Awaiting requested review from perlfu

@Pierre-vh Pierre-vh Awaiting requested review from Pierre-vh

@sgundapa sgundapa Awaiting requested review from sgundapa

@shiltian shiltian Awaiting requested review from shiltian

Assignees
No one assigned
Labels
backend:AMDGPU llvm:instcombine Covers the InstCombine, InstSimplify and AggressiveInstCombine passes llvm:transforms
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@arsenm @llvmbot @jayfoad