[llvm] c82a6a0 - [AMDGPU] Use correct vector elt type when shrinking mfma scale (#123043)

[via llvm-commits]

Author: Fraser Cormack
Date: 2025-01-15T14:28:42Z
New Revision: c82a6a025179e3b155c70f5ad8f84fa8ec2a9452

URL: https://github.com/llvm/llvm-project/commit/c82a6a025179e3b155c70f5ad8f84fa8ec2a9452
DIFF: https://github.com/llvm/llvm-project/commit/c82a6a025179e3b155c70f5ad8f84fa8ec2a9452.diff

LOG: [AMDGPU] Use correct vector elt type when shrinking mfma scale (#123043)

This might be a copy/paste error. I don't think this an issue in
practice as the builtins/intrinsics are only legal with identical vector
element types.

Added: 
    

Modified: 
    llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp b/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
index 8b1b398606583e..bac3bb5fde7b02 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
@@ -1346,7 +1346,7 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
 
     if (Src1Ty->getNumElements() > Src1NumElts) {
       Src1 = IC.Builder.CreateExtractVector(
-          FixedVectorType::get(Src0Ty->getElementType(), Src1NumElts), Src1,
+          FixedVectorType::get(Src1Ty->getElementType(), Src1NumElts), Src1,
           IC.Builder.getInt64(0));
       MadeChange = true;
     }