Skip to content

Conversation

@changpeng
Copy link
Contributor

No description provided.

@llvmbot
Copy link
Member

llvmbot commented Jul 15, 2025

@llvm/pr-subscribers-llvm-analysis
@llvm/pr-subscribers-backend-amdgpu

@llvm/pr-subscribers-llvm-ir

Author: Changpeng Fang (changpeng)

Changes

Full diff: https://github.com/llvm/llvm-project/pull/148989.diff

2 Files Affected:

  • (modified) llvm/include/llvm/IR/IntrinsicsAMDGPU.td (-1)
  • (modified) llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp (-1)
diff --git a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
index acd10af1709ea..eb2c63c24e4b5 100644
--- a/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
+++ b/llvm/include/llvm/IR/IntrinsicsAMDGPU.td
@@ -3710,7 +3710,6 @@ class AMDGPUWmmaIntrinsicModsAllDiff<LLVMType DstTy, LLVMType AB, LLVMType C> :
 >;
 
 defset list<Intrinsic> AMDGPUWMMAIntrinsicsGFX1250 = {
-def int_amdgcn_wmma_f64_16x16x4_f64       : AMDGPUWmmaIntrinsicModsAll<llvm_anyfloat_ty, llvm_anyfloat_ty>;
 def int_amdgcn_wmma_f32_16x16x4_f32       : AMDGPUWmmaIntrinsicModsAllReuse<llvm_anyfloat_ty, llvm_anyfloat_ty>;
 def int_amdgcn_wmma_f32_16x16x32_bf16     : AMDGPUWmmaIntrinsicModsAllReuse<llvm_anyfloat_ty, llvm_anyfloat_ty>;
 def int_amdgcn_wmma_f32_16x16x32_f16      : AMDGPUWmmaIntrinsicModsAllReuse<llvm_anyfloat_ty, llvm_anyfloat_ty>;
diff --git a/llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp
index 3cc4f3ace9686..cbbb57c6f8122 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp
@@ -4689,7 +4689,6 @@ AMDGPURegisterBankInfo::getInstrMapping(const MachineInstr &MI) const {
     case Intrinsic::amdgcn_swmmac_f32_16x16x32_fp8_bf8:
     case Intrinsic::amdgcn_swmmac_f32_16x16x32_bf8_fp8:
     case Intrinsic::amdgcn_swmmac_f32_16x16x32_bf8_bf8:
-    case Intrinsic::amdgcn_wmma_f64_16x16x4_f64:
     case Intrinsic::amdgcn_wmma_f32_16x16x4_f32:
     case Intrinsic::amdgcn_wmma_f32_16x16x32_bf16:
     case Intrinsic::amdgcn_wmma_f32_16x16x32_f16:

@changpeng changpeng requested a review from rampitec July 15, 2025 23:20
@llvmbot llvmbot added the llvm:analysis Includes value tracking, cost tables and constant folding label Jul 16, 2025
@arsenm arsenm changed the title AMDGPU: Remove a non-existence wmma instruction from gfx1250 AMDGPU: Remove a non-existent wmma instruction from gfx1250 Jul 16, 2025
@changpeng changpeng merged commit 01dd892 into llvm:main Jul 16, 2025
10 checks passed
@changpeng changpeng deleted the removal branch July 16, 2025 01:16
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

backend:AMDGPU llvm:analysis Includes value tracking, cost tables and constant folding llvm:ir

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants