[llvm] [AMDGPU][NFC] Remove duplicate code by using getAddressableLocalMemor… (PR #104604)

via llvm-commits llvm-commits at lists.llvm.org
Fri Aug 16 08:18:41 PDT 2024


llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Mariusz Sikora (mariusz-sikora-at-amd)

<details>
<summary>Changes</summary>

…ySize

---
Full diff: https://github.com/llvm/llvm-project/pull/104604.diff


1 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp (+1-5) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp b/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
index 0ca6266cc678b6..5b41a2cd731607 100644
--- a/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
@@ -906,11 +906,7 @@ unsigned getWavefrontSize(const MCSubtargetInfo *STI) {
 }
 
 unsigned getLocalMemorySize(const MCSubtargetInfo *STI) {
-  unsigned BytesPerCU = 0;
-  if (STI->getFeatureBits().test(FeatureLocalMemorySize32768))
-    BytesPerCU = 32768;
-  if (STI->getFeatureBits().test(FeatureLocalMemorySize65536))
-    BytesPerCU = 65536;
+  unsigned BytesPerCU = getAddressableLocalMemorySize(STI);
 
   // "Per CU" really means "per whatever functional block the waves of a
   // workgroup must share". So the effective local memory size is doubled in

``````````

</details>


https://github.com/llvm/llvm-project/pull/104604


More information about the llvm-commits mailing list