[clang] [llvm] [AMDGPU] modify named barrier builtins and intrinsics (PR #114550)

Matt Arsenault via cfe-commits cfe-commits at lists.llvm.org
Fri Nov 1 09:05:30 PDT 2024


================
@@ -920,6 +920,124 @@ class AMDGPULowerModuleLDS {
     return KernelToCreatedDynamicLDS;
   }
 
+  static GlobalVariable *uniquifyGVPerKernel(Module &M, GlobalVariable *GV,
----------------
arsenm wrote:

This still should be a separate PR 

https://github.com/llvm/llvm-project/pull/114550


More information about the cfe-commits mailing list