[Mlir-commits] [mlir] Implement gpu.subgroup_reduce with DPP intrinsics on AMD GPUs (PR #133204)

[Krzysztof Drewniak]

================
@@ -1214,6 +1214,7 @@ struct ConvertAMDGPUToROCDLPass
   using Base::Base;
 
   void runOnOperation() override {
+    llvm::errs() << " WHEN DOES AMDGPU TO ROCDL RUN\n";
----------------
krzysz00 wrote:

Nit: you left a debug print in

https://github.com/llvm/llvm-project/pull/133204