[PATCH] D152649: [AMDGPU] Enable Atomic Optimizer and Default to Iterative Scan Strategy.

[Pravin Jagtap via Phabricator via llvm-commits]

pravinjagtap added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp:278-281
 static cl::opt<bool>
     EnableAtomicOptimizations("amdgpu-atomic-optimizations",
                               cl::desc("Enable atomic optimizations"),
+                              cl::init(true), cl::Hidden);
----------------
arsenm wrote:
> The old option can be dropped, none should be a value for the strategy 
> The old option can be dropped, none should be a value for the strategy 

I think, dropping the old option (`amdgpu-atomic-optimizations`) will break LLPC at the moment. 

I think right way to do this is, transfer the responsibility of enabling the pass to `amdgpu-atomic-optimizer-strategy` and keep the old option `amdgpu-atomic-optimizations` till LLPC removes all the usage of it.

@foad 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152649/new/

https://reviews.llvm.org/D152649