[PATCH] D50991: [AMDGPU] Consider loads from flat addrspace to be potentially divergent

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp:556
+    return Load->getPointerAddressSpace() == ST->getAMDGPUAS().PRIVATE_ADDRESS
+      || Load->getPointerAddressSpace() == ST->getAMDGPUAS().FLAT_ADDRESS;
 
----------------
|| on pervious line


================
Comment at: test/CodeGen/AMDGPU/divergent-flat.ll:1-2
+; RUN: llc -march=amdgcn -mcpu=gfx900 -o - %s | FileCheck %s --check-prefix=ASM
+; RUN: llc -march=amdgcn -mcpu=gfx900 -stop-after=structurizecfg -o - %s | FileCheck %s --check-prefix=STRUCTURIZECFG
+
----------------
Test can be a lot simpler and put in test/Analysis/DivergenceAnalysis/AMDGPU with the rest


Repository:
  rL LLVM

https://reviews.llvm.org/D50991