[PATCH] D100281: [AMDGPU] Revise handling of preexisting waitcnt

[Austin Kerbow via Phabricator via llvm-commits]

kerbowa added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp:803
+/// \p OldWaitcntInstr and apply any extra wait from waitcnt that were added
+/// by previous passes. Currently this pass conservatively assumes that these
+/// preexisting waitcnt are required for correctness.
----------------
t-tye wrote:
> It seems it could delete them if it can prove they have no effect on execution due to no memory operations existing between them and a previous waitcnt. But if did that it must respect the WB/INV instructions that also effect waitcnts.
I'm planning to do this in a follow-up change.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D100281/new/

https://reviews.llvm.org/D100281