[PATCH] D136069: [AMDGPU] Scheduler: Don't revert the schedule if the register pressure isn't changed for a region

[Austin Kerbow via Phabricator via llvm-commits]

kerbowa added a comment.

Interesting. Why exactly does this improve compile time so much? I thought reverting scheduling wasn't exactly expensive and the RP tracking was the problem.



================
Comment at: llvm/lib/Target/AMDGPU/GCNSchedStrategy.cpp:992
       DAG.RegionsWithExcessRP[RegionIdx]) {
     LLVM_DEBUG(dbgs() << "New pressure will result in more spilling.\n");
     return true;
----------------
arsenm wrote:
> arsenm wrote:
> > Maybe adjust the debug statement to say it's unchanged
> Plus this is mayCauseSpilling and falsely claiming true, should this check be in the caller?
Agreed. Can we put it in the base class shouldRevertScheduling?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D136069/new/

https://reviews.llvm.org/D136069