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

[Valery Pykhtin via Phabricator via llvm-commits]

vpykhtin added a comment.

In D136069#3862399 <https://reviews.llvm.org/D136069#3862399>, @kerbowa wrote:

> 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.

The number of reverted regions dropped from 34813 to 1289 in my testcase but they're RP measured anyway and this is strange.


Repository:
  rG LLVM Github Monorepo

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

https://reviews.llvm.org/D136069