[PATCH] D134664: [nfc][mlgo] Lazily compute the regalloc reward

[Mircea Trofin via Phabricator via llvm-commits]

mtrofin added a comment.

In D134664#3816089 <https://reviews.llvm.org/D134664#3816089>, @jacobhegna wrote:

> so the point of this is just to avoid recomputing the reward for the two different regalloc models? the reward will always be computed the same way for both models?

the point is to avoid computing if noone needs it - which is always the case except when training.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D134664/new/

https://reviews.llvm.org/D134664