[PATCH] D147114: [LV] Use BFI to adjust cost of predicated instructions

[Dave Green via Phabricator via llvm-commits]

dmgreen added a comment.

The other thing I thought of was about scalar costs rounding down towards 0, as they are all stored as integers. I have no data to suggest that any one scheme is better or worse than another, but it might make sense to round towards nearest or round up. I'm just imagining costs possibly rounding to 0 in to many cases when the block frequencies are quite different.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D147114/new/

https://reviews.llvm.org/D147114