[PATCH] D85759: [SLPVectorizer] Fix regression from cost model refactoring

[Sam Parker via Phabricator via llvm-commits]

samparker added a comment.

> I took a little more look at this and all of this cost model is very confusing.

It is... people have lazily added stuff to TTI so I spent a few months trying to clean it up and the intrinsics was by far the hardest, and it's still a mess! I don't know why there was ever distinct paths for types and/or arguments, I think it may have grown from scalar and vector code paths... But the fact that it's still a mess is probably a different discussion from what this patch is trying to achieve though. This change looks good to me, purely because it looks like I just really messed up and this fixes it...


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D85759/new/

https://reviews.llvm.org/D85759