[PATCH] D98706: [SCEV] Add false->any implication to save compile time

[Philip Reames via Phabricator via llvm-commits]

reames added a comment.

This seems like a useful change to me purely on the transform impact.  The compile time appears to be a wash.  A refreshed patch which simply removed the misleading compile time comment would likely warrant an LGTM.

I would suggest auditing the callers to make sure perform checks in the right order though.  This could be surprising if the caller isn't expecting to be analyzing potentially dead code.

Maybe even we should assert this condition?  Not because it's wrong, but passing a trivial condition here really seems like a missing compile time optimization in each caller?  (This is only a lightly considered suggestion, feel free to ignore if a glance at code shows me off base.)


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98706/new/

https://reviews.llvm.org/D98706