[PATCH] D53157: Teach the IRBuilder about constrained fadd and friends

[Cameron McInally via Phabricator via llvm-commits]

cameron.mcinally added subscribers: craig.topper, uweigand, andrew.w.kaylor, hfinkel.
cameron.mcinally added a comment.

Adding the usual suspects: @andrew.w.kaylor @craig.topper @uweigand @hfinkel

I don't see the benefit of this change. The intrinsics are good enough to be functional right now (minus a few that are in progress), so I'm not sure we need IRBuilder functions. Am I missing something?

That said, the constrained intrinsics aren't a good long term solution for the FPEnv project. IMO, we'll hit a wall once we try to optimize them. Not sure if this patch helps with that problem though. We really need to toggle optimizations that can clobber side-effects on a case-by-case basis.


https://reviews.llvm.org/D53157