[PATCH] D43142: Experimental pass to convert all floating point operations to the equivalent constrained intrinsics

[Cameron McInally via Phabricator via llvm-commits]

cameron.mcinally added a comment.

In https://reviews.llvm.org/D43142#1284296, @kpn wrote:

> Then don't we have to deal with this sometime? If #pragma FENV_ACCESS ON is set for the main() example above then is constant folding allowed?


We do not want to constant fold when FENV_ACCESS=ON. We want the hardware instruction to execute and set the flags at runtime.

It's not really a problem though, since the constrained intrinsics won't constant fold.


https://reviews.llvm.org/D43142