[PATCH] D46863: [X86] Use __builtin_convertvector to implement some of the packed integer to packed flow conversion intrinsics.

[Eli Friedman via Phabricator via cfe-commits]

efriedma added a comment.

> I'm concerned about constant folding not taking into account runtime rounding mode

Changing the rounding mode is UB without FENV_ACCESS.  And with FENV_ACCESS, __builtin_convertvector should lower to @llvm.experimental.constrained.sitofp etc., which won't constant-fold.  (llvm.experimental.constrained.sitofp doesn't actually exist yet, but I assume it will eventually get added.)


Repository:
  rC Clang

https://reviews.llvm.org/D46863