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

[Simon Pilgrim via Phabricator via cfe-commits]

RKSimon added reviewers: efriedma, hfinkel.
RKSimon added a comment.

I'm all for keeping the scalar/vector behaviour the same but I'm concerned about constant folding not taking into account runtime rounding mode:

e,.g. SelectionDAG::getNode - we don't check the return status of convertFromAPInt from *INT_TO_FP - but then again the FP_TO_*INT constant folds only bail on invalid conversions.....

I haven't checked what other passes are doing.


Repository:
  rC Clang

https://reviews.llvm.org/D46863