[PATCH] D62757: [X86] Mutate fceil/ffloor/ftrunc/fnearbyint/frint into X86ISD::RNDSCALE during PreProcessIselDAG to cut down on pattern permutations

Craig Topper via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sun Jun 2 11:54:48 PDT 2019

craig.topper added a comment.

In D62757#1526692 <https://reviews.llvm.org/D62757#1526692>, @RKSimon wrote:

> I've no objection to this - what are you intending to do with the scalar versions? And where is the select code for the strict versions?

I think for scalar we would need to add patterns for a scalar X86ISD::RNDSCALE in addition to the X86ISD::RNDSCALES we use for intrinsics. Maybe we could use extractelement and movss/movsd matching to remove RNDSCALES?

StrictFP doesn't currently have proper selection code. Just before calling Select, SelectionDAGISel mutates them to remove the chain input/output which was a temporary solution. That will ultimately be removed.



More information about the llvm-commits mailing list