Page MenuHomePhabricator

[mlir][rocdl] Add math::Exp2Op lowering to ROCDL
ClosedPublic

Authored by akuegel on Jul 15 2021, 4:37 AM.

Details

Summary

There exists already a lowering to LLVM::Exp2Op.
On GPU we prefer to lower to an intrinsic instead.

Diff Detail

Event Timeline

akuegel created this revision.Jul 15 2021, 4:37 AM
akuegel requested review of this revision.Jul 15 2021, 4:37 AM
herhut accepted this revision.Jul 15 2021, 4:40 AM

Thanks!

This revision is now accepted and ready to land.Jul 15 2021, 4:40 AM
This revision was automatically updated to reflect the committed changes.