This is an archive of the discontinued LLVM Phabricator instance.

[ARM] Permit auto-vectorization using MVE
ClosedPublic

Authored by dmgreen on Jun 24 2019, 10:16 AM.

Details

Summary

Correctly report the number of registers available to the auto-vectorizer. Also allow FP auto-vectorization for MVE without -Ofast/-ffast-math, due to support for IEEE FP arithmetic and parity between scalar and vector FP behaviour.

Patch by David Sherwood.

Diff Detail

Repository
rL LLVM

Event Timeline

dmgreen created this revision.Jun 24 2019, 10:16 AM

This really depends on a lot of instruction selection being in place first.

OK. It's time. Lets do this.

Yep :-)

It's difficult to imagine, but maybe wait a day with committing this in case someone is unhappy or has doubts.

SjoerdMeijer accepted this revision.Aug 8 2019, 8:36 AM
This revision is now accepted and ready to land.Aug 8 2019, 8:36 AM
This revision was automatically updated to reflect the committed changes.