aws / aws-ofi-nccl

This is a plugin which lets EC2 developers use libfabric as network provider while running NCCL applications.
Apache License 2.0
140 stars 54 forks source link

Cleanup nccl_ofi.h / utility macros. #432

Closed bwbarrett closed 3 months ago

bwbarrett commented 4 months ago

While working on the pthread wrapper patch (https://github.com/aws/aws-ofi-nccl/pull/431), I got caught up in having to include nccl_ofi.h (and then breaking some bits) for OFI_LIKELY, which lead me down a rabbit hole.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

bwbarrett commented 4 months ago

Approved in spirit, though some tests are failing.

@aws-nslick's "make distcheck" tests save the day. I had forgotten to include the new header in the dist tarball.

sunkuamzn commented 3 months ago

bot:aws:retest