Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

WIP: Add nccl #8637

Closed
wants to merge 22 commits into from
Closed

WIP: Add nccl #8637

wants to merge 22 commits into from

Conversation

jakirkham
Copy link
Member

This is still very much WIP (because of the staged-recipes changes needed). Please do not review yet. Thanks. :)

This adds a package for NVIDIA's NCCL library. Requires the nvcc shim package for building. ( #8229 ) Leverages the recently added CUDA Docker images for building. ( conda-forge/docker-images#93 )

Note: This includes some other stuff that I used to build this locally. Please ignore that for the moment. It can be cleaned out later.

cc @jjhelmus @seibert @dantegd @JohnZed @kkraus14

Checklist

  • License file is packaged (see here for an example)
  • Source is from official source
  • Package does not vend other packages
  • Build number is 0
  • GitHub users listed in the maintainer section have posted a comment confirming they are willing to be listed there

This adds a shim package to help conda-build interact with an
**existing** nvcc install. Through this shim compiler package,
users are able to ensure an appropriate cudatoolkit version is
attached to the package and thus make sure that requirement is
met at install time.
Needed so packages compiled with them can find them easily.
Make sure that `libcudart` is a shared library dependency.
After talking with one of the NCCL developers, they confirmed that
NCCL will only make breaking changes at major versions. So have added
`run_exports` with a loose constraint here and reference to the
upstream issue where this was discussed.
@conda-forge-linter
Copy link

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipes/nccl, recipes/nvcc) and found it was in an excellent condition.

@hmaarrfk
Copy link
Contributor

What's the feeling on merging this in? Hardware support sounds super interesting

@jakirkham jakirkham mentioned this pull request Oct 1, 2019
6 tasks
@jakirkham
Copy link
Member Author

I'm replacing this with PR ( #9694 ). Please take a look 🙂

@jakirkham jakirkham deleted the add_nccl branch October 22, 2019 15:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Development

Successfully merging this pull request may close these issues.

3 participants