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

More fixes and tweaks to the cuda conda packages #3281

Merged
merged 5 commits into from
Jun 5, 2019

Conversation

abergeron
Copy link
Contributor

This brings some fixes to the conda cuda build so that it works once again.

I've also added support for cudnn in those packages since anaconda started packaging it.

There should really be a phase in the tests that builds the packages. If there is interest I can make an attempt at adding that to the Jenkinsfile.

@abergeron
Copy link
Contributor Author

@tqchen

@tqchen
Copy link
Member

tqchen commented Jun 3, 2019

Thanks, @abergeron for the contribution! Yes, a CI script that verifies the build will be nice. One question about the docker env. We normally use https://github.com/dmlc/tvm/tree/master/docker as a way to cache pre-built docker images so we don't have to build again and avoids some problems when there is an upstream change, can we also do that in conda packaging?

@abergeron
Copy link
Contributor Author

It would certainly be possible to adapt my build scripts to generate the GPU docker images and pre-build them.

As for the CPU packages they are currently built as-is, but I could make a docker image for that too.

I can add that to this PR or make a new one.

@tqchen
Copy link
Member

tqchen commented Jun 3, 2019

a new PR would be great since we build and cache the docker images

@tqchen tqchen merged commit 1f62d95 into apache:master Jun 5, 2019
@tqchen
Copy link
Member

tqchen commented Jun 5, 2019

Thanks, @abergeron , this PR is now merged

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants