Skip to content

Conversation

@awaelchli
Copy link
Contributor

What does this PR do?

Tries to fix #12314

Does your PR introduce any breaking changes? If yes, please list them.

Before submitting

  • Was this discussed/approved via a GitHub issue? (not for typos and docs)
  • Did you read the contributor guideline, Pull Request section?
  • Did you make sure your PR does only one thing, instead of bundling different changes together?
  • Did you make sure to update the documentation with your changes? (if necessary)
  • Did you write any new necessary tests? (not for typos and docs)
  • Did you verify new and existing tests pass locally with your changes?
  • Did you list all the breaking changes introduced by this pull request?
  • Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

  • Is this pull request ready for review? (if not, please submit in draft mode)
  • Check that all items from Before submitting are resolved
  • Make sure the title is self-explanatory and the description concisely explains the PR
  • Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

@awaelchli awaelchli added bug Something isn't working ci Continuous Integration priority: 0 High priority task labels Mar 15, 2022
@awaelchli awaelchli changed the title nonsense Fix CI failures due to Horovod and MPI dependencies Mar 15, 2022
@Borda
Copy link
Collaborator

Borda commented Mar 15, 2022

Seems to be parallel effort with #12318

@awaelchli
Copy link
Contributor Author

@Borda How do you force the GPU CI to use the changed docker build?

@Borda
Copy link
Collaborator

Borda commented Mar 15, 2022

@Borda How do you force the GPU CI to use the changed docker build?

No, because it doesn't compile correctly yet...

Copy link
Contributor

@carmocca carmocca left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is how you force the Docker images to get rebuilt and pushed to dockerhub:

cd83dd1

After they finish, you revert the above commit and CI should pull the image you just generated.

@mathemusician
Copy link
Contributor

👀

This reverts commit f3a9440.
@awaelchli awaelchli mentioned this pull request Mar 16, 2022
12 tasks
@awaelchli
Copy link
Contributor Author

My idea of including missing mpi libs did not work.

@mathemusician
Copy link
Contributor

“You think I’ll weep?
No, I’ll not weep. Storm and tempest.”

HOROVOD_WITHOUT_TENSORFLOW=1 \
HOROVOD_WITHOUT_MXNET=1 \
HOROVOD_WITH_GLOO=1 \
HOROVOD_WITHOUT_MPI=1
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Isn't this part of the problem?

@Borda Borda closed this in #12318 Mar 18, 2022
@Borda Borda deleted the ci/horovod branch November 5, 2022 19:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working ci Continuous Integration priority: 0 High priority task

Projects

None yet

Development

Successfully merging this pull request may close these issues.

AttributeError: module 'horovod.torch' has no attribute 'nccl_built'

4 participants