-
Notifications
You must be signed in to change notification settings - Fork 4.7k
Fix for distributed tests on pytorch>=1.12 #2141
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
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jeffra
approved these changes
Jul 29, 2022
This was referenced Aug 3, 2022
This was referenced Sep 6, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
With the recent release of torch 1.12, we saw all unit tests using the
@distirbuted_testdecorator break (see this issue).The problem involves changes in
torch.multiprocessingandtorch.distributedthat prevents usingforkto spawn new processes in distributed tests. We must now usespawnorforkserverto spin up the distributed processes. However, this change alone is not sufficient as the decorator causes issues when pickling the target function (e.g.,dist_initortest_*):AttributeError: Can't pickle local objectand_pickle.PicklingError: Can't pickle <function test_example at 0x7f6c4e5640d0>: it's not the same object as unit.test_example.test_exampleerrors are unavoidable with the distributed test decorator.The solution implemented here loosely follows how pytorch unit tests are done and adds a
DistributedTestclass. This will require refactoring all distributed tests like so:@jeffra, @tjruwase, and I will be refactoring existing tests over time to use this new test class. In doing so, we are also organizing the unit tests files to match the directory structure of the DeepSpeed source code.