switch more test cases to use MultithreadTestCase #108744
Labels
good first issue
module: dtensor
distributed tensor tag
triaged
This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Comments
wanchaol
added
triaged
|
Hi @wanchaol ! I would like to work on this issue. |
|
@wanchaol Are people still working on this? Or can I pick it up ? |
|
Hello, if the bug is still unresolved, I would be interested in contributing a potential solution. Let me know if there's an opportunity to assist with resolving this issue. |
fkouteib
added a commit
to fkouteib/pytorch
that referenced
this issue
Apr 27, 2024
This conversion covers all tests under 'test/distributed/_tensor' directory. Fixes pytorch#108744
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
MultithreadTestCase allow us to run less resouce by spawning threads instead of processes, which could make distributed tests run faster. We have the following test files still not using MultithreadTestCase, and we should switch those test case to use it.
[ ] https://github.com/pytorch/pytorch/blob/main/test/distributed/_tensor/test_math_ops.py
[ ] https://github.com/pytorch/pytorch/blob/main/test/distributed/_tensor/test_matrix_ops.py
[ ] https://github.com/pytorch/pytorch/blob/main/test/distributed/_tensor/test_tensor_ops.py
[ ] https://github.com/pytorch/pytorch/blob/main/test/distributed/_tensor/test_embedding_ops.py
Example test case that already uses multithreaded test case, see https://github.com/pytorch/pytorch/blob/main/test/distributed/_tensor/test_pointwise_ops.py#L75
one just need to extend the
DTensorOpTestBasefor the above test files, should be relatively simpleThe text was updated successfully, but these errors were encountered: