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

process_group: support all PG APIs #97

Closed
d4l3k opened this issue Feb 4, 2025 · 0 comments · Fixed by #114
Closed

process_group: support all PG APIs #97

d4l3k opened this issue Feb 4, 2025 · 0 comments · Fixed by #114
Labels
enhancement New feature or request

Comments

@d4l3k
Copy link
Member

d4l3k commented Feb 4, 2025

We currently only support a subset of collective APIs in torchft.ProcessGroup. We want to support all of them in all of our ProcessGroup implementations.

This can be done piece by piece and we don't need to add all of them right now.

The API definition for ProcessGroup is located at https://github.com/pytorch/torchft/blob/main/torchft/process_group.py#L123

We need to support these for at least ProcessGroupWrapper, ManagedProcessGroup, ProcessGroupBaby*.

We want to support all of the collectives in PyProcessGroup: https://github.com/pytorch/pytorch/blob/11f69808c64a65c68a4452250ba7719dcff27c78/torch/csrc/distributed/c10d/PyProcessGroup.hpp

  • allgather_into_tensor_coalesced
  • allreduce_coalesced
  • alltoall_base
  • barrier
  • reduce_scatter
  • reduce_scatter_tensor_coalesced
  • send/recv (trickier)

Testing:

We have an existing test suite for collectives. We should add tests for these new types to _test_pg and make sure all of our PG implementations support them.

https://github.com/pytorch/torchft/blob/main/torchft/process_group_test.py#L96

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
1 participant