Call .wait_tensor() in compiled region for dist.Work created in eager region #2485
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Outdated suggestions cannot be applied.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.
Summary:
In compiled region, instead of calling
dist.Work.wait()
, we will calltorch.ops._c10d_functional.wait_tensor()
on the dist.Work's output tensor. This way, we can capture thewait_tensor()
op within the torch.compile graph (instead of graph-breaking ondist.Work.wait()
), and the tensor will be waited on properly within the graph.This diff also depends on pytorch/pytorch#137763 to function properly.
Differential Revision: D64275115