Update send/recv tests to reflect intended use #11275
Closed
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
pietern
requested review from
colesbury,
ezyang,
gchanan,
soumith and
zdevito
as code owners
Contributor
Author
|
FYI, I'm waiting for CI to confirm the tests don't break for MPI and/or THD. I have confirmed they work with the upcoming gloo send/recv support. |
Contributor
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
pietern has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
The existing tests had every rank run send to every other rank and only then switch to recv mode. This only works if the send operations are non-blocking and the passed tensors are immediately copied to some kind of send buffer. Instead, every send must be matched with a recv on the other side, because from the API perspective they may block. E.g. imagine a 1GB tensor being sent to every other rank. It can only go through if there is a recv on the other side, or it will deadlock. This change reflects this in the send/recv unit tests.
pietern
force-pushed
the
distributed-send-recv-test
branch
from
37317d2 to
8b7bf53
Compare
Contributor
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
pietern has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
PenghuiCheng
pushed a commit
to PenghuiCheng/pytorch
that referenced
this pull request
Sep 11, 2018
Summary: The existing tests had every rank run send to every other rank and only then switch to recv mode. This only works if the send operations are non-blocking and the passed tensors are immediately copied to some kind of send buffer. Instead, every send must be matched with a recv on the other side, because from the API perspective they may block. E.g. imagine a 1GB tensor being sent to every other rank. It can only go through if there is a recv on the other side, or it will deadlock. This change reflects this in the send/recv unit tests. Pull Request resolved: pytorch#11275 Differential Revision: D9658197 Pulled By: pietern fbshipit-source-id: fb6a3fc03b42343a9dfeed0def30d94914e76974
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
The existing tests had every rank run send to every other rank and only
then switch to recv mode. This only works if the send operations are
non-blocking and the passed tensors are immediately copied to some kind
of send buffer. Instead, every send must be matched with a recv on the
other side, because from the API perspective they may block.
E.g. imagine a 1GB tensor being sent to every other rank. It can only go
through if there is a recv on the other side, or it will deadlock.
This change reflects this in the send/recv unit tests.