[c10d] Enable CudaEventCache by default and add multi device support #140975
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
[ghstack-poisoned]
pytorch-bot
bot
added
oncall: distributed
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/140975
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit 88fa03c with merge base d0fd42e ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
fduwjj
requested review from
c-p-i-o,
eqy,
kwen2501,
shuqiangzhang,
wconstab and
xw285cornell
and removed request for
c-p-i-o,
shuqiangzhang and
wconstab
eqy
approved these changes
Nov 18, 2024
eqy
reviewed
Nov 18, 2024
Collaborator
eqy
left a comment
eqy
left a comment
There was a problem hiding this comment.
Failures look real, perhaps there's an issue with the cache being shared across streams/devices?
Contributor
Author
|
@eqy I cannot repro it locally... |
kwen2501
approved these changes
Nov 19, 2024
Collaborator
kwen2501
left a comment
kwen2501
left a comment
There was a problem hiding this comment.
Approving.
As discussed, a fix for the CI failure would be to create one cache per device index seen, as DDP still supports multi-device modules. (Though we may start deprecating it)
We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
fduwjj
changed the title
kwen2501
reviewed
Nov 21, 2024
…ce support" We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
fduwjj
commented
Nov 21, 2024
fduwjj
commented
Nov 21, 2024
…ce support" We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
…ce support" We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
…ce support" We added `CudaEventCache` in #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files. cc H-Huang awgu kwen2501 wanchaol fegin wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
Contributor
Author
|
@pytorchbot rebase |
Collaborator
|
@pytorchbot started a rebase job onto refs/remotes/origin/viable/strict. Check the current status here |
Collaborator
|
Successfully rebased |
Contributor
Author
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
pobin6
pushed a commit
to pobin6/pytorch
that referenced
this pull request
Dec 5, 2024
…ytorch#140975) We added `CudaEventCache` in pytorch#133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating. Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files. Pull Request resolved: pytorch#140975 Approved by: https://github.com/eqy, https://github.com/kwen2501
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.
Stack from ghstack (oldest at bottom):
We added
CudaEventCachein #133727 and this is a feature which tries to reuse CudaEvent so that we don't call destroy of CudaEvent which causes hang in the past. We had a bunch of tests and testing on TorchTitan and internal workload already. So far no errors or crash are found at the moment so we decide to roll out to all OSS users. For internal workload, this PR would not affect it because of some internal gating.Also we observed some multi-device use cases in OSS, so that we want to bring back multi-device support originally proposed in https://github.com/pytorch/pytorch/pull/122732/files.
cc @H-Huang @awgu @kwen2501 @wanchaol @fegin @wz337 @wconstab @d4l3k @c-p-i-o