Add option for mutex timeout in distributed optimizer backward hook #9087
+27
−1
Conversation
This file contains 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
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
minitu
force-pushed
the
distopt-mutex-timeout
branch
from
dd78d22 to
f36140a
Compare
minitu
changed the title
8 tasks
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
minitu
force-pushed
the
distopt-mutex-timeout
branch
from
f36140a to
972d1b6
Compare
added 2 commits
May 1, 2024 20:33
This reverts commit 972d1b6. Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
|
@ShriyaPalsamudram Tested on Eos single-node, the updated code spits out a RuntimeError when the lock is not obtained within the given timeout period. |
ShriyaPalsamudram
previously approved these changes
May 2, 2024
|
How were you able to create a timeout situation in one node? Did you set the timer to very very small value? |
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
ericharper
reviewed
May 2, 2024
| self._lock.release() | ||
| else: | ||
| # Failed to acquire lock before timeout | ||
| print(f'MegatronDistributedFusedAdam: Failed to acquire lock within {lock_timeout} seconds.') |
There was a problem hiding this comment.
this should use logging
There was a problem hiding this comment.
The issue is that we want the print to happen even with NeMo logging disabled, for hang debugging purposes. @ShriyaPalsamudram
github-actions bot
pushed a commit
that referenced
this pull request
May 2, 2024
…9087) * Tim: Add option for timeout in distopt callback mutex Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> * Replace parent's _lock Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> * Revert "Replace parent's _lock" This reverts commit 972d1b6. Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> * Raise RuntimeError when timeout Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> * Change RuntimeError to print Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com>
michal2409
added a commit
that referenced
this pull request
Jun 12, 2024
…9087) (#9091) * Tim: Add option for timeout in distopt callback mutex * Replace parent's _lock * Revert "Replace parent's _lock" This reverts commit 972d1b6. * Raise RuntimeError when timeout * Change RuntimeError to print --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <minitu77@gmail.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Michal Futrega <mfutrega@nvidia.com> Co-authored-by: Pablo Garay <palenq@gmail.com>
galv
pushed a commit
to galv/NeMo
that referenced
this pull request
Jun 13, 2024
…VIDIA#9087) (NVIDIA#9091) * Tim: Add option for timeout in distopt callback mutex * Replace parent's _lock * Revert "Replace parent's _lock" This reverts commit 972d1b6. * Raise RuntimeError when timeout * Change RuntimeError to print --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <minitu77@gmail.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Michal Futrega <mfutrega@nvidia.com> Co-authored-by: Pablo Garay <palenq@gmail.com>
JesusPaz
pushed a commit
to JesusPaz/NeMo
that referenced
this pull request
Jun 18, 2024
…VIDIA#9087) (NVIDIA#9091) * Tim: Add option for timeout in distopt callback mutex * Replace parent's _lock * Revert "Replace parent's _lock" This reverts commit 972d1b6. * Raise RuntimeError when timeout * Change RuntimeError to print --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <minitu77@gmail.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Michal Futrega <mfutrega@nvidia.com> Co-authored-by: Pablo Garay <palenq@gmail.com>
rohitrango
pushed a commit
to rohitrango/NeMo
that referenced
this pull request
Jun 25, 2024
…VIDIA#9087) (NVIDIA#9091) * Tim: Add option for timeout in distopt callback mutex * Replace parent's _lock * Revert "Replace parent's _lock" This reverts commit 972d1b6. * Raise RuntimeError when timeout * Change RuntimeError to print --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <minitu77@gmail.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Michal Futrega <mfutrega@nvidia.com> Co-authored-by: Pablo Garay <palenq@gmail.com>
XuesongYang
pushed a commit
to paarthneekhara/NeMo
that referenced
this pull request
Jan 18, 2025
…VIDIA#9087) (NVIDIA#9091) * Tim: Add option for timeout in distopt callback mutex * Replace parent's _lock * Revert "Replace parent's _lock" This reverts commit 972d1b6. * Raise RuntimeError when timeout * Change RuntimeError to print --------- Signed-off-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Jaemin Choi <minitu77@gmail.com> Co-authored-by: Jaemin Choi <jaeminc@nvidia.com> Co-authored-by: Michal Futrega <mfutrega@nvidia.com> Co-authored-by: Pablo Garay <palenq@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
What does this PR do ?
Modified version of #9084, to debug a hang at
NeMo/nemo/core/optim/distributed_adam.py
Line 131 in f658b6f
Collection: NLP
Changelog
Usage
Run GPT, e.g. with the config at https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/language_modeling/conf/megatron_gpt_config.yaml.
Enable the distributed optimizer with
model.optim.name=distributed_fused_adamand set the timeout withmodel.optim.lock_timeout=<seconds>.Jenkins CI
The Jenkins CI system has been replaced by GitHub Actions self-hosted runners.
There's no need to comment
jenkinson the PR to trigger Jenkins CI.The GitHub Actions CI will run automatically when the PR is opened.
To run CI on an untrusted fork, a NeMo user with write access must click "Approve and run".
Before your PR is "Ready for review"
Pre checks:
PR Type:
If you haven't finished some of the above items you can still open "Draft" PR.
Who can review?
Anyone in the community is free to review the PR once the checks have passed.
Contributor guidelines contains specific people who can review PRs to various areas.
Additional Information