[DCP] Add option to use PrefixStore to create checkpoint background process#166560
Closed
kevinmtang wants to merge 1 commit intopytorch:mainfrom
Closed
[DCP] Add option to use PrefixStore to create checkpoint background process#166560kevinmtang wants to merge 1 commit intopytorch:mainfrom
kevinmtang wants to merge 1 commit intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/166560
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit 0fdfe61 with merge base 104b868 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
pytorch-bot
bot
added
oncall: distributed
|
@kevinmtang has exported this pull request. If you are a Meta employee, you can view the originating Diff in D84928180. |
kevinmtang
force-pushed
the
export-D84928180
branch
from
68e4a19 to
146388c
Compare
kevinmtang
added a commit
to kevinmtang/pytorch
that referenced
this pull request
Oct 30, 2025
…rocess (pytorch#166560) Summary: DCP checkpoint background process currently determines the port used for pg via get_free_port(). During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call. We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port. This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT. The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1". context: https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/ Test Plan: ``` buck test fbcode//mode/opt fbcode//caffe2/test/distributed/checkpoint:test_async_process_executor -- --run-disabled ``` https://www.internalfb.com/intern/testinfra/testrun/1407375340702140 https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_16-ETT-1348807470?job_attempt=1&version=0&tab=summary&env=PRODUCTION Differential Revision: D84928180
kevinmtang
force-pushed
the
export-D84928180
branch
from
146388c to
81468a3
Compare
kevinmtang
added a commit
to kevinmtang/pytorch
that referenced
this pull request
Oct 30, 2025
…rocess (pytorch#166560) Summary: DCP checkpoint background process currently determines the port used for pg via get_free_port(). During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call. We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port. This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT. The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1". context: https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/ Test Plan: ``` buck test fbcode//mode/opt fbcode//caffe2/test/distributed/checkpoint:test_async_process_executor -- --run-disabled ``` https://www.internalfb.com/intern/testinfra/testrun/1407375340702140 https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_16-ETT-1348807470?job_attempt=1&version=0&tab=summary&env=PRODUCTION https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_256-ETT-28070b842e?version=0&env=PRODUCTION Differential Revision: D84928180
kevinmtang
force-pushed
the
export-D84928180
branch
from
81468a3 to
6357f5c
Compare
kevinmtang
added a commit
to kevinmtang/pytorch
that referenced
this pull request
Oct 30, 2025
…rocess (pytorch#166560) Summary: DCP checkpoint background process currently determines the port used for pg via get_free_port(). During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call. We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port. This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT. The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1". context: https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/ Test Plan: ``` buck test fbcode//mode/opt fbcode//caffe2/test/distributed/checkpoint:test_async_process_executor -- --run-disabled ``` https://www.internalfb.com/intern/testinfra/testrun/1407375340702140 https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_16-ETT-1348807470?job_attempt=1&version=0&tab=summary&env=PRODUCTION https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_256-ETT-28070b842e?version=0&env=PRODUCTION Differential Revision: D84928180
meetv18
approved these changes
Nov 3, 2025
Contributor
meetv18
left a comment
meetv18
left a comment
There was a problem hiding this comment.
Thanks for adding support for this and improving reliability of process based async cp :)
…rocess (pytorch#166560) Summary: DCP checkpoint background process currently determines the port used for pg via get_free_port(). During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call. We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port. This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT. The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1". context: https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/ Test Plan: ``` buck test fbcode//mode/opt fbcode//caffe2/test/distributed/checkpoint:test_async_process_executor -- --run-disabled ``` https://www.internalfb.com/intern/testinfra/testrun/1407375340702140 https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_16-ETT-1348807470?job_attempt=1&version=0&tab=summary&env=PRODUCTION https://www.internalfb.com/mlhub/pipelines/runs/mast/aps-omnifm_0928_6kx_256-ETT-28070b842e?version=0&env=PRODUCTION Reviewed By: meetv18 Differential Revision: D84928180
kevinmtang
force-pushed
the
export-D84928180
branch
from
6357f5c to
0fdfe61
Compare
Contributor
|
@pytorchbot merge (Initiating merge automatically since Phabricator Diff has merged) |
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 |
pytorch-bot bot
pushed a commit
that referenced
this pull request
Nov 4, 2025
…rocess (#166560) Summary: DCP checkpoint background process currently determines the port used for pg via get_free_port(). During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call. We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port. This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT. The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1". context: https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/ Differential Revision: D84928180 Pull Request resolved: #166560 Approved by: https://github.com/meetv18
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
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.
Summary:
DCP checkpoint background process currently determines the port used for pg via get_free_port().
During checkpoint background process initialization, gloo pg init occasionally times out on the first call but succeeds in a subsequent call.
We hypothesized that the timeouts are related to the port being used, and the solution would be to create the pg with PrefixStore and reuse the master port.
This diff adds the option for checkpoint background process to use PrefixStore with MASTER_ADDR + MASTER_PORT.
The default behavior is unchanged. Enabling the new PrefixStore behavior requires setting "DCP_USE_PREFIX_STORE" env var to "1".
context:
https://fb.workplace.com/groups/319878845696681/permalink/1516883985996155/
Differential Revision: D84928180
cc @H-Huang @awgu @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @pragupta @msaroufim @dcci