Add a warning log when there is high skew of uneven inputs in DDP training #45238
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
…ining Adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/) [ghstack-poisoned]
… in DDP training" This request came up in feature review for DDP uneven inputs, so this PR adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/) [ghstack-poisoned]
… in DDP training" This request came up in feature review for DDP uneven inputs, so this PR adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/) [ghstack-poisoned]
… in DDP training" This request came up in feature review for DDP uneven inputs, so this PR adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks in #42577, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/) [ghstack-poisoned]
💊 CI failures summary and remediationsAs of commit fb43098 (more details on the Dr. CI page):
Extra GitHub checks: 1 failed
codecov.io: 1 failed
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 6 times. |
mrshenli
approved these changes
Sep 24, 2020
torch/nn/parallel/distributed.py
Outdated
| "other currently active ranks. This level of skew could " | ||
| "lead to performance degradation during training." | ||
| ) | ||
| warned = True |
Contributor
There was a problem hiding this comment.
Please feel free to ignore. This can also be done using warnings.simplefilter("once") IIUC.
Contributor
Author
There was a problem hiding this comment.
Thanks! This is better than toggling the boolean.
… in DDP training" This request came up in feature review for DDP uneven inputs, so this PR adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks in #42577, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/) [ghstack-poisoned]
rohan-varma
added a commit
that referenced
this pull request
Sep 24, 2020
…ining Pull Request resolved: #45238 Adds a warning when there is much higher than expected amount of discrepancy of inputs across different processes when running with uneven inputs. This is because a skew in the thousands can reduce performance a nontrivial amount as shown in benchmarks, and it was proposed to add this warning as a result. Tested by running the tests so the threshold is hit and observing the output. ghstack-source-id: 112773552 Differential Revision: [D23719270](https://our.internmc.facebook.com/intern/diff/D23719270/)
Codecov Report
@@ Coverage Diff @@
## gh/rohan-varma/179/base #45238 +/- ##
===========================================================
- Coverage 68.01% 68.00% -0.01%
===========================================================
Files 393 393
Lines 50847 50854 +7
===========================================================
+ Hits 34583 34584 +1
- Misses 16264 16270 +6
Continue to review full report at Codecov.
|
Contributor
|
This pull request has been merged in e57a081. |
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.
Stack from ghstack:
This request came up in feature review for DDP uneven inputs, so this PR adds a warning when there is much higher than expected amount of
discrepancy of inputs across different processes when running with uneven
inputs. This is because a skew in the thousands can reduce performance a
nontrivial amount as shown in benchmarks in #42577, and it was proposed to add this
warning as a result. Tested by running the tests so the threshold is hit and
observing the output.
Differential Revision: D23719270