float8: snapshot FSDP precomputed scale#4555
Open
nascheme wants to merge 1 commit into
Open
Conversation
Read the precomputed scale once in fsdp_pre_all_gather so concurrent updates between FSDP iterations cannot change which value is used after the branch is selected. Assisted-by: Claude Opus 4.7
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4555
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
fsdp_pre_all_gatheronWeightWithDynamicFloat8CastTensorchecksself._precomputed_scale is not Noneand then readsself._precomputed_scaleagain to use it.precompute_float8_dynamic_scale_for_fsdpcan reassign that attribute on the same shared instance between the check and the use, so under free-threaded Python the value used may differ from the one checked.Read
self._precomputed_scaleinto a local once and use that for both the check and the call.