Skip to content

Enable fx tracing for maybe_td_to_kjt #2991

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

jeetkanjani7
Copy link

@jeetkanjani7 jeetkanjani7 commented May 21, 2025

No description provided.

Summary:
This diff applies Sparse Data Distribution (SDD) to the CINT expert model. SDD distributes the feature to the right trainers before running lookup. We make the modules fx tracable to enable SDD which remove s the communication/computation overlap.

Local run logs: https://www.internalfb.com/intern/everpaste/?handle=GPCynx2gWWqpyV4DAFLy0frFtTpPbsIXAAAz&phabricator_paste_number=1811578826

Mast job: fire-linjianma-20250516-1506-3ae5c1eb (peak qps: 1.17)
Baseline:  fire-linjianma-20250426-2216-8530f5d8 (peak qps: 1.11)

Verified from the trace that SDD is applied correctly:
 {F1978480235}

Differential Revision: D74751782
@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label May 21, 2025
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74751782

@jeetkanjani7 jeetkanjani7 changed the title Enable SDD for standalone CINT expert model Enable fx tracing for maybe_td_to_kjt May 21, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants