Skip to content

[megatron]delete cpu sync#8447

Open
yangbofun wants to merge 1 commit intomodelscope:mainfrom
xlm-research:yb/opti/reduce_sync_cpu_in_rope
Open

[megatron]delete cpu sync#8447
yangbofun wants to merge 1 commit intomodelscope:mainfrom
xlm-research:yb/opti/reduce_sync_cpu_in_rope

Conversation

@yangbofun
Copy link
Copy Markdown

PR type

  • Bug Fix
  • New Feature
  • Document Updates
  • More Models or Datasets Support

PR information

  1. Delete the CPU sync when rope, which speeds up 1.5%.

before:
image

after:
image

Experiment results

Paste your experiment result here(if needed).

@gemini-code-assist
Copy link
Copy Markdown
Contributor

Summary of Changes

Hello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request aims to enhance the performance of the Rotary Positional Embedding (RoPE) application by eliminating an unnecessary CPU synchronization step. The change streamlines the logic for detecting batched RoPE usage, resulting in a reported speedup of 1.5%.

Highlights

  • Performance Optimization: Removed CPU synchronization logic related to Rotary Positional Embedding (RoPE) calculation, leading to a reported 1.5% speedup.
  • RoPE Batched Calculation: Simplified the determination of whether to use batched RoPE by directly comparing freqs.shape[0] with t.shape[0], eliminating reliance on context parallel group size and cu_seqlens.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature Command Description
Code Review /gemini review Performs a code review for the current pull request in its current state.
Pull Request Summary /gemini summary Provides a summary of the current pull request in its current state.
Comment @gemini-code-assist Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help /gemini help Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

  1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution.

@yangbofun yangbofun changed the title delete cpu sync [megatron]delete cpu sync Mar 26, 2026
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request simplifies the logic for determining use_batched_rope within the _apply_rotary_pos_emb_thd function in swift/megatron/init.py. The previous calculation involving cp_group, cp_size, and cu_seqlens_for_batched has been removed, and use_batched_rope is now directly determined by comparing freqs.shape[0] with t.shape[0]. There are no review comments to provide feedback on.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant