[Bugfix] fix decode oom with EAGLE

[alphabetc1]

fix #13741

ROOT CAUSE

The original implementation only checked memory usage in the decode stage and ignored EAGLE’s draft and verify stages.

SOLUTION

Calculate the memory requirements for all stages in check_decode_mem().

Motivation

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.
• Work with maintainers to merge your PR. See the PR Merge Process

[gemini-code-assist]

Summary of Changes

Hello @alphabetc1, 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 resolves an Out-Of-Memory (OOM) issue encountered when utilizing the EAGLE speculative decoding algorithm. The root cause was an incomplete memory check that only considered the final decode stage, neglecting the memory footprint of EAGLE's intermediate draft and verification processes. The solution involves enhancing the check_decode_mem() function to accurately calculate and reserve memory for all stages of EAGLE, ensuring stable operation and preventing OOM errors.

Highlights

• Memory Calculation for EAGLE: The memory calculation in the check_decode_mem function has been updated to correctly account for the additional memory required by EAGLE's draft and verify stages, which was previously overlooked.
• Bugfix for OOM Errors: This change directly addresses and fixes Out-Of-Memory (OOM) errors that occurred when using the EAGLE speculative decoding algorithm due to incomplete memory usage checks.

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.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

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. ↩

[gemini-code-assist]

Code Review

This pull request addresses a potential out-of-memory issue with EAGLE speculative decoding by correctly accounting for the memory required during the draft and verify stages. The change is implemented in check_decode_mem by adding the calculated memory needs for these stages to the total token count. The logic is clear, correct, and directly solves the described problem. The implementation is solid, and I have no further suggestions.

[hnyls2002]

This is due to we forget to pass the buf_multiplier in check_decode_mem in retraction. Has been fixed by #14939.