fix:buffer control

[Yael-X]

The size of the lowlatency buffer is calculated using the following formula: num_max_dispatch_tokens_per_rank * num_ranks * min(topk, local_expert_num)

The output of the modified bs is correct.

SGLANG_DEEPEP_NUM_MAX_DISPATCH_TOKENS_PER_RANK=24，ep=16, topk=6, local_expert_num = 256 / ep = 16

[gemini-code-assist]

Summary of Changes

Hello @Yael-X, 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 addresses a critical fix in the low-latency buffer control mechanism. The primary goal is to ensure that the buffer's maximum token capacity is accurately calculated, preventing potential issues related to incorrect memory allocation or data handling. By refining the formula for num_max_tokens, the change improves the robustness and efficiency of the buffer management system.

Highlights

• Buffer Size Calculation Fix: Corrected the calculation of num_max_tokens within the low_latency_dispatch function. The formula now correctly uses std::min(num_topk, num_local_experts) to determine the maximum number of tokens, aligning with the intended buffer size logic.
• Type Specification: Explicitly defined the type of num_local_experts as int32_t instead of auto for clarity and consistency.

🧠 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.

Changelog

• csrc/deepep/deep_ep.cpp
  • Modified the low_latency_dispatch function to update the num_max_tokens calculation, incorporating std::min(num_topk, num_local_experts).
  • Changed the declaration of num_local_experts from auto to int32_t.

Activity

• No specific activity (comments, reviews, or progress updates) has been recorded for this pull request yet.

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 corrects the buffer size calculation for the low-latency mode by taking the minimum of topk and local_expert_num, which seems to be the intended logic. However, the change of num_local_experts's type from auto to int32_t introduces a potential integer overflow vulnerability, which could lead to incorrect buffer allocation. I've added a comment with a suggestion to mitigate this risk.