[CUDA] fix bug in syevBatched ffi kernel

[inailuig]

fix #33062

[gemini-code-assist]

Summary of Changes

Hello @inailuig, 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 a critical bug within the CUDA syevBatched kernel's Foreign Function Interface (FFI) implementation. The fix guarantees correct batch processing by ensuring that the calculated batch step is never less than one, thereby preventing potential division-by-zero or incorrect batching logic that could arise with extremely large matrix sizes. This change enhances the robustness and reliability of batched eigenvalue decomposition operations on GPUs.

Highlights

• Bug Fix: Addresses a bug in the syevBatched FFI kernel for CUDA, specifically related to the calculation of batch_step.
• Batch Step Calculation: Ensures that the batch_step value is always at least 1, preventing potential issues when std::numeric_limits<int>::max() / matrix_size evaluates to zero for very large matrix_size.

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 fixes a bug in the syevBatched FFI kernel where batch_step could become zero, potentially leading to an infinite loop. The fix correctly ensures batch_step is at least 1. I've also identified a related critical issue: if n is 0, matrix_size becomes 0, which would cause a division-by-zero error. I've added a comment with a suggestion to fix this.

[gemini-code-assist]

This change correctly handles the case where matrix_size is large, but it doesn't account for the case where n is 0. If n is 0, matrix_size will be 0, leading to a division-by-zero error. Please add a check to prevent this.

Suggested change

	batch_step = std::max(int64_t(1), std::numeric_limits<int>::max() / matrix_size);
	batch_step = (n == 0) ? 1 : std::max(int64_t(1), std::numeric_limits<int>::max() / matrix_size);

[PhilipVinc]

ping @dfm , we'd really hope to get this in before the next jax release...

[hawkinsp]

Thanks for the PR. Sorry, I had missed this! I sent #33233 that includes both this change, but also incorporates Gemini's suggestion (which is a good one) and avoids another possible latent bug. It also adds a test.

Closing, let's use that PR.