Fix several bugs and update v0.5 release notes

[MC952-arch]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @MC952-arch, 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 primarily focuses on updating the project's documentation to reflect the latest v0.5 release features, including new hardware support and performance enhancements. It also addresses several minor bugs related to network resource management and standardizes code formatting and argument parsing within the testing framework. The changes aim to improve clarity, maintainability, and correctness across the codebase.

Highlights

• Release Notes Update: The README.md has been updated to include the release notes for v0.5, detailing new features such as AMD support, a unified network backend (flagcxNetAdaptor), zero-copy RDMA, automatic tuning via flagcxTuner, and integrated PyTorch API tests.
• AMD Support Documentation: Documentation for building the library has been expanded to explicitly include USE_AMD=1 as a valid option for platform-specific compilation.
• Network Resource Deallocation Logic: The memory deallocation logic in flagcxSendProxyFree and flagcxRecvProxyFree within flagcx/core/net.cc has been refined to conditionally free memory based on the network adaptor type (SOCKET or IBRC), ensuring correct resource management for different backends. Additionally, FLAGCXCHECK calls were removed for non-critical deallocation steps.
• Code Style and Formatting Improvements: Numerous minor code style and formatting adjustments have been applied across flagcx_tuner.cc and flagcx_tuner.h, including consistent pointer declaration spacing, static_assert formatting, and improved readability for comments and log messages.
• Test Tool Argument Renaming: The split_mask argument in the performance test tool (test/perf/tools.cc) has been updated, changing its short option from -m to -s for clarity and consistency.

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 introduces several bug fixes and updates the release notes for v0.5. The changes include adding AMD support, unifying network backends, and enabling zero-copy RDMA in the release notes. The build and test documentation has been updated accordingly.

My review focuses on the correctness and potential issues in the code. I've found a couple of minor issues in the README.md file, such as a placeholder date and a typo. More importantly, I've identified a potential risk in flagcx/core/net.cc where error checks have been removed, which could lead to silent failures. I've provided suggestions to restore these checks.

Other changes in flagcx/core/flagcx_tuner.cc, flagcx/core/flagcx_tuner.h, flagcx/flagcx.cc, and test/perf/tools.cc are mostly related to code formatting and bug fixes that seem correct and improve the codebase.

[gemini-code-assist]

The error checks (FLAGCXCHECK) for deregMr and closeSend have been removed. If these functions can fail, their return values should be checked to avoid silent failures and potential resource leaks or other issues. Please consider restoring the error checks.

  FLAGCXCHECK(resources->netAdaptor->deregMr(resources->netSendComm,
                                             resources->mhandles[0]));
  FLAGCXCHECK(resources->netAdaptor->closeSend(resources->netSendComm));

[gemini-code-assist]

The error checks (FLAGCXCHECK) for deregMr, closeRecv, and closeListen have been removed. If these functions can fail, their return values should be checked to avoid silent failures and potential resource leaks or other issues. Please consider restoring the error checks.

  FLAGCXCHECK(resources->netAdaptor->deregMr(resources->netRecvComm,
                                             resources->mhandles[0]));
  FLAGCXCHECK(resources->netAdaptor->closeRecv(resources->netRecvComm));
  FLAGCXCHECK(resources->netAdaptor->closeListen(resources->netListenComm));

[gemini-code-assist]

The month in the release date for v0.5 is 00, which appears to be a placeholder. Please update it to the correct month.

Suggested change

	- **[2025/09]** Released [v0.5](https://github.com/FlagOpen/FlagCX/tree/release/v0.5):
	- **[2025/01]** Released [v0.5](https://github.com/FlagOpen/FlagCX/tree/release/v0.5):

[gemini-code-assist]

There is a typo in "experimently". It should be "experimentally".

Suggested change

	- Introduces flagcxNetAdaptor to unify network backends, currently supporting SOCKET, IBRC, UCX and IBUC (experimently).
	- Introduces flagcxNetAdaptor to unify network backends, currently supporting SOCKET, IBRC, UCX and IBUC (experimentally).

[mikethegoblin]

LGTM

[yzhang35]

LGTM

[chencjcj]

LGTM