[Feature] SM level profiler

[Edenzzzz]

📌 Description

Simply add smid into the profiler tag. Now it looks like this

Also fixed profiler install in pyproject.toml

cc @happierpig @yzh119

🔍 Related Issues

🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull request, please make sure the following items are complete.

✅ Pre-commit Checks

• I have installed pre-commit by running pip install pre-commit (or used your preferred method).
• I have installed the hooks with pre-commit install.
• I have run the hooks manually with pre-commit run --all-files and fixed any reported issues.

If you are unsure about how to set up pre-commit, see the pre-commit documentation.

🧪 Tests

• Tests have been added or updated as needed.
• All tests are passing (unittest, etc.).

Reviewer Notes

[gemini-code-assist]

Summary of Changes

Hello @Edenzzzz, 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 introduces a significant upgrade to the existing profiling system by integrating Streaming Multiprocessor (SM) IDs into the collected trace data. This change provides a more detailed, hardware-aware view of execution, which is crucial for optimizing performance. Alongside this core feature, I've also addressed a packaging configuration issue to ensure the profiler module is correctly installed.

Highlights

• Profiler Enhancement: I've enhanced the profiler by incorporating the Streaming Multiprocessor (SM) ID into the profiling tags. This allows for more granular performance analysis, enabling a deeper understanding of execution at the hardware level.
• Data Interpretation and Visualization: The decode_tag function and the export_to_perfetto_trace logic have been updated to correctly interpret and visualize the new SM-level profiling data. This includes dynamically creating Perfetto process and thread groups, now organized by SM ID and block index for clearer trace views.
• Build System Fix: I've corrected the pyproject.toml configuration to ensure the flashinfer.profiler module is properly included during package installation, resolving a potential issue with the profiler's availability.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 or fill out our survey to provide feedback.

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 SM-level profiling capabilities by including the SM ID in the profiler tag. The changes involve modifications in both the C++ header and the Python code. A critical issue was identified in the Python profiler data processing logic where the decoding of the tag was not correctly updated, potentially leading to incorrect profiling results. Additionally, suggestions were made to improve type correctness in the C++ header and to add a data type check in the Python code.

[yzh119]

Great work, I encourage renaming it to threadblock_id or cta_id instead as there are cases that there are more than one cta in the same sm.
Ignore my comments, you were using ptx instruction to fetch sm_id, which looks correct to me!

[yzh119]

LGTM, thank you!