bugfix: Fix FLOPS calculation for bench_trtllm_gen_mla.py #1640
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @RayWang96, 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 an inaccuracy in the FLOPS calculation within a specific benchmarking script. By adjusting the formula to use the average sequence length, it ensures that the reported performance metrics are more precise and reflective of typical workloads, thereby enhancing the reliability of benchmark results.
Highlights
- FLOPS Calculation Correction: The script
bench_trtllm_gen_mla.pynow correctly calculates FLOPS by using the average sequence length (avg_seq_len) instead of the maximum sequence length (seq_len) in its formula.
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. 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
-
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. ↩
There was a problem hiding this comment.
Code Review
This pull request corrects the FLOPS calculation in the bench_trtllm_gen_mla.py benchmark script. The change involves calculating the average sequence length of the batch and using it in the FLOPS formula instead of the maximum sequence length. This is a correct change that leads to a more accurate performance metric for the benchmark, given its use of variable sequence lengths. The implementation is sound and effectively resolves the issue.
benchmarks/bench_trtllm_gen_mla.py
Outdated
| ms = np.median(measurements) | ||
| flops = ( | ||
| 2 | ||
| * batch_size |
There was a problem hiding this comment.
Or removing the batch_size here and multiply sum(seq_lens) instead?
There was a problem hiding this comment.
You're right. It is simpler.
yzh119
left a comment
yzh119
left a comment
There was a problem hiding this comment.
LGTM, thanks for the fix!
yzh119
changed the title
2 participants
Fix FLOPS calculation for bench_trtllm_gen_mla.py
Details in #1638