ROCm 7.0 update #9757
ROCm 7.0 update #9757
Conversation
- torch 2.9.0a0+git7bcbafe - triton 3.4.0+git02502c86 - aiter 0.1.5.dev149+g85f66157f - Mooncake dcdf1c784b40aa6975a8ed89fe26321b028e40e8
![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 @sogalin, 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 updates the ROCm base image to version 7.0.0 within the Dockerfile.rocm. This update involves refreshing the underlying base images for different GPU architectures (gfx950 and gfx942) and aligning various build parameters and dependency versions, including torch, triton, and aiter, to support the new ROCm environment.
Highlights
- ROCm Base Image Update: The core ROCm base images for both gfx950 and gfx942 architectures have been updated to a newer rocm7-vllm-20250821 version.
- Dependency Version Alignment: The AITER_COMMIT has been bumped to v0.1.5, and the MOONCAKE_COMMIT has been updated to a more recent hash, ensuring compatibility with the new ROCm environment.
- Build Configuration Adjustments: The build process for TRITON (for gfx942) and LLVM (for both architectures) has been disabled, and MOONCAKE build is now enabled for gfx950.
- Mooncake Build Streamlining: The explicit installation of the go language runtime has been removed from the Mooncake build steps, likely indicating it's no longer a direct dependency or is handled by the base image.
- Python Package Pinning: The torchao Python package installation is now explicitly pinned to version 0.9.0.
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 updates the ROCm image to version 7.0, along with its dependencies like torch, triton, and aiter. The Dockerfile is modified to use new base images and adjust build steps accordingly. My review focuses on the correctness of the Dockerfile changes. I found one issue with incorrect syntax that could break future builds.
| # ----------------------- | ||
| # llvm | ||
| RUN if [ "$BUILD_LLVM" = "1" ]; then \ | ||
| ENV HIP_CLANG_PATH="/sgl-workspace/llvm-project/build/bin/" \ |
There was a problem hiding this comment.
The ENV instruction is a Dockerfile instruction and cannot be used inside a RUN command's shell script. This will cause the build to fail if BUILD_LLVM is set to "1". To set an environment variable for the duration of the RUN command, you should use export and terminate the command with a semicolon.
export HIP_CLANG_PATH="/sgl-workspace/llvm-project/build/bin/"; \
- ROCm630 on MI30x - ROCm700 on MI30x - ROCm700 on MI35x
2 participants
Motivation
Update ROCm image to ROCm7.0.0
- torch 2.9.0
- triton 3.4.0
- aiter 0.1.5
Modifications
Accuracy Tests
Benchmarking and Profiling
Checklist