Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

rocm_root automatically set to /opt/rocm/ #4149

Closed
sigbjobo opened this issue Sep 20, 2024 · 0 comments · Fixed by #4150
Closed

rocm_root automatically set to /opt/rocm/ #4149

sigbjobo opened this issue Sep 20, 2024 · 0 comments · Fixed by #4150

Comments

@sigbjobo
Copy link
Contributor

Summary

Setting ROCM_ROOT via hipconfig --rocmpath can cause compatibility issues if /opt/rocm/ changes due to updates.

DeePMD-kit Version

devel

Backend and its version

tensorflow14, ROCM

Python Version, CUDA Version, GCC Version, LAMMPS Version, etc

No response

Details

When the ROCM_ROOT environment variable is set, its value is automatically inferred by executing hipconfig --rocmpath. This command typically resolves the path to /opt/rocm/. However, this behavior can lead to compatibility issues if the contents of /opt/rocm/ are altered due to updates or changes on the server, such as discussed (#2523 (reply in thread)).

@njzjz njzjz linked a pull request Sep 20, 2024 that will close this issue
@njzjz njzjz added enhancement and removed wontfix labels Sep 20, 2024
github-merge-queue bot pushed a commit that referenced this issue Sep 20, 2024
This small commit closes issue #4149 by checking if `rocm_root` is not
set, and then inferring it from `rocm_path`, if it exists.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

- **New Features**
- Enhanced flexibility in retrieving the ROCM root directory by checking
both `ROCM_ROOT` and `ROCM_PATH` environment variables.
- Updated installation documentation to clarify the fallback mechanism
for locating the ROCM toolkit, ensuring users have clear guidance on
configuration options.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: Sigbjørn Løland Bore <31211370+sigbjobo@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants