Click for detailed status
Conda
Conda is a powerful package and environment manager, widely used in the Python and R communities. It simplifies installing, running, and updating packages and their dependencies, making reproducible research and software management easier.
Conda on Euler¶
On Euler there's no centrally installed version of Conda. Users can install Conda distributions themselfs. See: https://docs.conda.io/projects/conda/en/stable/index.html
Euler's Lustre file system is not well-suited for Conda (see: Article), as Conda creates many small files. This can lead to performance issues and file quota problems.
Recommended storage locations:
- Home directory: Suitable for small environments, but note the 500,000 file limit.
- Scratch directory: Good for temporary work; be aware of the automatic cleanup policy.
- Project directory: Only available for shareholders.
Avoid:
- Work directory: Not recommended for Conda environments, as it is a Lustre file system.
Containers¶
To bypass file number limitations, consider using an Apptainer container with Conda installed inside. This approach encapsulates all files within a single container image, avoiding issues with file quotas in the Home directory.