Luthaf / rascaline

Computing representations for atomistic machine learning
https://luthaf.fr/rascaline/
BSD 3-Clause "New" or "Revised" License
44 stars 13 forks source link

Parallelize LODE calculation over samples #293

Closed Luthaf closed 5 months ago

Luthaf commented 5 months ago

This is only used when we have fewer systems than available threads.

For many system, this is between 5% slower and 20% faster than the previous code (likely due to the reorganization of the loop over samples). For a single system, this is up to 6x faster with 10 threads for the molecular_crystal benchmark file.

Requested by @andreagrisafi for use in SALTED!


📚 Documentation preview 📚: https://rascaline--293.org.readthedocs.build/en/293/

github-actions[bot] commented 5 months ago

Here is a pre-built version of the code in this pull request: wheels.zip, you can install it locally by unzipping wheels.zip and using pip to install the file matching your system