Simulation of Systems of interacting mean-field Particles with High Efficiency


The SiSyPHE library builds on recent advances in hardware and software for the efficient simulation of large scale interacting particle systems, both on the GPU and on the CPU. The implementation is based on recent libraries originally developed for machine learning purposes to significantly accelerate tensor (array) computations, namely the PyTorch package and the KeOps library. The versatile object-oriented Python interface is well suited to the comparison of new and classical many-particle models, enabling ambitious numerical experiments and leading to novel conjectures. The SiSyPHE library speeds up both traditional Python and low-level implementations by one to three orders of magnitude for systems with up to several millions of particles.

The project is hosted on GitHub, under the permissive MIT license.
PyPi version



Contributions to make SiSyPHE grow are warmly welcome! Examples of possible (and ongoing) developments include the following.

  • The implementation of new models.

  • The implementation of more complex boundary conditions and of models on non-flat manifolds.

  • An improved visualization method (currently only basic visualization functions relying on Matplotlib are implemented).

Contributions can be made by opening an issue on the GitHub repository, via a pull request or by contacting directly the author.



Antoine Diez, Imperial College London


The development of this library would not have been possible without the help of Jean Feydy, his constant support and precious advice. This project was initiated by Pierre Degond and has grown out of many discussions with him.

Table of contents

Tutorials and examples

Indices and tables