PySAGES (Python Suite for Advanced General Ensemble Simulations) is a Python implementation of SSAGES with support for GPUs.
PySAGES currently supports HOOMD-blue, LAMMPS, ASE, and OpenMM. We plan to add support for other engines that have a python API, including those supported by the original SSAGES.
Before installing PySAGES, you need to install one of the following plugins, depending on your molecular dynamics engine of choice:
- DLPack Plugin for HOOMD-blue
- DLPack Plugin for OpenMM
- DLPack Plugin for LAMMPS
- No plugin needed for ASE
You also need to install JAX, a library for high-performance numerical computing. Follow their installation guide and make sure you have it installed before using PySAGES. Depending on your local setup, you may need to install the CPU version or the CUDA compatible version.
For GPU support, JAX and any backend (HOOMD-blue, LAMMPS, or OpenMM) need to be built or installed with CUDA support.
You can follow our installation tutorial on Google Colab to see how PySAGES, HOOMD-blue,
and OpenMM can be built and installed in such an environment.
PySAGES provides a straightforward interface to set up collective variables and enhanced sampling methods in your MD simulations. See the documentation to learn more.
We provide ready-to-go examples for common methods. Check out the examples subfolder to look at different script and notebook examples. These include pre-set simulations and a tutorial on how to install PySAGES along with the supported MD engines.
We believe in good software engineering and collaboration. We welcome all contributions.
To ease collaboration, we use trunk as a development tool free for
open-source software. This includes version-checked linters that can be run automatically.
We ship a launcher for trunk with this repo: ./trunk, no installation required. For
details about how to use ./trunk fmt to format existing code into this style and
./trunk check to verify a consistent code style, check out the trunk documentation
page.