Introduction

GPUMD stands for Graphics Processing Units Molecular Dynamics. It is a general-purpose molecular dynamics (MD) package fully implemented on graphics processing units (GPU). In addition to several empirical interatomic potentials, it supports neuroevolution potential (NEP) models. GPUMD also allows one to construct the latter type of models using the nep executable.

GPUMD is also highly efficient for conducting MD simulations with many-body potentials such as the Tersoff potential and is particularly good for heat transport applications. It is written in CUDA C++ and requires a CUDA-enabled Nvidia GPU of compute capability no less than 3.5.

Python interfaces

There are several packages that provide Python interfaces to GPUMD functionality, including calorine, gpyumd, and pyNEP.

Github Discussions

Discussion groups

There are Chinese discussion groups based on the following QQ numbers: * 887975816 (close to be full) * 778083814 (available)