Graphics Processing Units Molecular Dynamics
Alternatives To Gpumd
Project NameStarsDownloadsRepos Using ThisPackages Using ThisMost Recent CommitTotal ReleasesLatest ReleaseOpen IssuesLicenseLanguage
Schnetpack638127 days ago4October 12, 20216otherPython
SchNetPack - Deep Neural Networks for Atomistic Systems
a month ago2June 21, 202215mitPython
NequIP is a code for building E(3)-equivariant interatomic potentials
Torchmd39112 months ago2October 22, 20206mitPython
End-To-End Molecular Dynamics (MD) Engine using PyTorch
14 hours ago13gpl-3.0Cuda
Graphics Processing Units Molecular Dynamics
5 years ago1November 08, 201716gpl-3.0Python
Tensorflow + Molecules = TensorMol
5 months agomitPython
Allegro is an open-source code for building highly scalable and accurate equivariant deep learning interatomic potentials
a month ago10mitPython
sGDML - Reference implementation of the Symmetric Gradient Domain Machine Learning model
2 months ago1March 15, 202224otherC++
High-performance operations for neural network potentials
2 years ago1mitJupyter Notebook
Unsupervised learning of atomic scale dynamics from molecular dynamics.
3 days ago20apache-2.0Python
UF3: a python library for generating ultra-fast interatomic potentials
Alternatives To Gpumd
Select To Compare

Alternative Project Comparisons


What is GPUMD?

  • GPUMD stands for Graphics Processing Units Molecular Dynamics. It is a general-purpose molecular dynamics (MD) code fully implemented on graphics processing units (GPUs).
  • Force evaluation for many-body potentials has been significantly accelerated by using GPUs [1], thanks to a set of simple expressions for force, virial stress, and heat current derived in Refs. [2, 3].
  • Apart from being highly efficient, another unique feature of GPUMD is that it has useful utilities to study heat transport [2, 3, 4, 5].
  • It can run MD simulations with the machine-learning based force constant potential (FCP) [6].
  • It can train the NEP machine-learning potential [7, 8, 9] and run MD simulations with it. See this nep-data Gitlab repo for some of the published NEP potentials and the related training/testing data.


  • You need to have a GPU card with compute capability no less than 3.5 and a CUDA toolkit no older than CUDA 9.0.
  • Works for both Linux (with GCC) and Windows (with MSVC) operating systems.

Compile GPUMD

  • Go to the src directory and type make. When the compilation finishes, two executables, gpumd and nep, will be generated in the src directory.


  • Go to the directory of an example and type one of the following commands:
    • path/to/gpumd
    • path/to/nep

Colab tutorial

  • We provide a Colab Tutorial to show the workflow of the construction of a NEP model and its application in large-scale atomistic simulations for PbTe system. This will run entirely on Google's cloud virtual machine. You can also check other offline tutorials in the examples.


Mailing list:

Python interfaces:

Package link
gpyumd AlexGabourie/gpyumd
pynep bigd4/PyNEP
somd initqp/somd


  • Before the first release, GPUMD was developed by Zheyong Fan, with help from Ville Vierimaa (Previously Aalto University) and Mikko Ervasti (Previously Aalto University) and supervision from Ari Harju (Previously Aalto University).
  • Below is the full list of contributors starting from the first release.
Name contact
Zheyong Fan
Alexander J. Gabourie
Ke Xu
Ting Liang
Jiahui Liu
Penghua Ying
Real Name ?
Real Name ?
Yanzhou Wang
Rui Zhao
Eric Lindgren
Junjie Wang
Yong Wang
Zhixin Liang
Paul Erhart
Nan Xu
Shunda Chen
Jiuyang Shi
Nicklas Österbacka


Reference cite for what?
[1] for any work that used GPUMD
[2-3] virial and heat current formulation
[4] in-out decomposition and related spectral decomposition
[5] HNEMD and related spectral decomposition
[6] force constant potential (FCP)
[7-9] neuroevolution potential (NEP)
[10] NEP + ZBL


[1] Zheyong Fan, Wei Chen, Ville Vierimaa, and Ari Harju. Efficient molecular dynamics simulations with many-body potentials on graphics processing units, Computer Physics Communications 218, 10 (2017).

[2] Zheyong Fan, Luiz Felipe C. Pereira, Hui-Qiong Wang, Jin-Cheng Zheng, Davide Donadio, and Ari Harju. Force and heat current formulas for many-body potentials in molecular dynamics simulations with applications to thermal conductivity calculations, Phys. Rev. B 92, 094301, (2015).

[3] Alexander J. Gabourie, Zheyong Fan, Tapio Ala-Nissila, Eric Pop, Spectral Decomposition of Thermal Conductivity: Comparing Velocity Decomposition Methods in Homogeneous Molecular Dynamics Simulations, Phys. Rev. B 103, 205421 (2021).

[4] Zheyong Fan, Luiz Felipe C. Pereira, Petri Hirvonen, Mikko M. Ervasti, Ken R. Elder, Davide Donadio, Tapio Ala-Nissila, and Ari Harju. Thermal conductivity decomposition in two-dimensional materials: Application to graphene, Phys. Rev. B 95, 144309, (2017).

[5] Zheyong Fan, Haikuan Dong, Ari Harju, and Tapio Ala-Nissila, Homogeneous nonequilibrium molecular dynamics method for heat transport and spectral decomposition with many-body potentials, Phys. Rev. B 99, 064308 (2019).

[6] Joakim Brorsson, Arsalan Hashemi, Zheyong Fan, Erik Fransson, Fredrik Eriksson, Tapio Ala-Nissila, Arkady V. Krasheninnikov, Hannu-Pekka Komsa, Paul Erhart, Efficient calculation of the lattice thermal conductivity by atomistic simulations with ab-initio accuracy, Advanced Theory and Simulations 4, 2100217 (2021).

[7] Zheyong Fan, Zezhu Zeng, Cunzhi Zhang, Yanzhou Wang, Keke Song, Haikuan Dong, Yue Chen, and Tapio Ala-Nissila, Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport, Phys. Rev. B. 104, 104309 (2021).

[8] Zheyong Fan, Improving the accuracy of the neuroevolution machine learning potentials for multi-component systems, Journal of Physics: Condensed Matter 34 125902 (2022).

[9] Zheyong Fan, Yanzhou Wang, Penghua Ying, Keke Song, Junjie Wang, Yong Wang, Zezhu Zeng, Ke Xu, Eric Lindgren, J. Magnus Rahm, Alexander J. Gabourie, Jiahui Liu, Haikuan Dong, Jianyang Wu, Yue Chen, Zheng Zhong, Jian Sun, Paul Erhart, Yanjing Su, Tapio Ala-Nissila, GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations, The Journal of Chemical Physics 157, 114801 (2022).

[10] Jiahui Liu, Jesper Byggmästar, Zheyong Fan, Ping Qian, and Yanjing Su, Large-scale machine-learning molecular dynamics simulation of primary radiation damage in tungsten Phys. Rev. B 108, 054312 (2023).

Popular Molecular Dynamics Projects
Popular Machine Learning Projects
Popular Science Categories
Related Searches

Get A Weekly Email With Trending Projects For These Categories
No Spam. Unsubscribe easily at any time.
Machine Learning
Neural Network
Physics Simulation
Molecular Dynamics
High Performance Computing