OpenMM is both a library and a stand-alone application which provides tools for modern molecular modeling simulation. As a library it can be hooked into any code, allowing that code to do molecular modeling with minimal extra coding. Moreover, OpenMM has a strong emphasis on hardware acceleration, thus providing not just a consistent API, but much greater performance than what one could get from just about any other code available. Different parts of the code are distributed under different licenses (read license).

Getting Started

  • New to MD? - information for those starting out with molecular dynamic simulations

  • Sample Scripts and Tutorials - these can serve as templates and examples to help you achieve what you want to do with OpenMM

  • Troubleshooting Guide for OpenMM - commonly asked questions and troubleshooting tips for installing and using OpenMM

  • Hardware Recommendations for OpenMM - recommended hardware set-ups for maximizing OpenMM performance


See benchmarks showing how OpenMM performs when simulating different types of systems.

Code Repository

There are many additional files and software that can be used with OpenMM to enhance its capabilities. These can serve as templates and examples to help you achieve what you want to do with OpenMM. Included in the repository are:

  • Tutorials
  • Python scripts for running simulations and testing out new algorithms within OpenMM
  • Plug-in modules that extend OpenMM's capabilities
  • Force field definitions provided through XML files
  • Applications built on top of OpenMM

A list of these are available through the Code Repository.


  • 5.1 (May 8, 2013): Significant performance improvements; GROMACS TOP file reader; Support for Intel's OpenCL; Integrator, State, and Force objects now serializable
  • 5.0.1 (February 27, 2013): bug fixes
  • 5.0 (January 31, 2013): a major update, including a completely new CUDA-based platform with significantly improved performance on recent NVIDIA GPUs, support for double precision, checkpointing, and a CustomCompoundBondForce class

  • 4.1.1 (July 6, 2012): bug fixes
  • 4.1 (May 8, 2012): new features, including model building tools, a CustomIntegrator class, virtual sites, and more water models

  • 4.0 (Jan. 12, 2012): introduces a new application layer, allowing users to easily run simulations out of the box; also includes major performance improvements and support for Ring Polymer Molecular Dynamics
  • 3.1.1 (Aug. 11, 2011): bug fixes
  • 3.1 (Aug. 1, 2011): performance improvements, especially to the AMOEBA plugin; support for parallelizing computations across multiple GPUs; support of direct polarization model by AMOEBA plugin; GB/VI force fully supported
  • 3.0 (Mar. 30, 2011): supports AMOEBA force field; provides an energy minimizer; CMAP torsions; improved performance, especially for running on CPUs; Python API wrappers
  • 2.0 (June 24, 2010): supports pressure coupling; provides custom Hbond forces; major performance improvements; works on ATI GPUs
  • 1.1.1 (Mar. 4, 2010): bug fixes
  • 1.1 (Feb. 12, 2010): provides custom torsion forces; bug fixes and other improvements
  • 1.0 (Jan. 20, 2010): provides new custom forces, including bonds, angles, external, and GB; improved OpenCL support
  • 1.0 Beta (Oct. 30, 2009): supports Particle Mesh Ewald (PME); custom nonbonded interactions, including algebraic and tabulated forces; preliminary OpenCL support
  • Preview Release 4 (Aug. 20, 2009): supports Ewald summation; GPU accelerated energy calculations; C and Fortran API wrappers; a faster constraint algorithm; and many minor enhancements
  • Preview Release 3 (May 19, 2009): provides support for explicit solvent on NVIDIA GPUs; includes periodic boundary conditions, cutoffs on non-bonded interactions, and new constraint algorithms.
  • Preview Release 2 (Jan. 26, 2009): provides support for accelerating molecular modeling simulations on ATI and NVIDIA graphics processor units (GPUs)

Future releases

FrontPage (last edited 2014-09-02 19:14:17 by joyku)