A new separate homepage of i-PI is now available containing more information, documentation and resources.
Gallery of GLE pictures

GLE4MD aims to be a collection of tools to simplify the use of colored-noise, generalized Langevin equation thermostats to enhance your molecular dynamics simulations. A colored-noise thermostat can be used to enhance canonical sampling, in particular in difficult cases such as path integral dynamics, to perform constant-temperature MD in a Car-Parrinello-like dynamics, to control resonances in multiple time step dynamics, to include nuclear quantum effects at a fraction of the cost of a path-integrals simulation, to compute the particle momentum distribution, to selectively excite a few normal modes, and a number of other applications which have not been invented yet.

On these pages you will find a collection of ready-to-use GLE parameters, and a handy online tool to prepare the GLE section of the input for different programs, tools to analyze and fit GLE matrices, sample code to include colored-noise thermostats in your MD code, tutorials and documentation to explain in detail how to use colored noise in your simulation. The easiest way to perform simulations using GLE thermostats with your favourite atomistic code uses i-PI, a Python interface to perform classical or path integral molecular dynamics with ab initio (and empirical) forces, that implements all of the colored-noise machinery.

Acknowledgements

The Royal Society and the Swiss National Science Foundation are gratefully acknowledged for financial support during the initial development of the GLE framework described on these pages. Continued theoretica developments and the i-PI software implementation were done thanks to a REA Marie-Curie fellowship, and a Junior Research Fellowship from Merton College.

Help with the implementation of GLE thermostats and i-PI in different external codes is acknowledged in the code section.

Quick links

Source code as a git repository

Homepage of the Laboratory of Computational Science and Modelling