Research data for "Machine-learned acceleration for molecular dynamics in CASTEP"

This repository contains data and code supporting the following work:

T. K. Stenczel, Z. El-Machachi, G. Liepuoniute, J. D. Morrow, A. P. Bartók, M. I. J. Probert, G. Csányi, V. L. Deringer, in preparation (2023).

The purpose is to allow readers to reproduce the results of the paper, and also to design their own benchmarks, tests, and more. It is not meant as the primary repository for the up-to-date version of the code: for this, please visit https://github.com/libAtoms/QUIP.

Contents

This repository contains the following:

• The code (hybrid-md.py) which is interfaced to CASTEP
• Simulation results for aluminium oxide (original CASTEP output and trajectories in concatenated .xyz format) and the carbon graphitisation runs (original CASTEP and LAMMPS outputs, fitted models, DFT-labelled structural data in exteneded '.xyz' format and relevant input files)