Molecular Dynamics made accessible

Titus A. Beu, Professor, Dr.

Molecular Dynamics (MD) has matured into one of the most powerful and reliable computational tools in present-day research, not only having the ability to explain in detail known behavior, but also to anticipate uncharted phenomenology. In spite of its wide-spread use, due the growing complexity of the models and methodologies amassed in state-of-the-art MD simulators, amplified by complications associated with preparation and post-analysis of the runs, employing and, moreover, developing MD software becomes increasingly intricate and risks to remain the apanage of “initiates”.

I put forward a concept and a software repository for introducing MD at graduate/post-graduate level, based on unifying data structures and programming techniques in Python and C/C++. The main aim is to enable acquisition of specific coding abilities in parallel with operation skills of an established MD code.

Essentially, Python scripting is used for building the initial configurations and post-analyzing the MD trajectories. The developed MD modules rely on published numerical and graphical libraries Introduction to Numerical Programming. On the other hand, C/C++ is used to build a suite of simple MD simulators mimicking the functionalities of an established MD code, in a sequence of gradually growing complexity, along with the introduction of new theoretical elements.

Fundamental aspects such as molecular models, force fields, time propagation, thermodynamic reservoirs, electrostatics in periodic systems, and acceleration techniques are given particular emphasis. The results of the simulation test runs are consistently compared with output from established MD codes to clarify the relevance of the various modeling aspects.