Event-Driven Molecular Dynamics in Parallel
S. Miller, S. Luding
Abstract
Although event-driven algorithms have been shown to be far more efficient than time-driven methods such as conventional molecular dynamics, they have not become as popular. The main obstacle seems to be the difficulty of parallelizing event-driven molecular dynamics. Several basic ideas have been discussed in recent years, but to our knowledge no complete implementation has been published yet. In this paper we present a parallel event-driven algorithm including dynamic load-balancing, which can be easily implemented on any computer architecture. To simplify matters our explanations refer to a basic multi-particle system of hard spheres, but can be extended easily to a wide variety of possible models.
Create a lesson
Related papers
How durable are high-performance racing shoes?
Jeremy A. McCulloch, Ellen Kuhl
Correlation-Free Transition Path Sampling through Shooting Point Generation Guided by Committor Learning
Maximilian Negedly, Sebastian Falkner, Alessandro Coretti et al.
Exergy-Anergy Representation of Turbomachine Performance Characteristics
Tihomir Varchev, Yiwen Yuan, Tobias Schateikis et al.
Mollified-sharp decomposition: a probabilistic regularization of parametric POD for shock-bearing flows
Oliver T. Schmidt
Load balancing for adaptive-precision interatomic potentials in materials science
David Immel, Godehard Sutmann
Braided endovascular implants for intracranial aneurysms: mechanics, hemodynamics, and clinical translation
Ratnadeep Pramanik, Duygu Dengiz, Mariya S. Pravdivtseva et al.