Efficient molecular dynamics simulations with many-body potentials on graphics processing units

Abstract

Graphics processing units have been extensively used to accelerate classical molecular dynamics simulations. However, there is much less progress on the acceleration of force evaluations for many-body potentials compared to pairwise ones. In the conventional force evaluation algorithm for many-body potentials, the force, virial stress, and heat current for a given atom are accumulated within different loops, which could result in write conflict between different threads in a CUDA kernel. In this work, we provide a new force evaluation algorithm, which is based on an explicit pairwise force expression for many-body potentials derived recently [Phys. Rev. B 92 (2015) 094301]. In our algorithm, the force, virial stress, and heat current for a given atom can be accumulated within a single thread and is free of write conflicts. We discuss the formulations and algorithms and evaluate their performance. A new open-source code, GPUMD, is developed based on the proposed formulations. For the Tersoff many-body potential, the double precision performance of GPUMD using a Tesla K40 card is equivalent to that of the LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) molecular dynamics code running with about 100 CPU cores (Intel Xeon CPU X5670 @ 2.93 GHz).

Thermal conductivity decomposition in two-dimensional materials: Application to graphene

“…Only homogeneous systems have been considered in this work, and it would be interesting to extend the formalisms to study interface heat transport in inhomogeneous systems. Computer implementation of the methods presented here will be made available in the near future through the GPUMD code [47]. 11404033).…”

“…We performed all the MD simulations using graphics processing units molecular dynamics (GPUMD) [40,46,47], an MD code which attains high performance on graphics processing units. To model the interactions between the carbon atoms, we use the Tersoff potential [48] with optimized parameters for graphene [49].…”

Influence of thermostatting on nonequilibrium molecular dynamics simulations of heat conduction in solids

“…The NEMD, HNEMD, and spectral decomposition methods have been implemented in the highly efficient Graphics Processing Units Molecular Dynamics (GPUMD) package [47][48][49], which is the code we used for most of the MD simulations. The simulations of thermal rectification in asymmetric graphene devices were performed using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) package [50].…”

Thermal conductivity reduction in carbon nanotube by fullerene encapsulation: A molecular dynamics study

“…To clarify these issues, here we investigate thermal transport in individual C 60 @(10, 10) CNP by using three different MD methods, including the aforementioned EMD and NEMD methods and the homogeneous nonequilibrium MD (HNEMD) method [25], all of which were implemented in an efficient open-source graphics processing units molecular dynamics (GPUMD) package [26,27]. The high efficiency of GPUMD allowed us to explore length scales directly comparable to the experimental situations [18], which are inaccessible to previous MD simulations.…”