GPU implementations of some many-body potentials for molecular dynamics simulations. (September 2017)

• Record Type: Journal Article
• Authors: Minkin, Alexander S.; Knizhnik, Andrey A.; Potapkin, Boris V.
• Is Part Of: Advances in engineering software. Volume 111(2017)
• Page Start: 43
• Page End: 51
• Publication Date: 2017-09
• Subjects: GPGPU -- GPU -- OpenCL -- Many-body interatomic potentials -- Tersoff potential -- Embedded-atom potential -- EAM potential -- Atomic operations -- Semaphores -- Critical sections; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2016.05.013
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: Using atomic operations is optimal for GPU implementation of Tersoff potential. Atomic operations do reduce the efficiency of GPU implementation of EAM potential. Numerical force computation algorithm shows the best GPU scaling. GPU are not always effective for small number of threads. Significant acceleration can often be seen for large number of threads. The performance of GPU algorithms depends on floating point precision. High GPU acceleration values do not always mean algorithmic efficiency. Abstract: We study the efficiency of OpenCL implementations for Tersoff and embedded-atom interatomic potentials. We show that Tersoff potential can be computed faster using atomic operations rather than using longer kernel code. On the contrary better performance can be gained for embedded-atom potential without atomic operations. Numerical force computation algorithm is the slowest but shows the best scaling with the highest GPU acceleration. The GPU acceleration of different algorithms was evaluated and compared to the serial implementations of similar algorithms. The performance of GPU implementations is superior to their serial counterparts and depends on the algorithm and arithmetic precision. The corresponding benchmarks and performance comparison were done using NVidia GPUs and Intel CPUs.