DGP: Dynamic Grouping of Particles for Parallel Molecular
Dynamics Simulations
Yoshiaki Takai
Computing Center of Hokkaido University, Sapporo, Japan 060-0811
Hironori Emaru
Hitachi Ltd., Systems Development Laboratory, Kawasaki, Japan 215-0013
Tsuyoshi Yamamoto
Graduate School of Engineering of Hokkaido University, Sapporo, Japan 060-8628
SUMMARY
Molecular dynamics simulations (MD) have become
very important in physics and chemistry. In this paper, we
propose dynamic grouping of particles (DGP) for effective
parallel processing of short-range MD. DGP is based on
a cyclic process of grouping the particles that are within
a registration distance of each other. Such a particle
group defines a parallel task that can be processed inde-
pendently during the registration step. We have imple-
mented our method on SR2201, a loosely coupled
parallel processor. Experimental results show that DGP
is twice as fast as a conventional method based on cell
partitioning. © 2000 Scripta Technica, Syst Comp Jpn,
31(9): 1017, 2000
Key words: Molecular dynamics simulations;
bookkeeping method; particle grouping; parallel
processing.
1. Introduction
In advanced science and technology, simulations by
means of high-performance parallel computers have been
very important. In particular, molecular dynamics simula-
tions play a crucial role in physical chemistry and material
sciences.
Molecular dynamics simulation (MD) is a kind of
molecular simulation in which a system is composed of a
large number of particles and their interactions. All of the
microscopic motions of every particle are traced to obtain
macroscopic properties of the system. In addition, MD is
also used for physically based modeling in computer graph-
ics [1, 2].
A simulation of many particles is an N-body problem
in general, which requires enormous computation time.
Well-designed parallel algorithms and data structures for
handling particles are necessary to process the simulation
in much shorter computation time.
In this paper, we focus on short-range molecular
dynamics simulations in which particle interactions are
limited within a short distance. We propose dynamic group-
© 2000 Scripta Technica
Systems and Computers in Japan, Vol. 31, No. 9, 2000
Translated from Denshi Joho Tsushin Gakkai Ronbunshi, Vol. J82-D-I, No. 6, June 1999, pp. 711717
10