Abstract
A new dynamic load balancing algorithm is proposed for particle simulations on clusters of workstations which operates at two levels; it both dynamically defines groups of particles in which the force calculations are localised and at the same time adaptively assigns groups to processors based on the group sizes and runtime CPU usages. It therefore adaptively balances using both the dynamics of the simulation and the load usage of the cluster at runtime.
The algorithm is implemented in the POOMA framework and applied to a particle-in-cell approximation of a three dimensional elastic particle collision model. Load balancing metrics and parallel scalability is determined on an 8 quad- processor 833MHZ Compaq Alpha cluster connected by a gigabit ethernet.
The author thanks Alvin Chua and Steven Capper for their careful proof reading and useful comments on this article.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Hockney, R.W., Eastwood, J.W.: Computer Simulation using Particles. Taylor and Francis Inc, Abington (1988)
Kohring, G.A.: Dynamic Load Balancing for Parallelized Particle Simulations on MIMD Computers. Parallel Computing 21(4), 683–694 (1995)
Watts, J., Taylor, S.: A Practical Approach to Dynamic Load Balancing. IEEE Transactions on Parallel and Distributed Systems 9(3), 235–248 (1998)
Cummings, J.C., Humphrey, W.: Parallel Particle Simulations Using the POOMA Framework. In: Heath, M., et al. (eds.) Proceedings of the 8th SIAM Conference on Parallel Processing for Scientific Computing, SIAM, Philadelphia (1997)
Oldham, J.D.: POOMA Manual, A C++ Toolkit for High-Performance Parallel Scientific Computation, http://www.codesourcery.com (2002)
Lan, Z., Taylor, V.E., Bryan, G.: A Novel Dynamic Load Balancing Scheme for Parallel Systems. J. Parallel Distrib. Computing 62, 1763–1781 (2002)
Pearce, F.R., Couchman, H.M.P.: Hydra: a parallel adaptive grid code. New Astronomy 2, 411–427 (1997)
Othmer, C., Schüle, J.: Dynamic load balancing of plasma particle-in-cell simulations: The taskfarm alternative. Computer Physics Communications 147, 741–744
Stijnman, M.A., Bisseling, R.H., Barkema, G.T.: Partitioning 3D space for parallel manyparticle simulations. Computer Physics Communications 149(3), 121–134 (2003)
Rauber, T., Rünger, G., Scholtes, C.: Execution behavior analysis and performance prediction for a shared-memory implementation of an irregular particle simulation method. Simulation Practice and Theory 6(7), 665–687 (1998)
Hendrickson, B., Devine, K.: Dynamic load balancing in computational mechanics. Computer Methods in Applied Mechanics and Engineering 184(2-4), 485–500 (2000)
Kalé, L., et al.: NAMD: Greater Scalability for Parallel Molecular Dynamics. Journal of Computational Physics 151, 283–312 (1999)
Di Martino, B., et al.: Parallel PIC plasma simulation through particle decomposition techniques. Parallel Computing 27, 295–314 (2001)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Dixon, M.F. (2006). A Runtime Adaptive Load Balancing Algorithm for Particle Simulations. In: Dongarra, J., Madsen, K., Waśniewski, J. (eds) Applied Parallel Computing. State of the Art in Scientific Computing. PARA 2004. Lecture Notes in Computer Science, vol 3732. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11558958_86
Download citation
DOI: https://doi.org/10.1007/11558958_86
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-29067-4
Online ISBN: 978-3-540-33498-9
eBook Packages: Computer ScienceComputer Science (R0)