Parallelization of Simulations for Various Magnetic System Models on Small-Sized Cluster Computers with MPI
The topic of parallelization of physical simulations has become an important part of scientific work today. However, except for the simple Ising spin model, simulations of various magnetic systems, e.g. the Heisenberg model, on small to moderate size cluster computers were not in strong focus within the field of Computational Physics. The work presented in this paper is a contribution to fill exactly this gap. The feasibility and the benefits of distributing such simulations among several processes are demonstrated by means of simulations of three physical models in this context: a 2-dimensional Ising model, the Heisenberg model, and a magneto-dipolar glass model. Herein we present these models and the applied parallelizational techniques. In the following, we show that with our parallelization scheme an almost ideal speed-up can be achieved on cluster computers by using MPI.
KeywordsIsing Model Cluster Computer Heisenberg Model Metropolis Algorithm Parallelization Scheme
Unable to display preview. Download preview PDF.
- 3.Ising, E.: Zeitschrift Fuer Physik. 31, 253 (1925)Google Scholar
- 5.Grove, D.A., et al.: Cluster Computing with iMacs and Power Macintoshes. In: Proc. Parallel and Distributed Computing Systems (PDCS 1999), Fort Lauderdale (1999)Google Scholar
- 6.Fox, G., et al.: Solving Problems on Concurrent Processors. Prentice Hall, Englewood Cliffs (1988)Google Scholar
- 8.Amdahl, G.: Validity of the single processor approach to achieving large-scale computing capabilities. In: AFIPS Conference Proceedings, vol. 30, pp. 483–485 (1967)Google Scholar
- 9.Quinn, M.J.: Chapter Performance Analysis in Parallel programming in C with MPI and OpenMP. McGraw-Hill, Boston (2003)Google Scholar
- 13.Metropolis, N., et al.: J. Chem. Phys. 21, 1087 (1953)Google Scholar
- 15.Berkov, D.V., Gorn, N.L.: Susceptibility of the disordered system of fine magnetic particles: Langevin-dynamics study. J. of Physics Condensed Matter 13 (2001)Google Scholar
- 16.Burns, G., Daoud, R., Vaigl, J.: LAM: An open cluster environment for MPI. In: Proceedings of Supercomputing Symposium 1994. Univ. of Toronto, pp. 379–386 (1994)Google Scholar
- 17.Karp, A.H., Flatt, H.P.: Measuring parallel processor performance. CM 33(5), 539–543 (1990)Google Scholar