Magnetohydrodynamics on Heterogeneous architectures: a performance comparison
Department of Physics, University of Toronto, Toronto ON, M5S 1A7, Canada
arXiv:1004.1680 [cs.PF] (10 Apr 2010)
@article{pang163magnetohydrodynamics,
title={Magnetohydrodynamics on Heterogeneous architectures: a performance comparison},
author={Pang, B. and Pen, U. and Perrone, M. and Treibig, J. and Wellein, G. and Hager, G. and Altisen, K. and Liu, Y. and Moy, M. and Wang, Y. and others},
journal={Architecture},
volume={163},
number={323},
pages={643}
}
We present magneto-hydrodynamic simulation results for heterogeneous systems. Heterogeneous architectures combine high floating point performance many-core units hosted in conventional server nodes. Examples include Graphics Processing Units (GPU’s) and Cell. They have potentially large gains in performance, at modest power and monetary cost. We implemented a magneto-hydrodynamic (MHD) simulation code on a variety of heterogeneous and multi-core architectures — multi-core x86, Cell, Nvidia and ATI GPU — in different languages, FORTRAN, C, Cell, CUDA and OpenCL. We present initial performance results for these systems. To our knowledge, this is the widest comparison of heterogeneous systems for MHD simulations. We review the different challenges faced in each architecture, and potential bottlenecks. We conclude that substantial gains in performance over traditional systems are possible, and in particular that is possible to extract a greater percentage of peak theoretical performance from some systems when compared to x86 architectures.
November 11, 2010 by hgpu
