Issue No.02 - March/April (2005 vol.7)
Matthias K. Gobbert , University of Maryland, Baltimore County
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MCSE.2005.29
Beowulf clusters use commodity components in their individual compute nodes, such as Intel Xeon CPUs, yielding a good return on investment. With a high-performance communication network and extended hard-disk storage, they become viable platforms for large-scale parallel applications. The judicious combination of a numerical algorithm, its efficient implementation, and the right hardware can achieve parallel computing's two fundamental goals: to solve problems faster and to solve larger problems than on a serial computer.
reaction-diffusion equation, finite-element method, matrix-free iterative method, cluster computing, high-performance interconnect
Matthias K. Gobbert, "Configuration and Performance of a Beowulf Cluster for Large-Scale Scientific Simulations", Computing in Science & Engineering, vol.7, no. 2, pp. 14-26, March/April 2005, doi:10.1109/MCSE.2005.29