MRI: Acquisition and Analysis of a Multidisciplinary Beowulf Cluster

Calvin University, Grand Rapids MI

Investigators

Abstract

EIA-0079739 Adams, Joel C. Calvin College MRI: Acquisition and Analysis of a Multidisciplinary Beowulf Cluster It is proposed to design, acquire, build, and test a ``Beowulf cluster`` multiprocessor that will provide a high performance computing environment for multidisciplinary scientific research and research training at Calvin College. The objective is to determine what cluster design is best suited for research projects in multiple disciplines. More precisely, it is proposed to spend three years studying the performance of three common cluster topologies on different kinds of research problems. The aim of the study is to identify the cluster topologies that are best suited for particular kinds of problems, as well as identify what topology provides the best overall performance for multidisciplinary research machine.