A Beowulf cluster is a group of everyday computers that is connected together to function as a single, more powerful computer. By harnessing the resources of many computers, a Beowulf cluster is capable of performing many tasks in parallel, or at the same time, and can solve many problems much faster than traditional computers.
UAA's Computer Science Beowulf cluster consists of 13 custom-built boxes, each containing a dual processor 1 Ghz Pentium III and 768 Mb of shared memory. This gives us a total of 27 nodes, including the master. The cluster is running NetBSD 2.0F and is connected through a 100Mbps switch.
The system is configured with MPICH, a common framework for parallel programming in C or C++. The cluster will be used in conducting a number of computer science courses, but it is also available for faculty use on research projects. The cluster was formely housed at the Alaska Experimental Forecast Facility and was used for atmospheric modeling and simulation.
If you are interested in working on the cluster, contact Dr. Kenrick Mock (firstname.lastname@example.org) or one of the Computer Science Lab Technicians (email@example.com).
Parallel Programming Examples, with a flavor toward complex systems simulation, written by Selah Lynch.