Beowulf Project

Beowulf Project

Alvin Sherman Library 1054

Academic computing has long relied on supercomputers, such as Cray, to provide the computing power needed for complex scientific computations. Even though such computers in many cases supplied the computational power needed, such systems were also very expensive, and relied upon proprietary components. Beowulf systems offer a new, inexpensive solution to the problem of limited processing power. Consisting of networked desktops and commodity components Beowulf systems attempt to provide supercomputer performance at a reasonable cost.

A NASA contractor built the first Beowulf to address problems associated with the large data sets in space science applications. Since then the system has been put to use to try and tackle various scientific problems ranging from nuclear engineering applications to applications for the Center for Disease Control. Government agencies and universities are major factors in the system’s proliferation into mainstream technology culture.

Beowulf recipes vary in accordance to their creators. Most Beowulf systems utilize publicly available software such as a Linux operating system, GNU compilers and programming tools, and MPI and PVM message passing libraries. Pentium processor machines connected by 100Mbit/s Ethernet and 100Mbit/s Ethernet switches generally make up the hardware components of a Beowulf system.

In recent years, Beowulf has gained recognition as a parallel processing multi-system high performance computing solution. This presentation will include what has been learned about Beowulf, the necessary components for a Beowulf system, Beowulf installation and Beowulf Project goals.