Welcome to the Texas A&M Tensor Cluster

Funded by a National Science Foundation Major Research Instrumentation grant, the Tensor cluster represents the first large scale beowulf cluster based on AMD's Opteron processor. This cluster will accelerate and enhance scientific discovery at Texas A&M in a very cost effective manner. The Opteron processor in a cluster environment brings new levels performance to our researchers' applications allow scaling to much larger problem sizes.

Supplied by Appro International, the cluster nodes are housed in Appro's hyperBlade chassis allowing us to achieve a sub-1U node density. The primary interconnect is Gigabit Ethernet with a subset of the nodes utilizing 4X InfininBand connectivity provided by InfiniCon Systems. Details of cluster are available on our configuration page.

Installation was completed in October of 2003, followed by two months of a development period. The stability and performance of the cluster has been quite satisfactory and we entered a production computing environment in January, 2004. Today we are running at or above 90% capacity, providing valuable computing cycles to our researchers.

This project is funded by the National Science Foundation, the Texas A&M Office of the Vice President for Research, and the Texas A&M College of Science. We would like to acknowledge our partnerships with AMD, Appro, and InfiniCon. Special thanks to the Texas A&M Computing and Information Services for housing the cluster in their facility.