Acquisition of a High-Performance Computing Cluster Environment in Support of Scientific and Engineering Applications Grant uri icon

abstract

  • East Carolina University (ECU) request funding for the acquisition of a distributed memory high performance computing (HPC) cluster computing system designed to foster the application of computational methodologies to complex problems in support of chemistry, physics, biology and engineering. The work highlighted in the proposal cites three core projects and four ancillary projects that require a High Performance Computing (HPC) Cluster. This diverse set of core projects include detailed investigations of important catalytic reactions, reconstructing the phylogenetic relationships of mega-diverse animal groups, and modeling the complex fluid dynamics of internal combustion engines. At present, East Carolina University lacks a high performance computing cluster to address the growing computational needs of its researchers and educators. The acquisition targets a configuration that includes: 1. 128 node (each node dual processor) Opteron system (similar to the X326 E class server) 2. the node is a form 1 u with 30 to 40 nodes per housing 3.Memory configuration must be 2-4GB/node (8-16GB/frame depending on memory costs) 4.System DASD must have a fast external I/O using GPFS or BlueArc with 10000rpm SAT A drives 5. total storage should approximately 10 TByte with GigE or Myrinet interconnect 6. OS must be Red Hat Enterprise Server or Suse Enterprise Linux Currently, competitive quotes are being sought from IBM and SUN. Vendor selection will not be made pending award as systems configuration and pricing changes are likely to change substantially between submission and award.

date/time interval

  • September 2006 - August 2009