COMPUTATIONAL CORE

Princeton University, Princeton NJ

Investigators

Linked publications & trials

Abstract

Computational Core Facility: The goal of the computational core is to provide resources required by the scientists in the Center and the university to be at the cutting edge of computational biology in both research and education. We have an open, flexible computational space (-1,800 sq. ft.) in the first floor of the Cari Icahn Labs at Princeton that houses nine bioinformatics staff members. We have a 1,000 sq. ft. server room in the Icahn basement, complete with cooling systems and uninterrupted power supplies (UPSs);this facility houses tfie development and production servers for databases, web and application services, high-performance computing and networked storage systems. The Center has two primary compute resources for Center participants. The first system is available for custom applications that require Intensive computation;it is now composed of 54 nodes (108 CPU, plus two more for the head node). Since January 1, 2008, this system has run 54,534 jobs in 398,498,039 CPU seconds (496,052,183 real), or 12.6 CPU years (15.7 real). Based on user feedback, very heavy demand, and the near obsolescence of the current cluster, the Center plans to augment this system with a 384-CPU 48-node Linux-based cluster. Elsewhere on campus we own a 25% share of two supercomputers operated by Princeton's Office of Information Technology (Delia, a 768-processor cluster and an IBM BlueGene supercomputer). Center participants have used Delia a total of 103,103 real hours (11.8 years) in 130,930 jobs between January 1, 2008, and April 30, 2008. Data storage needs of the Center are met by Network Appliance redundant clustered storage systems that provide over 100 terabytes of raw storage capacity.