Bell Cluster

Named in honor of Clara Bell Sessions, minority advocate and Professor and Director of Continuing Education of Nursing, the Bell Cluster would become a workhorse for medical and community research. Like many other University Research Centers, Purdue was limited by the size of their facility and original building layout which prevented the addition of newer generation, higher density compute nodes. Direct Liquid Cooled Dell servers would allow them to focus on performance and providing researchers the best services possible, while working with their existing building limitations. Replacing the previous Conte (https://www.rcac.purdue.edu/compute/conte) compute system, Bell (https://www.rcac.purdue.edu/compute/bell/) is a Community Cluster with theoretical peak performance of 5.3 PFLOP’s, optimized for communities running traditional, tightly-coupled science and engineering applications. Bell was built through a partnership with Dell and AMD over the summer of 2020, and consists of Dell compute nodes with two 64-core AMD Epyc 7662 “Rome” processors (128 cores per node) and 256 GB of memory.

• #44 on the Green500 list and #431 on the Top500 list.
• Supporting critical medical research
• Bell liquid cooling technology:
  • 10x racks with direct liquid cooling Dell EMC PowerEdge C6525 servers with factory-installed Passive Coldplate Loops
  • 10x Stainless steel Rack Manifolds, featuring 100% dry-break quick disconnect technology
  • 2x row-based CDUs (fully redundant) managing 10x racks together
  • Custom secondary fluid network above the racks