This facility provides computational resources for mechanics research of Brown faculty and external collaborators. It operates as a cost center administered by CAMR with a full-time director (Scheuerman). The major equipment includes a 58 node Opteron dual processor-dual core (232 processors) High Performance Computing (HPC) cluster, 2 Polyserve scalable fileserver, 2 Red Hat cluster suite routers, MSA1000 SAN with ~1 TB of user storage and an MSL5026 tape library. To ensure optimal performance and utilization of all compute nodes, we have deployed Standard LSF (a job scheduler) and Fairshare (a job queuing algorithm) from the HP XC cluster software suite. We have divided the HPC cluster into two partitions for serial and parallel processing to best utilize the additional memory needed for serial jobs. We have implemented several layers of high availability solutions (hardware and software) to avoid any system downtime and loss of data. Subsequent yearly upgrades to the facility are estimated at $50,000. A new high-performance facility is being installed in conjunction with the recent University HPC facility built in collaboration with IBM. The CMRF portion will more than double the capability of the existing facility, enabling far larger and longer computations to be performed locally at Brown. Details of this facility will be provided once up and running.
The facility has no instruments.