Computational Resources

The Center houses state-of-the-art computational clusters that give students and faculty access to sophisticated hardware tools that can be leveraged to solve large-scale parallel computing and data-intensive tasks. Available computing resources include:


  • MGHPCC (remote UMass-wide resource): The machine has over 16,000 x86 processor cores. The MGHPCC wiki provides system details including the form to request access.
  • HPCC (local CSCVR resource): A 4 year old 80-node IBM iDataPlex system with each node supporting 8 Xeon-cores and 24 GB memory and an Nvidia Tesla "Fermi" GPU. There is 30 TB attached storage. The system runs RHEL 5.5 along with standard (mostly) open-source software. The job manager is PBS torque.
  • ARNiE (local CSCVR resource): This is our newer 40-node Supermicro system with each node supporting 8 Core-i7-cores and 24 GB memory and an Nvidia Tesla "Fermi" GPU. There is 200 TB attached storage. The system runs CentOS 6.6 (part of Rocks 6) along with standard open-source software. The job manager is Sun Grid Engine.
  • RPS (local CSCVR resource): A rapid prototyping server, i.e. a single high-end Linux server for prototyping / testing code quickly for all CSCVR faculty and students. Specs of the server: 32-cores Intel Xeon @ 2.2 GHz; 32 GB RAM; Nvidia Tesla K40 GPGPU (12GB vRAM); Ubuntu Linux 15.04; MATLAB & Jupyter server installed.
  • ELROY & CONDOR (local CSCVR resources): Two unique clusters built entirely out of 300 Sony PlayStations (PowerPC Cell architecture) and 32 Nvidia Tegra tablets (ARM mobile-devices). They run an architecture-specific Linux. Utilizing them effectively typically involves low-level architecture specific coding.