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. We are also members of the XSEDE Campus Champions program. Available computing resources include:

• Unity (shared, external resource): Shared HPC cluster managed by UMass Amherst Please see the Unity website for more.
• Carnie (local CSCDR resource): A $650K Center grant is funding an on-campus cluster with about 1300 CPU-cores, 20 GPUs, 132TB storage, and 3TB of memory. 50 compute nodes will have 24-core Intel skylake cores, 48GB memory, and 1TB of SSD local disk (15 of these will have one Nvidia V100 GPU each). There will be 2 OpenPOWER nodes, a 32-core IBM POWER9 nodes, 128GB memory, 1TB SSD local disk (2 Nvidia V100 GPUs with NVLINK2). Nodes will connected with Infiniband EDR + 10G Ethernet. Software will include TotalView license (128 processes), PGI compilers with OpenACC (floating license), and Intel compilers / libraries.
• ARNiE (local CSCDR resource): This is our older 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. Please see the Carnie wiki for more.
• RPS (local CSCDR resource): RPS (local CSCDR resource): A rapid prototyping server, i.e. a single high-end Linux server for prototyping / testing code quickly for all CSCDR faculty and students. Specs of the server: 16-cores (32 threads) Intel Xeon E5-2660 @ 2.2 GHz; 32 GB RAM; Nvidia Tesla K40c GPU (12GB RAM); Ubuntu Linux 16.04 LTS; Popular rapid prototyping software such as anaconda python 2 and 3, Julia, MATLAB with parallel computing toolbox, and Mathematica with GPU support are installed.
• RPS2 (local CSCDR resource): A rapid prototyping server 2, i.e. a single high-end Linux server for prototyping / testing code quickly for all CSCDR faculty and students. Specs of the server: 16-cores (32 threads) Intel Xeon E5-2667 @ 3.2 GHz; 256 GB RAM; Nvidia Pascal Titan X GPU (12GB RAM); Ubuntu Linux 16.04 LTS; Popular rapid prototyping software such as anaconda python 2 and 3, Julia, MATLAB with parallel computing toolbox, and Mathematica with GPU support are installed.
• ELROY & CONDOR (local CSCDR 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.

For CSCDR faculty looking to include information about CSCDR computing resources in their proposals, we have provided some templates you may find helpful in this OneDrive folder. Additional proposal preparation tools are also included.