The Center for Scientific Computing and Data Science Research (CSCDR) completed the purchase and installation of a new computer cluster, which was installed as part of the UNITY collaboration. This cluster was funded by a $600K award from the Air Force Office of Scientific Research (AFOSR) Defense University Research Instrumentation Program (DURIP), and will foster the development and testing of novel mathematical methods suitable for large-scale parallel scientific computing and data science applications. This is the fourth equipment grant award the CSCDR has secured, resulting in the purchase of over $1.6M in shared research computing hardware.
The proposal was developed by a team of computational scientists affiliated with the CSCDR: Alireza Asadpoure, Collin Capano, Yanlai Chen, Zheng Chen, Geoffrey W. Cowles, Scott Field, Sigal Gottlieb, Alfa Heryudono, Gaurav Khanna, Arghavan Louhghalam, Maricris Mayes, Mehdi Raessi, and Mazdak Tootkaboni. The projects that will be enabled by the new cluster focus on the development of innovative computational mathematics and data science algorithms that are of interest to the AFOSR, including the development of novel and efficient algorithms; mathematical methods for model complexity reduction; data analytics approaches including machine learning and similar approaches; development of novel materials including metamaterials, multi-phase architected materials, quantum materials, and materials for energy storage.
The new cluster consists of 53 compute nodes, each with 64 Intel x86 cores connected by a high speed, low latency Infiniband network, 3 energy efficient nodes, each with an 80-core ARM architecture processor, and 1 GPU node with 4x NVIDIA A100 GPUs on an NVLINK interconnect. The installed hardware comprises a total of 3,696 compute cores, 4 GPUs, and 15TB of memory and has an aggregate theoretical capability of completing 300 trillion double-precision calculations each second (Tflops FP64). The cluster was designed and built by Microway Inc., Plymouth MA.
The hardware has been integrated into the UNITY Machine at the Mass Green High Performance Computing Center in Holyoke, MA. UNITY represents a partnership of UMass Dartmouth, UMass Amherst, and the University of Rhode Island and offers researchers over 12,000 cores and 1000 GPUs as well as IT and facilitator support. UMassD faculty all have access to the entire Unity cluster.