MRI: Acquisition of a High Performance GPU/CPU Cluster for Research and Innovation in Computational Sciences and Engineering

点击翻译按钮获取中文摘要

FIT is acquiring a high-performance computing (HPC) cluster with multi-core Intel CPUs, NVIDIA V100 GPUs, and several layers of storage at Florida Institute of Technology. The key research thrusts include natural sciences (chemistry, physics, mathematics, and meteorology), computer science and engineering, and medicine. The infrastructure would be added to the previous cluster funded by a 2009 MRI and will help establish the Institute for Computational Engineering and Sciences (ICES) at FIT.The design of this HPC cluster has taken the cue from the CERN laboratory in Switzerland, arguably the largest scientific center in the world. The CERN strategy to meet computing challenges has a 3-prong approach: scale out capacity with public clouds and HPCs; increase data center performance with hardware accelerators such GPUs, FPGAs and TPUs; utilize new computational techniques based on Artificial Intelligence (AI) and Machine Learning (ML). The present design comprises a computing architecture that includes an on-campus HPC cluster (the ICES HPC), which is based on GPU machine learning accelerators, as well as CPUs, with highly efficient networking to ensure rapid bi-directional access to other supercomputers and clouds to allow super-large computational tasks beyond the local HPC capacity to run on off-site hyperscale providers, such as those managed by Extreme Science and Engineering Discovery Environment XSEDE), after development and testing at the ICES HPC facility. The intellectual merit of the project pertains to the ICES as a multidisciplinary and interdisciplinary computational center, with a computer cluster equipped with the latest technologies for highto-performance computing and data storage that could allow Florida Tech to work in a collaborative environment to create new and innovative algorithms and techniques utilizing AI/ML for conducting computational-intensive and big-data research in science and engineering disciplines.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.