MRI: Acquisition of a Cyberinstrument for Interdisciplinary Computational Science and Engineering

Clemson University, Clemson SC

Investigators

Abstract

This project, acquiring, deploying, and maintaining a high performance computing (HPC) cluster, aims to support a broad range of data-enabled science and research training activities. Motivated by both fundamental and applied science agendas, the instrument will be utilized by at least the fourteen collaborative projects cited below: - Quality and Scalability of Topic Models; - Multigrid and Multiscale Optimization for Machine Learning and Networks; - Computational Studies of Multifunctional Polymers; - Gene Network Alignment at High-scale; - Modeling of Peptide-functionalized Polyethylene-glycol (PEG)-based Hydrogels; - Computer Simulations of Highly Thermostable Copolymer-enzyme Conjugates; - Enabling Modeling of Large Macromolecular Assemblages; - Computational Studies of Biomolecular Evolution; - Lattice Boltzmann Simulation of Vascular Blood Flow with Stents and Flow Diverters; - Automated Parallel Parameterization Methods; - Large-scale Simulations of Rare Events in Biological Systems; - Performance Optimization and Load Balancing for Large-scale Metal Cation Catalysts; - Fundamental Insights into Alkane Selective Oxidation in Atomic-scale Metal Cation Catalysts, and - Computation Determination of Chemical Stability of Nuclear Materials. The instrument is the primary resource for the projects in computer science, including scalable machine learning and methods for utilization of emerging memory and Graphics Processing Unit (GPU) technologies. Moreover, the cluster constitutes a much needed platform for collaborative development and execution of GPU-accelerated scientific applications. The solid-state disks (SSDs) and memory paths available through Omni-path enable high scalability of research applications that manage complex operations, Transforming the way scientific knowledge is attained and implemented to drive new technologies, new developments in computing, coupled with traditional computational methods, advance efforts in chemistry, physics, biology, and materials science. Broader Impacts: The instrument serves as an essential tool for advancing inquiry for at least 20 faculty in the collaborative team from Clemson and Claflin University, an Historically Black University partner. Three hundred fifty or more graduate and undergraduate students participate in the research proposed and also in the educational projects that will use the instrument. Furthermore, configured as part of the Clemson "Palmetto" supercomputer cluster, the system is accessible to other faculty, staff, postdocs, and students and provides benefit to world-class operational user support and offers a path for sustainability for the instrument. The team includes nine women researchers and faculty participants who serve as role models. The participation of women, as well as in graduate programs at this institution, exceeds the national average. Outreach, education, and training activities will reach dozens of undergraduate participants at Claflin University. Current and proposed courses and seminars are planned at all levels.