NSF-1228312, MRI: Acquisition of High Performance Computing Instrument for Collaborative Data-Enabled Science (Senior Investigator)

This project, acquiring a high performance computing (HPC) for collaborative data-enabled science, aims to enable fundamental and applied research in areas of experimental HPC, including scalable file systems, accelerator technologies, middleware architectures, and cluster scheduling policies. The instrument will be used as a production resource to support several compelling science applications in materials and biomolecular modeling, bioinformatics (and others) and will contribute a significant new component of the institution?s well established and managed Palmetto cluster while providing critically needed file storage, and accelerator systems not currently available. Supported projects include:

• Scalable file system implementation
• Exploiting and analyzing concurrency in heterogeneous computing environments
• High performance cluster modeling, analysis and characterization
• File replication and consistency maintenance
• Cloud architecture
• Properties of advanced materials and biomolecular systems
• Molecular dynamic simulation studies of poly-dots, new light-emitting nanoparticles
• Biomolecular evolution
• Optimizing catalysts for direct chemical remediation from water
• Protein-surface interactions
• Map/Reduce framework for next generation sequencing data assembly
• Searching for identification of significantly conserved protein blocks
• Digital manufacturing
• Statistical inference in nonparametrics models of production
• Analysis of the structures of hospital and banking industries
• Multiple scatter rendering
• Digital scholarship
• Applications of computing to the geosciences.

Regular planned input from the scientific team is expected to facilitate the development of policies for use of the instrument as a test bed for experimental high performance computing systems research as well as a shared production resource for scientific applications.

Broader Impacts:

The instrument serves as a tool for the collaborative team that includes more than 20 faculty from 4 institutions, and more than 250 graduate and undergraduate students. The system will also be utilized by the already established ?Desktop to Teragrid? program that reaches more than a dozen universities across South Carolina, an EPSCoR state, and with regional partners. The proposal team includes 6 women senior personnel. The institution has outreach, education, and training activities, and has engaged SC?s only public HBCU, South Carolina State University, as well as other minority institutions within the region. Development of interdisciplinary curricula will support computational and data-enabled science and engineering.