Computational Study of Ligand-Protein-Nanafiber Interactions for the Development of Biosensors
Name: Dr. Jianping Zhu
Department: Department of Theoretical and Applied Mathematics
Position: Professor and Chair
Institution: University of Akron
The requested Itanium cluster will be used for an interdisciplinary research project on the computational study of ligand-protein-nanofiber interactions for the development of biosensors, and a number of other ongoing research projects in the five departments represented by the PIs at the University of Akron. The main approach of the research is multi-scale molecular dynamics (MD) and quantum mechanics (QM) simulations using high performance cluster computing and grid computing. More specifically, the PIs will (1) develop efficient parallel algorithms for MD simulations; (2) develop multi-scale MD and QM simulation models for ligand-protein-nanofiber interactions; (3) develop a grid-based simulation system with new parallelization method; (4) develop efficient loop scheduling techniques; and (5) combine computational simulations with experimental results to study ligand-protein-nanofiber interactions for the design and construction of novel nanostructure-based biosensors.
The requested Itanium cluster and the Cluster Ohio project would have an enormous and immediate impact on the proposed research. They will also significantly enhance interdisciplinary collaborations at the University of Akron, provide a valuable resource for teaching of parallel computing and simulation methods in upper division and graduate level courses, and facilitate incorporation of the latest research into existing courses in mathematics, computer science, biochemical engineering, physics, and polymer science. The cluster would be stored in our dedicated server room with ample space, power, and dedicated climate control system. The data could be backed on both a large RAID5 array and magnetic tape as required.
Advancement of Computational Research at Miami University
Name: Dr. Stephen Wright
Department: Department of Mathematics and Statistics
Position: Associate Professor
Institution: Miami University
This proposal requests a 32-processor cluster in support of a new Center for the Advancement of Computational Research being established this year at Miami University. The goals of the new Center include expanding opportunities for computation in research and supporting the research efforts of both faculty and students in the sciences and engineering. Obtaining the requested cluster would be an important first step in establishing our new Center. The forty faculty members who have agreed to participate in the new Center belong to a dozen different academic departments, with a diverse range of disciplinary and interdisciplinary research interests, so only a representative set of five projects is described in this proposal.
Distributed Computing to Support Interdisciplinary Research in Space Geodesy and Earth Sciences Using the OSC 16-node SGI Cluster
Name: Dr. C.K. Shum
Department: Department of Civil and Environmental Engineering and Geodetic Science
Position: Associate Professor
Institution: The Ohio State University
We propose to host the Ohio Supercomputer Center Cluster Ohio Project supplied 16-Node SGI 1400 Itanium-based cluster system to: (1) conduct interdisciplinary research using cluster-style distributing computing in the areas of interdisciplinary research in space geodesy and geosciences, including geophysical inverse problems of gravity field modeling, atmospheric water vapor retrieval, and ice-stream velocity retrieval, using advanced gravity mission measurements, data from spaceborne GPS-equipped Low-Earth Orbiters (LEOs) and GPS network, and Synthetic Aperture Radar interferometry (InSAR) data; (2) develop parallel and distributed computing algorithms and tools using publicly available utilities like Message Passing Interface (MPI) [Gropp et al., 1994] and Parallel Linear Algebra Package (PLAPACK) [van de Geijn, 1997] and OSC numerical libraries (e.g., ScaLAPACK, PBLAS, etc.) for geophysical inverse problems in Earth gravity field determination, 4-D data processing to retrieve GPS water vapor, and
processing of radar data associated with InSAR for mapping of ice stream velocities; and (3) develop visualization tools specifically applicable to GPS meteorology and InSAR applications. The tools will be developed based on existing software systems currently operational or under development at the Laboratory for Space Geodesy and Remote Sensing Research, located at Bolz Hall 246, Department of Civil and Environmental Engineering and Geodetic Science, The Ohio State University.
The three interdisciplinary research foci (gravity field modeling, GPS meteorology, and InSAR) are relevant to currently funded investigations by NSF, NASA, NIMA, DLR (German Space Agency and Remote Sensing Facility), and NASDA (Japanese Space Agency). The proposal intends to request a 16-node SGI 1400 cluster to be hosted at the Laboratory, and the researchers will conduct research using this system and other systems available on the OSC cluster system as well as OSC's Cray SV1 and teraflop HP zx6000 cluster platforms. The specific software systems to be modified to adapt to distributed computing include LLISS (Large Linear System Solver), PLAPACK, and the ASF (Alaskan SAR Facility) SAR/InSAR Processing Package. The developed tools will be made available to the science community and unused cycles on the cluster will be made available to statewide users.
Cluster Computing for Large Scale Neural Networks, Data Mining, Simulation, Fault Tolerance and Sequence Analysis
Name: Dr. Nikolaos Bourbakis
Department: Information Technology Research Institute and Department of Computer Science and Engineering
Position: Director and Professor
Institution: Wright State University
The efficient function of today's technological society has drastically increased the computational needs in data processing and interpretation, knowledge extraction, communication, computing, health, etc. In order for these needs to be satisfied huge amounts of computational power not available in a simple PC or workstation today are required. Thus, the scientific community has to develop new computational methodologies (parallel algorithms) able to efficiently work and highly perform in parallel machines, where the computational power is available. At the same time, the scientists have also to carefully evaluate the cost of these parallel computing machines to be as low as possible in order to make their use more feasible and accessible to a large variety of applications.
This proposal consisted of five projects, such as large
scale neural networks processing, data mining of large scale databases, simulation of large scale networks, fault tolerance of distributed parallel computing systems, and analysis of large sets of biological data (sequences), and deals with these particular issues and require parallel computing support from OSC through the Cluster Ohio Project to efficiently produce results (software methodologies) valuable for economic development, health care, and communication.
Computational Science at Otterbein College: A Proposal to the Cluster Ohio Grant Project
Name: Dr. David Robertson
Department: Physics and Astronomy
Position: Assistant Professor
Institution: Otterbein College
In this proposal, we request a grant of an 8-node (16 CPU) Itanium cluster to support a developing effort in computational science at Otterbein College. The proposed system will support ongoing faculty research in the Departments of Chemistry and Physics, as well as provide a resource for teaching in all disciplines with an interest in computational science.