MRI/RUI: Acquisition of a Linux Cluster for Numerical Simulations of Excitable Media

Montclair State University, Montclair NJ

Investigators

Abstract

ABSTRACT Proposal No. CTS-0319555 Principal Investigator: R. Zaritski, Montclair State University The grant will facilitate acquisition of a Linux cluster with a total peak performance of 150 billion floating point operations per second. Tentatively, the cluster will consist of 32 dual AMD Athlon nodes interconnected by Gigabit Ethernet. Its main intended use is for numerical simulations of spiral and scroll wave interactions in excitable media. This highly multidisciplinary research project is a collaborative effort between the principal investigator, Dr. R. Zaritski (Montclair State University), and Dr. A. Pertsov (SUNY Upstate Medical University). It involves Applied and Computational Mathematics, Chemistry and Biology, Physics and Computer Science. To increase utilization, the cluster will also be used for numerical simulations of disease dynamics, an auxiliary research project headed by Dr. L. Billings (Montclair State University), and, possibly, for other auxiliary projects. Examples of excitable media include a variety of physical, chemical, and biological systems. Two most well-known examples are Belousov-Zhabotinsky (BZ) chemical reaction and electrochemical activity in the cardiac tissue. Virtually all known excitable media support special vortex-like waves known as spiral waves in two dimensions and scroll waves in three dimensions. Spiral and scroll waves in human heart correspond to cardiac arrhythmias, the leading cause of death in industrialized countries. Interacting spiral and scroll waves are believed to be the primary mechanism of fibrillation, the most deadly cardiac arrhythmia. Spiral and scroll wave interactions have been studied for many years, but are not yet well understood. Numerical simulations of such interactions are very computationally intensive and, especially in the three-dimensional case, require supercomputing resources. The broader impact of this project at Montclair State University, which is a major public, predominantly undergraduate institution serving the diverse New Jersey population, will include: enhancing the university high-performance computing infrastructure, attracting more students to hands-on multidisciplinary research, and establishing and strengthening a cross-institutional research partnership between a teaching and a research university. Eventually, computer modeling of excitable media may lead to better prediction and control of heart arrhythmias and to new computing paradigms, based on chemical reactions. Results of the proposed investigation will be disseminated through national and international professional journals and meetings.