SCREMS: High Performance Scientific Computing Environment

Kent State University, Kent OH

Investigators

Abstract

The Departments of Mathematical Sciences and of Computer Science at Kent State University will purchase a high performance parallel cluster computer to enhance the research and educational activities in the area of applied mathematics and scientific computing at both departments. Scientific computing has become an important investigation tool in many areas of science and engineering. High performance computing on large parallel clusters is routinely required for many research and application problems. The Departments of Mathematical Sciences and of Computer Science at Kent State University have research areas which include parallel numerical algorithms for solving boundary value problems for partial differential equations, iterative solution methods for large linear systems of equations, iterative methods for the computations of a few eigenvalues or singular values of a large matrix, solution of large-scale ill-posed problems, image processing, simulation and analysis of liquid crystals, etc. The cluster computer will contain one master node and 12 computation/visualization nodes connected through a high speed InfiniBand network in a single rack. Each computation node contains 4 processors (two 2.8GHz dual-core AMD CPUs) with dual PCI-E graphics cards and 32GB of RAM. This high performance computing environment will provide sufficient computing support for high-quality research activities in the area of applied mathematics and scientific computing at both departments.