Eversole Jay, Anderson Wendell
MatLab™, Octave, OpenMP, parallel methods, biological warfare, optical sensors
The Center for Computational Science (CCS) of the US Naval Research Laboratory (NRL) conducts leading edge research in High Performance Computing. CCS currently has two SGI Altix 3700s, one Altix 4700, and one SGI ICE machine. Recently the center has seen an increased interest from scientists at NRL who have been running MatLab™ on their local PC's and workstations but who need more computational power. One such application is the development of biological warfare point detection sensors where time-to solution for a single run can take over 30 hours to complete and many runs are necessary during the development process. This paper describes the issues that were encountered in the port of this code to the SGI High Performance Computing (HPC) computers at NRL and provides a paradigm for moving other computationally intensive MatLab™ programs to HPC machines.