Introduction to Parallel Programming and pMatlab v2.0

Abstract : The computational demands of software continue to outpace the capacities of processor and memory technologies, especially in scientific and engineering programs. One option to improve performance is parallel processing. However, despite decades of research and development writing parallel programs continues to be difficult. This is especially the case for scientists and engineers who have limited backgrounds in computer science. MATLAB?, due to its ease of use compared to other programming languages like C and Fortran, is one of the most popular languages for implementing numerical computations, thus making it an excellent platform for developing an accessible parallel computing framework. The MIT Lincoln Laboratory has developed two libraries, pMatlab and MatlabMPI, that not only enables parallel programming with MATLAB in a simple fashion, accessible to non-computer scientists. This document will overview basic concepts in parallel programming and introduce pMatlab.