On the parallelization of a commercial PSE for scientific computing

High performance computing is the paradigm of choice in the solution of many real world problems, where the complexity of large-scale phenomena or the need for real time results requires the use of increasingly powerful computational resources. Although low cost off the shelf supercomputers, namely Beowulf class systems, have gained a widespread popularity, an effective use of high performance computing resources and its diffusion still depends on the availability of easy-to-use software tools enabling a transparent access to computing resources. In that scenario many projects are devoted to the implementation of problem solving environments (PSEs) for high performance computers. The authors describe the parallelization of IDL, a commercial product by Research Systems, that enables an intuitive use of a Beowulf supercomputer and we assess its performance with wall clock execution times and some well known parameters.