Mecánica Computacional, Volume XXVIII. Number 5. High Performance Computing in Computational Mechanics

A Tool for Automatic Parallelization of CPU-Intensive Java Applications on Distributed Environments

Cristian Mateos, Alejandro Zunino, Marcelo Campo

Abstract


Grid Computing, a relatively new paradigm for distributed computing, delivers the necessary computational infrastructure –the so-called Computational Grids– to perform resource intensive computations such as the ones that solve the problems scientists are facing today. Exploiting Computational Grids comes at the expense of explicitly adapting the conventional software implementing such problems to take advantage of Grid resources, which requires knowledge on Grid programming. The recent notion of gridifying conventional applications, which is based on semi-automatically deriving the Grid-aware version from the compiled code of an ordinary application, promises users to be relieved from the requirement of manual usage of Grid APIs within their codes. This paper describes a novel Java-based gridification tool, which allows users to automatically parallelize applications on local-area and wide-area Grids. Experiments confirm that BYG effectively exploits such Grids while delivers competitive performance with respect to manually using Grid APIs to gridify conventional software.

Full Text: PDF