Evaluation of a methodology for the reverse engineering and parallelization of sequential code

Abstract

A general methodology based on software engineering principles is proposed for the parallelization of existing sequential code. The utility of the proposed methodology is evaluated through a case study involving a numerically intensive application in the domain of petrochemical exploration. The methodology does not assume the existence of detail design documentation for the sequential code. The methodology involves three basic phases: (1) reverse engineering; (2) parallel design; and (3) parallel implementation. The process iterates between phases two and three until the values of the performance metrics satisfy project requirements. In addition to the methodology itself considerable detail related to the experiences and lessons learned in performing the case study, are included.