Software engineering methods for parallel and distributed scientific computing

Abstract

In this paper, we present an interdisciplinary research project whose central objective is to develop new software engineering (SWE) methods for distributed memory parallel scientific computing. Our emphasis is on putting into practice and evaluating the proposed methods. The main test case for their definition and evaluation is the parallelization of an industrial CFD software package. A major concern is to achieve maximum portability, i.e. cover a large number of target systems ranging from networks of workstations (NOWs) to massively parallel systems (MPPs). In order to optimize utilization of NOWs, a resource management system is being designed, which runs parallel applications in batch mode and dynamically assigns available resources to the tasks of the parallel applications. Besides giving an overview of the objectives followed in the project, we give a progress report, which details the achievements reached so far.