This paper reports experience and describes the tools used in parallelizing large application programs of two kinds : Monte--Carlo simulation and linear programming optimization. These techniques are used in a large number of applications in power system operations and planning, and in various fields of engineering and science. Parallelism is analysed on the subroutine level and the programmer will be responsible for adapting the existing application code to the parallel environment. The tools comprise a multi-processor computing system with parallel processing software facilities, a concurrent processing simulator for helping on problem partition and debugging, and a data-base containing information on the application's variables and subroutines. The goal is to reduce execution times taking into account the solution structure. Three case studies illustrate the application of the above tools and the development methodology. Results obtained show a very high efficiency in tbe use of the concurrent processors.

http://www.lbd.dcc.ufmg.br/colecoes/sbes/1989/004.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web Processando consulta.