Estudos comparativos de ferramentas de programação para grades computacionais

Abstract

Parallel and distributed computing is widely used to solve problems that demand high computational power. There are many platforms that allow executing parallel and distributed programs. One of these platforms is the computational grids, composed by many computers dispersed in many sites. There are many tools to help programmers to develop computational grid applications. The present work intended to compare studies of some of the computational grid programming tools. This comparison aim to help programmers to choose between these tools, decreasing the necessity for practical tests. To execute this comparison, practical tests with an implementation using these tools were run. An application that solve N-Queens problem was created using ProActive and POP-C++. Therefore, both tools have shown good performance, but both need to improve their features. ProActive is a Java tool, but this fact does not interfere on the application performance, so suggest that Java could be a choice for computational grid programming.