Some results on ongoing research on parallelimplementation of graph algorithms

In high performance computing, three recognized important points are us-ability, scalability and portability. No models seemed to satisfy these three steps till recently: a few proposed models try to fullll the previous goals. Among them, the BSP-like CGM model seemed adapted to us to facilitate the way between algorithms design and real implementations. Many algorithms have been designed but few implementations have been carried out to demonstrate the practical relevance of this model. In this article, we propose to test this model actually on an irregular problem. We present the results of implementations of permutation graph algorithms written in two diierent models: the PRAM and the BSP-like CGM model. These implementation have been made on a CM5 and a PC cluster. We compare the results of these implementations with the performances of sequential code for this problem. With a classical problem in gaph theory, we validate BSP-like CGM model: it is possible to write portable code on two kinds of platform; this code is eecient, even on a PC-cluster, and handles very large data; real behaviour are in agreement with theory.