Dynamic Topology Selection for High Performance MPI in the Grid Environments

MPI (Message Passing Interface) is getting more popular and important even in the Grid, but its performance still remains a problem, which is caused by the communication bottleneck on wide area links. To overcome such performance wall problem, we propose a dynamic topology selection which is a kind of resource selection method. It provides an effective resource selection service based on four principles which are derived from wide area message passing paradigm. Consequently, the proposed method can improve overall application performance by reducing the communication load. To demonstrate the proposed method, we executed parallel benchmark programs and measured each execution time in five geometrically distributed clusters by changing resource selection method. When using topology selection method, experimental results show that performance gain can be achieved by up to 200%.