Performance analysis of MPI collective operations

Previous studies of application usage show that the per-formance of collective communications are critical for high-performance computing and are often overlooked when compared to the point-to-point performance. In this paper, we analyze and attempt to improve intra-cluster col-lective communication in the context of the widely deployed MPI programming paradigm by extending accepted mod-els of point-to-point communication, such as Hockney, LogP/LogGP, and PLogP. The predictions from the mod-els were compared to the experimentally gathered data and our findings were used to optimize the implementa-tion of collective operations in the FT-MPI library. 1