摘要：Rmpi is an R interface (wrapper) to the Message- Passing Interface (MPI). MPI is a standard application interface (API) governed by the MPI forum (http://www.mpi-forum.org) for running parallel applications. Applicable computing environments range from dedicated Beowulf PC clusters to parallel supercomputers such as IBM’s SP2. Performance and portability are the two main strengths of MPI. In this article, we demonstrate how the MPI API is implemented in Rmpi. In particular, we show how interactive R slaves are spawned and how we use them to do sophisticated MPI parallel programming beyond the “embarrassingly parallel”.
