A static performance estimator in the Fortran D programming system

Abstract The choice of the data decomposition scheme is an important factor in determining the available parallelism and hence the performance of an application on a distributed memory multiprocessor. In this paper, we present a performance estimator and discuss its role in the interactive Fortran D programming system. The estimator statically evaluates the relative efficiency of different data decomposition schemes for any given program on any given distributed memory multiprocessor. Our method is not based on a theoretical machine model, but instead uses a set of kernel routines to “train” the estimator for each target machine. We also describe a prototype implementation of this technique and discuss an experimental evaluation of its accuracy.