Multicomputers have the potential to deliver gifaflop performance on many scientific applications. Initial implementations of parallel programs on these machines, however, are often inefficeint and reruire signitifant oprimization before they can harness the potential powes of the machine. Performance prediction tools can provide valuable information on which optimizations will result in increased perfoemance. This paper describes an analytical performance. This paper describes an analytical performance data compilers, programmers and sysem architects to assist them in making choices which will lead to more efficient implementations. Efficient performance prediction tools can provide information which will help programmer's make better use of the power of multicomputers