Transformation of systolic algorithms for interleaving partitions

Abstract

A systematic method to map systolizable problems onto multicomputers is presented in this paper. A systolizable problem is a problem for which it is possible to design a Systolic Algorithm. This method selects and transforms the Systolic Algorithm into a parallel algorithm with high granularity. The communications requirements are reduced and thus the performance can be increased. The proposed scheme requires a classification of dependences, and it is based in the interleaved execution of several partitions of the Systolic Algorithm. The code to be executed in a processing element of the multicomputer system is obtained through application of the proposed systematic transformations to the original sequential code. By applying this method to the APP we illustrate their main features, and several performance measures for a torus of transputers system are presented, considering the various algorithms which are unified by the APP.