Author(s)

M.P. Bekakos

Abstract

The principal theme herein is the performance evaluation of a fast direct hardware implementation of an alternate matrix procedure for the solution of linear systems Ax=b, where A is a compact dense (nxn) matrix, on a simu- lated special-purpose shared memory machine utilizing the dewavefront con- cept (rf. Evans[5]]). The simulated DEWavefront Array Processor Machine (DEWAP) consists of a square mesh of processors on which, and in distinct stages, the factorization and the solution phases of the method are performed. 1 Introduction The introduction of VLSI circuit technology has offered the means for processing very large scale data on multiprocessor arrays in a highly parallel manner. The efficient utilization, however, of such an advanced technology in large processor array structur

Keywords