Improving the performance of optimum parallel algorithms on SIMD array processors: Programming techniques and methods

Abstract

In this paper we describe parallel programming techniques and methods applied to the implementation of two parallel algorithms on a SIMD array processor. These programming techniques and methods can be employed to parallel numerical algorithms for improving their performance efficiency. First were implemented the column sweep algorithm to solve a lower triangular system of linear equations and then we implemented an algorithm based on Given's rotations for reducing a rectangular matrix into lower triangular form. The estimated execution time of all algorithms has been obtained and analyzed.