Polynomial Evaluation Using Parallel Processing and Specialized Hardware for Scientific Applications

摘要

An algorithm for fast evaluation of polynomials using parallel processing is presented. Feasibility demonstrations at various stages are described, and a timetable for completion within approximately one and one-half years is presented. Implementation of the algorithm will result in increases in speed of polynomial evaluation by approximately a factor of two over conventional linear techniques. By maintaining a modular design, interface to various minicomputer systems and program environments is facilitated. (ERA citation 05:011017)