ABSTRACT
Since the early 1980s, there has been signicant progress in computer hardware design. Besides designing faster computer chips, computer vendors have introduced vector and parallel hardware, which can increase the computer performance (GFLOPS:Giga FLoating point OPerations per Second) by orders of magnitudes. The performance improvement due to vector or parallel processing is highly dependent on the algorithm, i.e., how well it suits and/or how well it is designed for vector/parallel architectures.
