作者: S.M. Boxer , B.G. Batchelor
关键词:
摘要: A linear array of microprocessors provides a powerful computing system that is particularly well suited to many pattern-recognition and cluster-analysis algorithms. These often rely heavily upon the calculation distances in high-dimensional vector spaces: can be computed at high speed by an identical processing elements, operating parallel under command central controller. To achieve speeds those pattern recognition algorithms which refer input each member set stored reference vectors, elements should contain some `local? storage. Of course, not all are parallel, accomodate these, may required operate autonomously. Nevertheless, controller must, times, able force entire its control again. The third mode, namely acting as pipe-line processor, useful situations (e.g. polynomials) for transferring data between array's local store rectangular even faster than one, but is, more expensive. cost performance Intel 8080 compared other systems.