期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：1988
卷号：XXVII Part B2
页码：465-478
出版社：Copernicus Publications
摘要：High rate data processing is extremely challenging since it requires fast computation,large data storage, and fast data access simultaneously. In order to meet these challenges,computer technology has evolved toward concurrent architectures where multiple processingelements are integrated for parallel execution of data. Processors may communicate through ashared memory system or may communicate directly to neighboring processors via a communicationnetwork. The performance characteristics of a system depend on the processorconnection topology, memory organization, and synchronization mechanism of the system.Implementation of a data processing algorithm in a concurrent system requires decompositionand/or reorganization of the algorithm according to the characteristics of the concurrentsystem. It is critical to employ a concurrent system whose systematic characteristics are compatiblewith algorithmic properties in order to minimize the decomposition/reorganizationeffort. This paper presents a methodology of analyzing the algorithmic properties with respectto their implementational constraints in a concurrent architecture
