并行系统的通讯效率问题

巨量并行处理（ＭＰＰ）强调并行系统结构和并行算法的可扩放性。在一个可扩放的并行系统结构上，可扩放的并行算法应该能够有效地利用不断增加的处理机，算法的有效性通常以算法运行时的处理机效率来衡量。一个被普遍忽视的因素是通讯效率，这是一个具有一般性的问题。本文给出了通讯效率的定义，研究了它与处理机效率的关系，并通过对一个典型算法的运行情况分析，研究了几个常见的并行系统结构的通讯效率。本文的结果表明：处理机效率和通讯效率的综合才能全面地评价算法的可扩放性并指导并行系统结构的设计。

On the Communication Efficiency of Parallel Systems

The study of massively parallel processing emphasizes the scalability of parallel algorithms and architectures,which is a measure of the capability of the algorithm to effectively utilize an increasing number of the processors in the scalable architecture.In scalability analysis.the validity of the algorithms is usually measured with processor efficiency although it has been noticed that the processor efficiency does not reflect the cost effectiveness of the architectures to run the algorithms.This paper defines the communication efficiency,which is directly related to the cost efficiency,and studies the relationship between the communication efficiency and the processor efficiency when they are applied to scalability analysis.An example of algorithm is given to analyze several architectures.The problem of communication efficiency is a general issue in parallel processing.Further research is necessary.