一种异构NOWs上负载均衡的适应式并行调度算法  

Adaptive-partitioned parallel scheduling arithmetic based on equal load isomerous NOWs

作　　者：冯景超[1] 

机构地区：[1]西安邮电学院,陕西西安710061

出　　处：《计算机工程与设计》2004年第12期2291-2293,共3页Computer Engineering and Design

摘　　要：分析了并行作业调度算法,针对异构工作站群机系统(Network of Workstations,NOWs)提出了一种负载均衡的适应式划分(Adaptive Partition,AP)策略的并行调度算法,它利用了短作业需求优先(Shortest Demand First,SDF)的策略,主要思想是让系统中所有作业尽可能均分同时结束,以达到均衡负载。实验结果证明该算法公平和高效。A method of parallel job scheduling arithmetic is analyzed, and a new adaptive-partitioned parallel scheduling arithmetic based on equal load isomerous NOWs is brought forward. The policy of shortest demand first (SDF) is utilized, and its main idea is to make all of job finish at the same time, in order to arrive at equal load. It is proved that the arithmetic is correct and high efficient based on the result of experimentation.

关 键 词：NOW 并行调度 异构 负载均衡 算法 群机系统 SDF 作业 适应 证明 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP311[自动化与计算机技术—计算机科学与技术]