摘要
分析了多处理机系统中的集中式、分散式和全分布式这三种工作池的动态平衡的工作原理,通过仿真实验,提出了利用工作池技术,及时将工作池技术中待处理的工作任务分派给松耦合多处理机系统中处于空闲状态的处理器,使松耦合多处理机系统(尤其是异构机群平台)中各机器能够有机地协调工作,有效地提高海量数据的处理速度.
The paper analyzes the dynamic balance working principles of centralized work pool, decentralized work pool and distributed work pool in the multiprocessor system. Through the simulation experiment, it brings, forward a looselycoupled multiprocessor system using the work pool technology to assign the work tasks to the idle processors in a timely manner, which makes each machines of the loosely coupled multiprocessor system （especially for heterogeneous cluster platform） work coordinately. The speed of massive data processing is efficiently improved.
出处
《计算机系统应用》
2016年第7期278-281,共4页
Computer Systems & Applications
基金
福建省莆田市科技项目(2013G16)
福建省教育厅B类科技项目(JB14195)
关键词
海量数据
多处理机系统
工作池技术
massive data
multiprocessor systems
work pool technology
