摘要
近年来,随着多核芯片的产生和现代科学计算需求的不断增大,并行计算系统被广泛研究和使用,旨在提高多核芯片的利用率,提升系统的计算执行效率。该文针对多核平台间的并行计算系统,首先,介绍了并行计算系统的硬件组成结构,然后,针对系统的硬件结构,提出了相应的并行编程模型的基本编程接口和支持机制,并从3个角度,即并行性分解、任务划分和任务调度介绍并行计算系统的设计原理;最后总结并展望了并行计算系统未来的研究方向。
In recent years, with the development of multi-core chips and the increasing demand of modern scientific computing,Parallel computing systems have been widely studied and used in order to improve the utilization of multi-core chips, and to improve the computational efficiency of the system. In this paper, a parallel computing system is aimed at the multi core platform,First, the hardware structure of parallel computing system is introduced;Then, the basic programming interface and supporting mechanism of the parallel programming model are put forward;The design principles of parallel computing system are introduced from 3 aspects, namely, parallel decomposition, task partitioning and task scheduling; Finally, the future research directions of parallel computing systems are summarized and discussed.
出处
《电脑知识与技术》
2016年第2Z期244-246,共3页
Computer Knowledge and Technology
关键词
并行计算系统
并行编程模型
并行任务划分
并行任务调度
parallel computing system
parallel programming model
parallel task partitioning
parallel task scheduling
