基于时间帧的处理器PFair调度改进算法  

PFair processor scheduling algorithm based on time frame

作　　者：蒋丽丽[1] 陈国彬[1] 张广泉[2,3] 

机构地区：[1]重庆工商大学融智学院大数据应用研究所,重庆400030 [2]苏州大学计算机科学与技术学院,江苏苏州215006 [3]中国科学院计算机科学国家重点实验室,北京100080

出　　处：《计算机应用研究》2016年第4期1058-1061,共4页Application Research of Computers

基　　金：重庆市教委科学技术研究项目(KJ133103);江苏省自然科学基金资助项目(BK2011152);中国科学院计算机科学国家重点实验室开放课题(CSYSKF0908)

摘　　要：为了解决PFair算法进行交互任务调度时,由于忽略了不同阶段的周期性任务而导致多个线程之间任务的迁移问题以及空间和时间的浪费问题,提出了基于时间帧的处理器PFair调度改进算法。该算法基于周期性任务系统的特点,引入时间帧控制和改变本地周期性任务调度来限制任务迁移,从而实现对PFair算法的改进。为了评估算法的迁移开销和公平性,通过实验对普通PFair算法及所提出的改进算法ERfair进行对比实验,结果表明,改进算法ERfair能够通过时间帧内调度和分区控制大大降低任务在不同处理器间的迁移次数。基于时间帧的处理器PFair调度改进算法在保证公平性的同时提高了系统效率,应用于多核处理器上的任务调度是可行的、有效的。When the PFair algorithm makes task scheduling,it may ignore the periodic tasks of different phases so that there will be problems of task migration among multiple threads and the waster of space and time. In view of the above problems,this paper put forward an improved PFair scheduling algorithm for the processor based on time frame. Based on the characteristics of periodic task system,the algorithm,introduced the time frame to control and changed the local periodic task scheduling and to limit task migration,so as to realize the improvement of the PFair algorithm. To assess the migration expense and fairness of the algorithm,it carried out comparison experiment between the ordinary PFair algorithm and improved ERfair algorithm proposed. The results show that,the improved ERfair algorithm can greatly reduce the times of task migration between different processors through scheduling in the time frame and partition control. The PFair scheduling algorithm for processor based on time frame can guarantee fairness,and also improve the efficiency of the system. It is feasible and effective when apply to the scheduling of multi-core processors.

关 键 词：多处理器 实时系统 任务迁移 任务调度 

分 类 号：TP302.1[自动化与计算机技术—计算机系统结构]