硬实时环境下固定优先级调度的离线优化  

Off-Line Optimization of Fixed Priority Scheduling in Hard Real-Time Environment

作　　者：罗玎玎[1] 赵海[1] 孙佩刚[1] 林恺[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《东北大学学报（自然科学版）》2008年第9期1257-1261,共5页Journal of Northeastern University(Natural Science)

基　　金：国家火炬计划项目(2002EB010154)

摘　　要：RM调度算法具有简单的实现机制和较低的调度开销,被广泛应用于硬实时调度领域.然而这类算法的固定优先级特征使其在高任务负载环境下具有极高的抢占次数,从而导致了较大的系统开销,因此提出一种方法来减少RM调度的抢占次数.该方法通过离线计算任务集的最优属性来减少基于RM调度的系统在运行时的抢占次数,进而降低系统的抢占开销.仿真结果表明,该方法可以在不付出额外调度开销的前提下有效减少RM调度的抢占次数,降低实时系统的抢占开销.The RM algorithm has widely been used in hard real-time environment,mainly due to its simple implementation mechanism and low scheduling overheads.However,the heavy system overheads resulting from the highly frequent occurrences of preemption under heavy workload are characterized by the fixed priority scheduling of the RM algorithm.So,a method is proposed to reduce the occurrences of preemption in running due to the scheduling by RM algorithm via optimizing the attribute by off-line computation of task set,thus reducing the relevant overheads. Simulation results show that the occurrences of preemption and relevant overheads in real-time systems can be reduced efficiently by the method proposed without extra scheduling overheads.

关 键 词：实时系统 硬实时调度 固定优先级 RM算法 抢占开销 

分 类 号：TP316.2[自动化与计算机技术—计算机软件与理论]