摘要： 随着多处理器系统规模的不断扩大,如何节能成为一个亟待解决的重要问题。为此,基于多处理器系统提出一种针对随机任务的在线节能实时调度算法。使用统计方法,根据已有任务的到达时间和计算量估计新任务在空闲处理器上执行的电压/ 频率,使还未到达的任务能够满足截止期限并有效节能。在考虑单个处理器上执行的任务时,计算执行这些任务所需的平均电压/ 频率,使所有任务的执行速度尽量均衡,当某些任务不能满足截止期限要求时,则调高未执行任务的电压/ 频率。实验结果表明,与EDF,HVEA,MEG 和ME-MC 算法相比,该算法在满足截止期限和节能方面具有明显的优势。

关键词: 多处理器系统, 随机任务, 动态电压/ 频率调整, 在线, 实时, 节能调度

Abstract: With the continuous expansion of the scale of the multi processor system,the issue of energy consumption becomes more and more important. How to save energy becomes an important problem to be solved. Based on the multiprocessor system,On-line Energy-efficient Real-time Scheduling Algorithm(OERSA) aiming at a random task is proposed. According to the arrival time and calculated amount of the existing task,the algorithm estimates the executive voltage / frequency of the new task in the idle processor by using statistical methods,which can meet the deadline and save the energy effectively for not yet arrived tasks. At the same time,considering the task executed on a single processor,the algorithm first calculates the average voltage / frequency required to perform these tasks,thus making all the task execution speed equal as much as possible. When some tasks can not meet the deadline requirements,voltage / frequency for previous not executed tasks will be adjusted high. Experimental results show that OERSA has obvious advantages in the aspect of meeting deadlines and energy consumption saving compared with EDF,HVEA,MEG and ME-MC algorithm.

Key words: multiprocessor system, random task, dynamic voltage / frequency scaling, on-line, real-time, energy-efficient scheduling

中图分类号: 

• TP316. 4