具有双电压调节处理器的最优DVS算法

为了解决嵌入式实时系统如何节省电能的问题,针对具有两种运行状态的微处理器提出了最优的动态电压调节(DV S)算法。对具有周期性硬实时任务集的嵌入式系统,引进了部分降频的新概念,建立了精确实用的数学模型,在此基础上证明了该算法的最优性。分别从实例实验、仿真试验两个方面,证实了该算法的正确性和有效性。实验结果表明,采用这种最优DV S算法的系统,比不采用DV S算法的系统明显地节省能量,具体减少的数值依赖于任务集的构成和处理器状态,但是减少量平均达到20%以上。和其他类似的调度算法相比,也能更好地减少能耗。

Optimal DVS algorithm for real-time systems with double voltage scalable processor

The power consumption in real-time embedded systems can be improved with an optimal dynamic voltage scaling(DVS) algorithm for micro-processors with double scalable voltages.Partial frequency scaling was used for the real-time periodic task sets.A mathematical model of the scheduling system was used to prove the optimality.Analyses and simulations further prove the effectiveness of the algorithm.The test results show that the DVS algorithm significantly reduces the power consumptions.Although the power savings depends greatly on the task sets and processor states,it averages 20% or more,more than previous DVS algorithms.