基于任务级的动态电源管理设备调度策略

在嵌入式和便携式系统的低功耗设计中,动态电源管理(Dynamic Power Management,DPM)是一个非常重要的技术.然而,传统的DPM策略都仅仅建立在对目标设备的工作负载进行观测的基础上,从而带来了诸如无谓等待过程和预测错误等缺陷.本文则从任务级的角度出发提出了一种在线设备调度策略EODSA(Energy Optimal Device Scheduling Algorithm)及其改进策略IEODSA1(Improved EODSA1)、IEODSA2(Improved EODSA2).实验结果表明,本文所提出的三种策略都具有稳定性;而且和传统DPM策略相比,EODSA、IEODSA1、IEODSA2在满足对应的假设和前提条件下都能够显著降低外设的平均功耗.

Dynamic power management device scheduling policy on task level

Dynamic power management (DPM) is an important technique to reduce power consumption in embedded andportable systems. However, traditional DPM policies are based solely on the observation of requests at the target device andnone of these policies can make distinction on the request source, which generate unavoidable drawbacks, such asunnecessary waiting and error prediction etc. In this paper,an on-line device scheduling policy called EODSA (EnergyOptimal Device Scheduling Algorithm) is presented, and its improved policies IEODSA1, IEODSA2 are also suggested. Theexperimental results show that these policies have stability. Moreover, all of these policies can further reduce the averagepower consumption of peripheral with the corresponding assumption and premises condition compared to traditional DPMpolices.