基于平衡因子的动态偶发任务低功耗调度算法
发布时间：2018-12-13 02:56
【摘要】：为了更好地降低实时系统的功耗,本文以偶发任务为调度对象,提出了一种基于平衡因子的动态偶发任务低能耗调度算法(LP-DSAFST)。LP-DSAFST算法分为两个阶段,第一阶段根据释放任务的负载调整处理器速度,充分利用处理器的资源;第二阶段通过高优先级任务提前完成剩余的空闲时间动态更新处理器速度,并且当处理器没有任务执行时,根据平衡因子关键速度调度策略,判断是否使用DPM技术将处理器进入休眠状态以达到进一步降低能耗的目的。实验表明,新算法与DVSST和DSTLPSA算法相比节能效果更好,可以分别节约10.3%~62.5%和5.4%~20.6%的能耗。
[Abstract]:In order to better reduce the power consumption of the real-time system, this paper presents a dynamic contingency task low-energy-consumption scheduling algorithm (LP-DSAFST) based on a balance-factor. The LP-DSAFST algorithm is divided into two stages, and the first stage adjusts the processor speed according to the load of the releasing task. fully utilizing the resources of the processor; the second stage dynamically updating the processor speed by the high-priority task ahead of time, and scheduling the policy according to the key speed of the balance factor when the processor does not have a task to execute, it is judged whether to use the dpm technology to put the processor into the sleep state to achieve the purpose of further reducing the energy consumption. The results show that the new algorithm has better energy-saving effect than DVSST and DSTLPSA, and can save the energy consumption of 10.3% ~ 62.5% and 5.4% ~ 20.6%, respectively.
【作者单位】： 中国科学院沈阳计算技术研究所;中国科学院大学;华侨大学计算机科学与技术学院;
【基金】：国家科技重大专项项目(2013ZX04007-011) 辽宁省博士启动基金项目(201501040)
【分类号】：TP301.6
，
本文编号：2375748
[Abstract]:In order to better reduce the power consumption of the real-time system, this paper presents a dynamic contingency task low-energy-consumption scheduling algorithm (LP-DSAFST) based on a balance-factor. The LP-DSAFST algorithm is divided into two stages, and the first stage adjusts the processor speed according to the load of the releasing task. fully utilizing the resources of the processor; the second stage dynamically updating the processor speed by the high-priority task ahead of time, and scheduling the policy according to the key speed of the balance factor when the processor does not have a task to execute, it is judged whether to use the dpm technology to put the processor into the sleep state to achieve the purpose of further reducing the energy consumption. The results show that the new algorithm has better energy-saving effect than DVSST and DSTLPSA, and can save the energy consumption of 10.3% ~ 62.5% and 5.4% ~ 20.6%, respectively.
【作者单位】： 中国科学院沈阳计算技术研究所;中国科学院大学;华侨大学计算机科学与技术学院;
【基金】：国家科技重大专项项目(2013ZX04007-011) 辽宁省博士启动基金项目(201501040)
【分类号】：TP301.6
，
本文编号：2375748
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2375748.html
