摘要
当前处理器由于较高的能量消耗,导致处理器热量散发的提高及系统可靠性的降低,已经成为目前计算机领域较为关心的问题.然而目前一些有效降低能量消耗的技术大多针对单处理器系统,较少考虑多处理器系统.提出的调度算法针对多处理器计算环境,以执行时间最快的任务优先调度为基础,结合其它有效技术(共享空闲时间回收),使得实时任务在其截止期内完成的同时能够有效地减低整个系统的能量消耗.针对独立任务集及具有依赖关系的任务集,提出两种针对同构计算环境的算法:STFBA1(Shortest-Task-First-BasedAlgorithm)及STFBA2,及两钟针对多任务集的算法HSA1(HybridSchedulingAlgorithm)及HAS2.在单任务集计算环境下,与目前所知的有效算法相比,算法具有更好的性能(调度长度及能量消耗).在多任务集计算环境下,基于混合调度策略的算法能够明显改进调度性能.
At present the high power consumption of modern processors becomes a major concern due to the fact that it leads to increased heat dissipation and decreased reliability of systems. Many techniques have been proposed to reduce power consumption for uniprocessor systems, while less work have taken multiprocessor systems into account. The algorithms proposed in this paper are based on the strategy of least execution time first, focusing on multiprocessor systems and combining with other efficient technique（shared slack reclamation）, consequently, not only the real-time tasks can be completed before deadline but also the global power consumption of systems will be reduced. In this paper, we present two algorithms ： STFBA1 （Shortest Task-First-Based Algorithm） and STFBA2 to cope with independent task sets and task sets with precedence in homogeneous systems respectively. Moreover, we extend our strategy to multiple task sets, and present two algorithms：HSA1 （Hybrid Scheduling Algorithm） and HSA2 Compared to the efficient algorithm for single task set so far, our algorithms show much better scheduling performance in terms of makespan and power consumption. For multiple task sets, the algorithms based on hybrid strategy obviously improve the scheduling performance.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第5期866-872,共7页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60273075
60503048)资助
关键词
实时系统
多处理器系统
调度算法
能量消耗
real-time systems
multiprocessor systems
scheduling algorithm
power consumption
