摘要
在单处理机系统中,由于计算高优先级任务抢占的时间相对比较简单,所以单处理机调度理论取得了长足的进步.提出一个端到端时间约束的实时任务调度算法,当实时任务到达系统时,算法为任务的每个子任务在相应的处理机上预约一定的计算资源,把端到端的多处理机调度问题转换成单处理机调度问题,从而可以利用单处理机调度理论判定实时任务的可调度性.实验表明,该算法明显地提高了CPU利用率和任务接收率.
Uniprocessor schedulability theory made great strides,in part,due to the simplicity of composing the delay of a task from the execution time of higher-priority tasks that preempt it. In this paper,a scheduling algorithm of real-time tasks with end-to-end time constraints is proposed. When a real-time task arrives at the system,the algorithm reserves computing time for each of its subtasks on the processor assigned,so that the multiprocessor schedulability has been transformed to uniprocessor schedulability. The wealth of schedulability analysis techniques derived for uniprocessors can then be applied to decide the schedulability of real-time task with end-to-end time constraints. The experimental results demonstrate that the CPU utilization ratio and tasks accept ratio increase greatly.
出处
《小型微型计算机系统》
CSCD
北大核心
2010年第2期239-242,共4页
Journal of Chinese Computer Systems
基金
因特尔:异构多可重构计算体系结构研究
关键词
调度算法
实时调度
端到端时间约束
EDF算法
scheduling algorithm
real time scheduling
end-to-end time constraints
EDF algorithm