基于异构多核处理器的静态任务调度研究

Research on static task scheduling based on heterogeneous multi-core processors

针对现存任务调度算法优先级选取过于单一、冗余任务处理较晚的问题,提出一种基于加权优先级的任务调度算法——WPTS算法。该算法综合考虑任务3个属性的加权值以决定任务被处理的先后次序,从而克服了任务选取时的单一性问题。在将任务分配到处理器的过程中,保证任务优先调度到完成时间最早的处理器上。同时,引入冗余任务处理过程,及时消除冗余任务,达到对处理器空闲时间段进行有效回收、减少处理器调度长度的效果。性能对比实验表明,WPTS算法较CPFD算法、HCPFD算法和HDEFT算法能取得更好的性能。

Aiming at the problems that singleness of priority selection and later dispose of redundant task existed in, a task scheduling algorithm based on the weighted priority named WPTS is proposed. To overcome the shortcomings of the singleness, this new algorithm takes the method of adding weighted value of the three task properties value to determine the priority of the tasks to be handled. In the phase of mapping the tasks to the processors, the task will be scheduled to the processor with the minimum execution time. Meanwhile, a redundant task elimination process is introduced to make full use of idle processors and reduce the schedule length as well as eliminate the redundant tasks timely. In the phase of mapping the task to the processors, make sure that schedule tasks to the processor which can minimum the completion time, the comparison experiments show that WPTS algorithm can achieve better performance than CPFD, HCPFD and HDEFT algorithms.