A High Efficient Task Scheduling Algorithm Based on Heterogeneous Multi-Core Processor
Hui Cheng
Abstract
Hui Cheng
Abstract
Task scheduling is an important part of high performance multi-core building. The shortcomings of existing task scheduling algorithms is analyzed, and a new efficient heuristic task scheduling algorithm, namely,HCDDSL is proposed in this paper. Firstly, the new algorithm optimizes DAG graph by using clustering, then the nodes are descended by the values of Succ_sum,the task schedule has been processed in the case of insert interval and task duplication are considered. The weight of task in the DAG is considered in the process of schedule. By this means, the result of task schedule is optimized effectively,and the makespan of all tasks is reduced also.The proposed task scheduling algorithm has been analyzed through simulation, the results show that compare to old scheduling algorithm,this new algorithm can shorten the scheduling length of the task graph greatly.
OpenAlex reports 18 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Task scheduling is an important part of high performance multi-core building. The shortcomings of existing task scheduling algorithms is analyzed, and a new efficient heuristic task scheduling algorithm, namely,HCDDSL is proposed in this paper. Firstly, the new algorithm optimizes DAG graph by using clustering, then the nodes are descended by the values of Succ_sum,the task schedule has been processed in the case of insert interval and task duplication are considered. The weight of task in the DAG is considered in the process of schedule. By this means, the result of task schedule is optimized effectively,and the makespan of all tasks is reduced also.The proposed task scheduling algorithm has been analyzed through simulation, the results show that compare to old scheduling algorithm,this new algorithm can shorten the scheduling length of the task graph greatly.
Key concepts: Computer science, Fixed-priority pre-emptive scheduling, Dynamic priority scheduling, Fair-share scheduling, Rate-monotonic scheduling, Directed acyclic graph, Job shop scheduling, Two-level scheduling