Research on Hierarchical Multi-core Scheduling Algorithm Based on Task Replication

International Journal of Advanced Network, Monitoring, and Controls

article

Yin Haijing¹  Wang Jianguo¹  Huang Shujuan¹ 

[1] School of Computer Science and Engineering Xi’an Technological University Xi’an

关键词: Load Balancing;    Task Scheduling;    Task Duplication;    Hierarchical Scheduling;   

DOI  :  10.21307/ijanmc-2021-038

学科分类：社会科学、人文和艺术（综合）

来源: Asociación Regional De Diálisis Y Trasplantes Renales

【 摘 要 】

The rapid development of multi-core systems makes task scheduling in multi-core systems a new research topic. While tasks are running in parallel, how to improve the efficiency of the system and maintain the load balance of the system is the focus of research in the new era. Aiming at the problem that the multi-core scheduling algorithm based on task duplication does not consider the load balance of each CPU, which leads to the problem of reduced CPU utilization. This paper combines a hierarchical idea on the basis of task replication, and proposes a new hierarchical multi-core scheduling algorithm TDLS algorithm based on task replication. This algorithm is based on the idea of hierarchical scheduling. According to the fact that there is no dependency relationship between tasks at the same layer after layering, the task scheduling sequence is adjusted to reduce the waste on the core, shorten the waste between cores caused by communication time, and reduce the number of processors. , Thereby greatly improving the CPU utilization rate, using the least time and the least number of cores to complete scheduling, making the load of multi-core scheduling more balanced. Experiments show that under the same experimental conditions, compared with the traditional multi-core scheduling algorithm based on task replication, the improved algorithm TDLS reduces the number of processor cores, and also shortens the scheduling length of the total task. Its performance is better than the traditional multi-core scheduling algorithm based on task replication.

【 授权许可】

CC BY-NC-ND