一种多核处理器中混合关键级任务半分区调度算法  

Hybrid Critical-level Task Half-partition Scheduling Algorithm in Multi-core Processors

作　　者：李俊何 杨康 张凤登[1] LI Junhe;YANG Kang;ZHANG Fengdeng(School of Optical-Electrical&Computer Engineering,University of Shanghai for Science&Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《小型微型计算机系统》2024年第5期1264-1271,共8页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(71840003,61773083)资助;上海市自然科学基金项目(15ZR1429300)资助。

摘　　要：目前用于混合关键级系统的任务调度算法存在资源分配不合理、中低关键级任务的运行受限等情况,造成系统整体服务质量被降低.本文提出了有效改进系统服务质量的半分区调度算法,该调度算法将用于多核处理器的空闲资源回收计算,能够在触发了高关键级模式的情况下,为被抛弃的低关键级任务分配空闲时间片.此外,该算法还可以利用系统中的空闲时间片推迟系统关键等级的提升,并提出系统关键等级平稳回落策略,两者可以确保系统关键等级不会过于频繁变化.最后,通过算法仿真和实验证明了本文所提方法在混合关键级任务调度问题上的有效性和优越性.The current task scheduling algorithm for hybrid critical-level systems suffers from unreasonable resource allocation and restricted operation of low and medium critical-level tasks,resulting in a degradation of the overall quality of service of the system.In this paper,we propose a half-partition scheduling algorithm that can effectively improve the quality of service of the system.This scheduling algorithm will be used for the idle resource recovery calculation of multi-core processors and can allocate idle time slices for discarded low-critical-level tasks when a high-critical-level mode is triggered.In addition,the algorithm can also make use of the free time slices in the system to postpone the system criticality level increase and propose a smooth system criticality level fallback strategy,both of which can ensure that the system criticality level does not change too frequently.Finally,the effectiveness and superiority of the proposed method in the hybrid critical level task scheduling problem is demonstrated through algorithm simulation and experiments.

关 键 词：混合关键级系统 动态需求边界函数 半分区划分算法 多核处理器 任务调度 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]