华中科技大学学报（自然科学版）2026，Vol.54Issue(5)：149-156,8.DOI:10.13245/j.hust.250304
基于任务同步的异构多核实时任务划分算法
Task partitioning on heterogeneous multi-core real-time systems based on task synchronization
摘要
Abstract
To address the problem of simultaneously optimizing task synchronization and system energy consumption in heterogeneous multi-core environments,an algorithm named least loss energy density-synchronization task partitioning(LLED-STP)based on task synchronization was proposed.The algorithm prioritized the task with the greatest current impact on system energy consumption by analyzing and sorting the energy density differences of tasks across different cores.In the subsequent stage of assigning processor cores to tasks,the algorithm comprehensively considered both resource correlation and energy density difference to select the optimal core for each task,and the schedulability constraint was always ensured,thereby guaranteeing the overall schedulability of the system while reducing the total energy consumption of system scheduling.Experiment results show that under the experimental sets,the proposed algorithm has a schedulability ratio of 3％to 28％higher than other algorithms,and saves energy consumption of more than 13％.关键词
实时系统/异构多核处理器/任务同步/任务划分/节能调度Key words
real-time system/heterogeneous multi-core processor/task synchronization/task partitioning/energy-efficient scheduling分类
机械制造引用本文复制引用
赵小松,康玉龙,黄超,高杰,冯源..基于任务同步的异构多核实时任务划分算法[J].华中科技大学学报（自然科学版）,2026,54(5):149-156,8.基金项目
湖北省科技攻关计划资助项目(2023BAA018). （2023BAA018）
