康鹏, 刘从新, 沈绪榜. 一种基于分组的多核嵌入式实时调度算法[J]. 微电子学与计算机, 2016, 33(10): 32-35.
引用本文: 康鹏, 刘从新, 沈绪榜. 一种基于分组的多核嵌入式实时调度算法[J]. 微电子学与计算机, 2016, 33(10): 32-35.
KANG Peng, LIU Cong-Xin, SHEN Xu-Bang. Multicore Embedded Real-Time Scheduling Algorithm Based on Gang Scheduling[J]. Microelectronics & Computer, 2016, 33(10): 32-35.
Citation: KANG Peng, LIU Cong-Xin, SHEN Xu-Bang. Multicore Embedded Real-Time Scheduling Algorithm Based on Gang Scheduling[J]. Microelectronics & Computer, 2016, 33(10): 32-35.

一种基于分组的多核嵌入式实时调度算法

Multicore Embedded Real-Time Scheduling Algorithm Based on Gang Scheduling

  • 摘要: 针对亲和多核调度方法中容易出现的负载不均衡问题, 在原有多核亲和调度算法中, 提出了一种新的调度算法.该算法在每个核上增加统计任务对CPU进行实时监控, 同时在系统任务中增加全局负载均衡任务进行第三级调度, 相对于已有的亲和调度算法, 在保证原有操作操作系统不被破坏的基础上, 操作系统核间负载更加均衡.通过组内亲和任务融合, 来减小系统功耗, 满足了操作系统中任务调度的实时性及负载均衡的需要.

     

    Abstract: Aimed at solving the unbalanced problem caused by the Real-time task scheduling on multi-core, this paper proposed an advanced scheduling based on affinity scheduling. This algorithm monitors CPU by increasing a statistical task. Meanwhile, it can process the third scheduling by adding global payload task. Compared with the former scheduling, this method guarantees the balance of every CPU and the operation system undamaged. Real-time task scheduling and payload balance can be satisfied by mixing the different CPU unites and reducing the system power consumption.

     

/

返回文章
返回