摘要： 提出一种主从调度模型在Linux下的实现方法。在该模型中，系统中的每一个进 程由某一个指定的CPU来运行，从而减少了进程切换的次数。在内核中，每一个CPU单独拥有 一个可运行队列，各个CPU各自进行调度，使得内核调度线程能够同时在多个CPU上运行，多 个CPU充分并行，从而提高系统性能。

关键词: 操作系统, 主从调度, Linux

Abstract: This paper presents the implementation of Master/Slave scheduling mo del in Linux. In Master/Slave system, every process is dedicated to one CPU, so times of process switching is decreased. In OS kernel, every CPU having a runque ue of itself, the schedule thread can run on two or more CPUs simultaneously, so parallelity and performance of operating system are improved.

中图分类号: 

• TP316.81