Abstract: In order to enhance the execution efficient of thread scheduling and parallel performance in Chip Multi-processor(CMP), a kind of scheduling method of heuristic Particle Swarm Optimization(PSO) algorithm is proposed. This algorithm is based on the module of designed thread scheduling module and uses Direct Acyclic Graph(DAG) chart to express the dependence between threads. Meanwhile, it uses the new method for thread scheduling. Experimental result shows that the execution efficient of this method is prior to the genetic algorithm and can better reduce the task execution time and take full advantages of benefit of the multi-core structure.

Key words: Chip Multi-processor(CMP), thread scheduling, Particle Swarm Optimization(PSO) algorithm, global optimum, local optimum, Direct Acyclic Graph(DAG), scheduling method

摘要： 为提高片上多核处理器(CMP)架构中线程调度的执行效率，发挥CMP的并行性能，提出一种基于混合粒子群优化算法的线程调度方法。根据设计的线程调度模型，利用有向无环图表述线程及线程间的相互依赖关系，并采用改进的混合粒子群算法对其进行合理调度。实验结果表明，该方法的执行效率优于现有的遗传算法，能有效地降低任务的执行时间，充分发挥多核架构的优势。

关键词: 片上多核处理器, 线程调度, 粒子群优化算法, 全局最优, 局部最优, 有向无环图, 调度方法

CLC Number: 

• TP273