任务驱动的众核资源调度技术研究

发布时间：2018-01-28 21:58

本文关键词： 复杂实时系统众核架构资源利用率任务调度效率　出处：《小型微型计算机系统》2014年06期 　论文类型：期刊论文

【摘要】：在采用众核架构的复杂实时系统中,对众核资源进行合理的调度,使得片内资源的分配与系统任务的需求相适应,对提高资源利用率和系统处理能力至关重要.为了使众核运算单元满足不同应用任务的计算需求,本文提出了一种基于众核架构的任务资源调度模型,该模型主要由应用任务集、实时计算管理逻辑和众核处理机三部分组成,基于该模型设计了众核资源调度算法MRSA(Many-core Resource-Schedule Algorithm).最后搭建了PowerPC主处理机和众核处理机的仿真实验环境,测试了系统的核资源利用率及任务调度效率,结果表明,MRSA算法能够有效适应众核架构系统的任务需求,提高系统资源的利用率.
[Abstract]:In complex real-time systems using multi-core architecture, reasonable scheduling on multi-core resources, makes the system task allocation and on-chip resources to adapt the needs of improving the utilization rate of resources and processing capacity of the system is very important. In order to make the nuclear unit meet the computational demands of different application tasks, this paper proposes a the task scheduling model based on multi-core architecture, the model is mainly composed of application tasks, real-time computing management logic and many core processor is composed of three parts, the model design of all nuclear resource scheduling algorithm based on MRSA (Many-core Resource-Schedule Algorithm). Finally, we build a simulation environment of PowerPC and the main processor core processor, testing system nuclear resource utilization and task scheduling efficiency, the results show that the MRSA algorithm can effectively adapt to many core system architecture requirements, improve the system of resources Utilization rate.

【作者单位】：西北工业大学计算机学院陕西省嵌入式系统技术重点实验室;第二炮兵工程大学信息工程系;
【基金】：国家自然科学基金项目(61103004)资助青年科学基金项目(F020308)资助
【分类号】：TP332
【正文快照】： 1引言随着集成电路工艺技术的不断进步,单芯片上可集成上百个甚至上千个处理器核(众核处理器).众核处理器计算资源密度更高,片上通信开销显著降低,更多处理器核可使应用的性能得到提升,因此众核处理器逐渐成为学术界研究的主要对象.在嵌入式实时系统中,对于计算复杂的应用任务

【参考文献】