与体系结构匹配的多级可重构任务划分方法

Multi_reconfigurable Task Partitioning Method on Architecture Matching

异构系统是高性能计算发展的主要模式,云计算是异构计算的典型实例。其优势在于异构处理器能各尽其能,但在实际应用中异构系统的性能往往不能充分发挥,因为处理器特征与应用程序特征不匹配,造成系统效率低下。因此借助重构思想,提出与体系结构结合的多级可重构任务划分方法。定义了多级可重构的概念,分析了异构匹配的原理,给出异构特征分析过程,提出了基于异构特征匹配的多级可重构任务划分方法。最后通过仿真实验说明,与体系结构匹配的划分方法适合当前的异构系统。

Heterogeneous systems have been the main model of high-performance computing currently, such as cloud platform. Its advantage is that heterogeneous system is feasible. Yet its performance cannot be fully utilized due to not matching between tasks executed and its architecture. So using reconfigurable ideas, the paper provided multi reconfigu- rable partitioning method combining architecture. Firstly the paper defined the multi-level reconfiguration concept. Sec- ondly the paper analysed heterogeneous matching theory and a process of heterogeneous feature, further proposed heter- ogeneous multi-level task partitioning method. And finally the paper verified the effectiveness of the method.