基于通用计算的GPU-CPU协作计算模式研究  

On the Basis of General-purpose Computing GPU-CPU Cooperative Computing Method Research

作　　者：崔雪冰[1,2] 张延红[1] 李国徽[2] 

机构地区：[1]河南城建学院计算机科学与工程系,河南平顶山467001 [2]华中科技大学计算机科学与技术学院,湖北武汉430074

出　　处：《微电子学与计算机》2009年第8期30-33,共4页Microelectronics & Computer

基　　金：国家自然科学基金项目(60873030)

摘　　要：依据GPU计算特点和任务划分的特点,提出一种类似主从模型的GPU-CPU协作计算的处理模式,通过把问题或算法划分成多子任务,并对划分的子任务给出合理的调度算法,使GPU和CPU各自发挥特点,从而发挥较高效率的GPU通用计算能力,通过测试验证该协作模式是有效的.According to the basis of features about GPU computing and tasks dividing, the article tries to bring forward a method of similar Master/Worker GPU-CPU cooperative computing. A logical scheduling algorithm is proposed to divide an issue or algorithm into more little tasks to fulfill the features of the GPU and CPU and General-purpose computing ability of the GPU . Finally, the Cooperative Computing Method is verified by experiment.

关 键 词：GPU通用计算 并行计算 协作计算模式 划分 任务调度 模型 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]