摘要
在软硬件协同设计中,将数据流图作为软硬件划分问题的输入,建立了以数据流图为基础的软硬件划分数学模型.对于规模较大的划分问题,为了满足系统性能,达到功耗和成本最小化,提出了0-1背包分解算法,采用并行求解的方式,排除了绝大多数无最优解的子问题,从而改善了大规模划分问题的求解效率.实验结果表明在大规模问题和硬件约束条件下,该算法能获得行之有效的划分方案.
In the software hardware co-design,a mathematical model based on data flow diagrams is established by using the data flow diagram as the input of the software hardware partition problem.For the large scale partition problem,a partition algorithm based on 0-1 knapsack is presented to meet the system performance and minimize power consumption and cost.It can exclude vast majority of sub-problems through parallel search and improve the productivity of the algorithm for the large scale problem.Our experiment results show that on the constraint of large scale problem and hardware,this algorithm can get effective partition scheme.
出处
《烟台大学学报（自然科学与工程版）》
CAS
北大核心
2011年第3期214-217,222,共5页
Journal of Yantai University(Natural Science and Engineering Edition)
基金
国家自然科学基金资助项目(60970088)
关键词
软硬件划分
0-1背包
数据流图
hardware-software partition
0-1 package
data flow diagram
