A genetic and configuration-prefetching scheduling based software/hardware partition algorithm
The automated software/hardware partition of application is an important issue of reconfigurable compiling research.Aimed at current methods not making full use of CPU and FPGA working in parallel,a genetic and configuration prefetching based scheduling software/hardware partition algorithm (GCPSPA) is presented.The algorithm combines applications partition and scheduling to take full advantage of CPU core and FPGAs parallel working,and by configuration prefetching to hide the hardware tasks configuration overhead.Experimental results show that compared with existing automated partition method,the proposed method can effectively reduce applications execution time.
software/hardware partition configuration-prefetching genetic algorithm
Jin Qu Rongcai Zhao Lin Han Dan Zhang Yuan Yao
National Digital Switching System Engineering and Technological Research Center,Zhengzhou,Henan,China
国际会议
秦皇岛
英文
401-404
2010-11-05(万方平台首次上网日期,不代表论文的发表时间)