一种FPGA配置数据压缩算法的参数选择方法  

A parameter selection method for FPGA configuration data compression algorithm

作　　者：杨鹏[1] 李仁发[1] 吴强[1] 

机构地区：[1]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《传感器与微系统》2011年第1期31-34,共4页Transducer and Microsystem Technologies

基　　金：国家"863"计划资助项目(2007AA01Z104)

摘　　要：随着现场可编程门阵列(FPGA)配置数据规模的显著增加,片外配置存储的低带宽成为了运行时可重构系统配置的瓶颈。考虑片外配置存储和片上配置接口对配置数据流传输速度的双重限制,建立了配置压缩系统的配置加速比模型,并实现了一种基于LZSS算法的压缩和解压系统。实验结果表明:配置加速比高的压缩算法相对压缩率高的算法能够达到更高的配置速度,因此,配置加速比作为评价配置压缩系统性能的指标比压缩率更具有指导意义。With the dramatic increasing in size of FPGA＇ s configuration data, the low bandwidth of off-chip nonvolatile configuration memory becomes a main bottleneck during the reconfiguration of run-time reconfiguration （RTR） system. A new model named configuration speedup ratio was built for RTR system by taking into account the speed limits on two sides which were configuration memory and on-chip configuration interface. Compression and decompression systems were constructed based on LZSS encoding. Experimental results showed a higher configuration data ratio of a RTR system when using an algorithm with high configuration speedup ratio rather than that with good compression ratio. Configuration speedup ratio is much more competitive than compression ratio for evaluating the performance of a configuration compression svstem.

关 键 词：现场可编程门阵列 运行时可重构 配置压缩 配置加速比 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]