摘要
集成电路规模持续增长对数字波形的保存、打开以及传输带来困难,亟需设计高效且无损的数字波形压缩算法,以减少相关开销。针对数字波形压缩,提出了一种的基于同类数据重组增强算法。首先将所有波形数据按时钟域聚合重组,然后在各时钟域中按时间戳、信号ID和信号这3大类聚合重组,再通过预测编码、字典编码等技术对波形进行压缩。通过重组,极大降低了预测误差、增大字典匹配概率和匹配长度。引入并行策略,使得压缩和解压缩过程能够以多线程方式实现,提升速度。应用于典型数字波形的实验表明,所提方法的压缩率最大达到3275倍;相较于现有方法,文件压缩率最大提升近8倍。在多线程下相比现有方法能获得3倍以上的压缩加速和10倍以上的解压加速。该方法对于提高当前超大规模集成电路仿真能力有着重要意义。
The scale of integrated circuits continuously grow bringing challenges in saving,opening,and transmitting digital waveforms.There is an urgent need to design efficient and lossless digital waveform compression algorithms to reduce related costs.This paper proposes for the first time a novel algorithm based on similar data reconstruction enhancement.First,all waveform data are aggregated according to the clock domains.Then,within each clock domain,they are aggregated according to time stamp,signal ID,and signal.Finally,the waveform is compressed by predictive coding and dictionary coding etc.Through reconstruction,the prediction error is greatly reduced,the dictionary matching probability is increased,and the matching length is increased.Finally,by introducing parallel strategies,the compression and decompression processes can be implemented in a multithreaded manner to improve speed.Experimental results on typical digital waveforms show that the compression ratio of this method can reach up to 3275 times,which is nearly 8 times higher than existing methods.Also,it can achieve more than 3 times compression speedup and more than 10 times decompression speedup over existing methods under multi-threading manner.This technology is of great significance for improving the efficiency of present very large scale integrated circuit simulation.
作者
刘必慰
赵振宇
蒋汶乘
施嘉庚
周杰
LIU Biwei;ZHAO Zhenyu;JIANG Wencheng;SHI Jiageng;ZHOU Jie(Institution of Microelectronics and Microprocessor,School of Computer,National University of Defense Technology,Changsha 410073,China)
出处
《微电子学与计算机》
2024年第11期68-74,共7页
Microelectronics & Computer
基金
国家自然科学基金(重点)(62034005)。
关键词
数字波形
压缩
字典
重组
digital waveform
compression
dictionary
reorganization
