基于翻转信息熵的组合逻辑电路功耗分析  

Power Analysis of Combinational Circuits Based on Transition Entropy

作　　者：张盛[1] 周润德[1] 羊性滋[1] 

机构地区：[1]清华大学微电子学研究所,北京100084

出　　处：《电子学报》2004年第8期1256-1259,共4页Acta Electronica Sinica

基　　金：国家自然科学基金 (No 59995550 1 ) ;清华大学 985关键研究基金

摘　　要：基于信息熵的复杂度分析方法是在VLSI设计流程的高层次抽象阶段对组合逻辑电路功耗和面积进行分析估计的可行方法之一 .本文通过提出新的利用翻转信息熵进行电路实现面积和功耗估计的理论方法 ,改善了面积和功耗估计精度 .大量基于随机电路和BENCHMARK电路的实验结果表明 ,上述方法能够使面积和功耗估计的相对误差分别降低 2 4 3% (从 12 74 %到 9 6 5 % )和 15 4 % (从 13 6 7%到 11 5 7% ) .For control circuits and random logic, signal complexity evaluated by information entropy (or informational energy) had been calculated to estimate the power consumption of circuits in the past research in the high-level abstract step of VLSI design flow. This paper improved the estimation accuracy for power and area costs of combinational circuits at the behavioral design level, with transition information entropy. Lots of experiments had been implemented to prove the improved estimation accuracy compared to the traditional analytic method. The experimental results with BENCHMARK circuits show that relative errors of area and power estimation have been reduced 24.3% (12.74% to 9.65%) and 15.4% (13.67% to 11.57%).

关 键 词：组合电路 低功耗设计 信息熵 

分 类 号：TP331[自动化与计算机技术—计算机系统结构] TN431[自动化与计算机技术—计算机科学与技术]