An Effective Algorithm for Average Power Estimation of CMOS Sequential Chircuit  

作　　者：LIYueping TANGPushan ZHAOWenqing 

机构地区：[1]CADLaboratory,DepartmentofElectronicEngineering,FudanUniversity,Shanghai200433,China

出　　处：《Chinese Journal of Electronics》2003年第1期65-70,共6页电子学报（英文版）

摘　　要：An incremental probabilistic algorithm is proposed for estimating average power of CMOS sequential circuit.We facilitate the flrst-order Taylor expansion to consider the spatial and temporal correlation among the internal nodes of the seauential circuits.Regarding finite state machines as non-decomposable and aperiodic Markov Chains,the steady-state probabilities exist.Consequently there have the steady probabilities of state lines.Thus the signal probability and switching activity of state line can be gotten through Picard-Peano iteration method.Sequential modules are separated from the whole circuit to shorten the runtime of our algorithm.We unroll the sequential module to accurately estimate the signal probability of state lines.Unilke the algorithms bassed on global BDD,the runtime of computing signal probability and switching activity of our algorithm does not depend on the circuit size.Experimental results show that our algorithm is much faster than the Monte-Carlo simulation method with the error below 10%.

关 键 词：CMOS时序电路 马尔可夫链 布尔近似法 平均功率估计 

分 类 号：TN432[电子电信—微电子学与固体电子学]