基于概率表达式的MPRM电路功耗计算方法

Probability Expression Based Power Estimation Method for MPRM Circuits

采用基于信号概率的功耗计算模型进行MPRM(Mixed Polarity Reed-Muller)电路功耗优化,信号概率计算是功耗计算的关键.提出一种基于概率表达式的MPRM电路功耗计算方法.该方法兼顾信号概率计算的时间效率和准确性,对MPRM电路中不存在空间相关性的信号通过在电路中传播信号概率的方式计算其信号概率,存在空间相关性的信号则利用概率表达式计算其信号概率,并在电路中传播概率表达式以解决空间相关性问题,在此基础之上根据基于信号概率建立的解析动态功耗和静态功耗计算模型计算电路功耗.为进一步提高时间效率,该方法采用二元矩图表示概率表达式.使用基准电路对所提出方法进行了验证,并与其他采用不同信号概率计算方法的MPRM电路功耗计算方法进行了比较.结果表明所提出方法准确有效.

Signal probability calculation is the key to power estimation when optimizing power of MPRM(Mixed Polarity Reed-Muller)circuits by using signal probability based power estimation models.A probability expression based power estimation method is proposed for MPRM circuits.The proposed method takes into account both efficiency and accuracy of signal probability calculation,for signals having not spatial correlation in MPRM circuit,their signal probabilities are computed by means of signal probability propagation in circuit,whereas for signals having spatial correlation,probability expressions are utilized to calculate their signal probabilities and are propagated in circuit to resolve spatial correlation problem,then the dynamic and static power of the circuit are computed respectively by using the established analytical power estimation models based on signal probability.In order to further improve time efficiency,the proposed method utilizes binary moment diagram to represent probability expression.The proposed method is validated by using several benchmark circuits,and compared to other power estimation methods using different signal probability calculation methods for MPRM circuits.Results show that the proposed method is accurate and effective.