论文部分内容阅读
为了提高在信息论基础上的逻辑电路面积和功耗分析方法的估计精度,在传统输出信息熵估计理论基础上,提出了利用翻转信息熵进行电路实现复杂度和面积估计的理论方法。概率的方法被用于对组合逻辑电路的输入输出信号翻转行为的相关和相似进行量化分析。在此基础上实现了翻转信息熵面积估计算法,对随机生成的大量电路和标准benchmark电路进行的实验结果表明,该方法带来了至少3%左右的估计精度改善。
In order to improve the estimation accuracy of logic circuit area and power consumption analysis method based on information theory, a theoretical method of circuit complexity and area estimation using flip information entropy is proposed based on the traditional output information entropy estimation theory. The method of probability is used to quantitatively analyze the correlation and similarity of the input-output signal flip behavior of the combinational logic circuit. On this basis, the algorithm of flip entropy area estimation is implemented. Experimental results on a large number of randomly generated circuits and standard benchmark circuits show that the proposed method can improve the estimation accuracy by at least 3%.