多级混合极性Reed-Muller逻辑电路功耗优化
Power Optimization of Multi-Level MPRM Logic Circuits
-
摘要: 针对多级混合极性Reed-Muller(MPRM)电路功耗优化问题, 提出一种基于列表法和onset法的多级MPRM功耗优化算法. 对于电路的各个割集, 结合列表法和多输入AND/XOR门低功耗分解模型, 遍历所有极性下的二级MPRM最优功耗函数; 然后在二级MPRM函数的基础上, 通过onset法得到多级MPRM函数, 对其中的多输入AND/XOR门进行低功耗分解. MCNC和EPFL测试集的实验结果表明, 与原电路相比, 所提算法的平均功耗优化率达到49.90%和27.87%; 与二级MPRM功耗优化算法相比, 该算法的平均面积优化率为20.52%, 平均功耗优化率达到21.24%.Abstract: In this paper, multi-level mixed-polarity Reed-Muller (MPRM) power optimization is addressed by tabular technique and onset table methods. The low-power model of multi-input AND/XOR gates and tabular technique enable us to traverse the optimal power circuit of the two-level MPRM under all polarities for each cut set of the circuit. The onset table is then used to optimize multi-level MPRM based on two-level MPRM circuits. Based on our experimental results on the MCNC and EPFL benchmark suites, our algorithm improved power by 49.90% and 27.87% respectively, compared to the original circuit. Compared to the two-levels MPRM power optimization algorithm, the average area optimization rate of our algorithm is 20.52%, and the average power optimization rate is 21.24%.