State assignment for power optimization of sequential circuits based on a probabilistic pairwise swap search algorithm

摘要

State assignment (SA) for Finite State Machines (FSMs) has a significant impact on the area and power of synthesized sequential circuits. In this paper, we propose a novel probabilistic pairwise swap search (PPSS) state assignment algorithm. The algorithm is based on assigning probabilities for each pair of code swaps and intelligently updating these probabilities such that potentially useful code swaps will get higher chance of being explored. Experimental results demonstrate the effectiveness of the proposed algorithm by achieving significantly better power results in comparison to other evolutionary state assignment algorithms.