Reliability Estimation of Approximate Circuits Based on Probabilistic Gate Model

The existing analytical methods for estimation of the reliability of approximate circuits are based on probability transfer module (PTM) method and Monte Carlo (MC) method. However, the PTM-based method is confined to small-scale approximate circuits and large circuits with weak signal correlation, and the MC method is time-consuming to obtain accurate results. This paper presents two methods for reliability estimation of gate-level approximate circuits based on probability gate model (PGM). The first is non-process correlation algorithm that does not consider the correlation among signals to obtain an approximate value of circuit reliability, and its time complexity keeps a linear relation with the number of gates. The second is the processing correlation algorithm which can estimate the correlation caused by fanout nodes of the approximate circuits, and it has the obvious advantage on accuracy, but its time complexity is exponential with the number of fanout nodes in the circuit. Experimental results demonstrate the validity and accuracy of the proposed methods.