Reliable area index: A novel approach to measure reliability of Markov Random Field based circuits

Markov Random Field (MRF) is a probabilistic circuit design approach that transforms simple CMOS to MRF-CMOS digital circuits. The transformed circuits are proved to be extremely noise-tolerant in the previous research literature. In this paper, we have quantified the noise tolerance capability of MRF circuits as compared to CMOS counterparts. The tradeoff for this noise-immunity is the increase in transistor-count of the circuit. This tradeoff is modelled by a novel factor in this research work called as `reliable area index'. The results show that the value of this index is 30 times higher for MRF than CMOS which proves that the MRF design approach still maintains a suitable tradeoff between noise-tolerance and area overhead of the circuit.