High-Order Temporal Effects in Finite-State Machine Analysis

The objective of this paper is to illustrate, analytically and quantitatively, the effect of high-order temporal correlations on steady-state and transition probability calculations in Finite State Machines (FSMs) analysis. As the main theoretical contribution, it is shown that, if the sequence feeding the target circuit has temporal correlations of order k, then a lag-k Markov chain model of the sequence will suffice to model correctly the joint transition probabilities of the primary inputs and internal states in the target circuit. From the experimental point of view, it is shown that assuming temporal independence or using firstorder temporal models is not sufficient, that is, the inaccuracies induced in steady-state and transition probability calculations are significant for most of the analyzed benchmarks.