Fault-Simulation Methods - Extensions and Comparison

2235 Therefore, we have attempted to extend the existing methods, wherever necessary, before making the comparisons between methods. Before proceeding to the analysis of the methods, we present a brief description of each method. Historically, parallel simulation was the first method that simulated a number of faults simultaneously. This method, which is perhaps the most widely used, takes advantage of word-oriented operations in the host computer and packs together several faulty circuit values into one or more computer words. Although this method is quite efficient, multiple passes are required for simulating large numbers of faults. Deductive simulation attempts to eliminate the need for multiple passes by computing normal signal values in the circuit and deducing the faulty values by manipulating lists of faults. Associated with each signal is a fault list, which is a set of faults, any one of which will cause the signal value to be different from the normal value. The effects of faults are propagated through the circuit by an algebra of sets. The multilist method associates two or more lists of faults with each signal. Conceptually, the number of lists associated with a signal is equal to the number of logic values simulated. Thus, for two logic values, there will be a O-list and a 1-list associated with each signal, the former being the set of faults in whose presence (individually) the signal will have the value 0, and the latter those that result in a value of 1. Set algebra is necessary for manipulating these lists also.