A New Approach For Factorizing FSM's

Abstract

Exact Factors as defined in [2], if present in an FSM can result in most effective way of factorization. However, it has been found that most of the FSM's are not exact factorizable. In this paper, we have suggested a method of making FSM's exact factorizable by minor changes in the next state space while maintaining the functionality of the FSM. We have also developed a new combined state assignment algorithm for state encoding of Factored and Factoring FSM's. Experimental results on MCNC benchmark examples, after running MISII on the Original FSM, Factored FSM and Factoring FSM have shown a reduction of 40% in the worst case signal delay through the circuit in a multilevel implementation. The total number of literals, on an average is the same after factorization as that obtained by running MISII on the original FSM. For two-level implementation, our method has been able to factorize Benchmark FSM's with a 14% average increase in overall areas, while the areas of combinational components of Factored and Factoring FSM's have been found to be significantly less than the area of the combinational component of the original FSM.