Algebraic decomposition of MCNC benchmark FSMs for logic synthesis

Abstract

An interactive application of algebraic structure theory is presented to decompose finite state machines (FSMs) for logic synthesis purposes. Parallel and serial decompositions with symbolic states are achieved. CASTOR, JEDI ESPRESSO, and MIS are utilized to perform state assignment and logic minimization. Experimental results of two-level and multi-level implementations of MCNC benchmark FSMs are presented.<>