An efficient algorithm for constrained encoding and its applications

Abstract

An efficient algorithm and its implementation, ENCORE, are presented for finding approximate solutions to dichotomy-based constrained encoding, a problem fundamental to the synthesis of combinational logic circuits, and synchronous and asynchronous sequential circuits. ENCORE adopts a greedy strategy to find an encoding bit by bit, and then uses an iterative method to improve the solution quality. The novelty of the algorithm lies mainly in a linear-time heuristic to select each individual bit; this problem was previously solved in quadratic time. ENCORE has been applied to a variety of practical problem instances. For a number of examples found in the literature on the synthesis of asynchronous sequential machines, ENCORE consistently obtains optimal or near-optimal results. For the optimum state assignment of the MCNC FSM benchmarks, ENCORE generates the same or even shorter encoding lengths than the programs KISS, NOVA and DIET, but takes much less CPU time. It is demonstrated for the first time that PLA implementations of synchronous FSMs using dichotomy constraints compare very favorably with respect to area with those based on traditional group constraints.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>