Array folding using heuristics and simulated annealing

Abstract

Array folding is a technique to reduce the area of a logic array by exploiting its sparsity. Since the problem is NP-complete, heuristic algorithms that yield a near optimal solution have to be employed. In this paper, a method employing a combination of simulated annealing and heuristic algorithms has been used to find a near optimal solution for both simple and multiple folding of logic arrays. The algorithms developed have been implemented in a computer program called GAMIN-SA. When compared to PLEASURE, GAMIN-SA was seen to perform as good or better with regard to quality of solution and, for the bigger arrays, it was better in terms of run-time as well (multiple folding).<>