New efficient clique partitioning algorithms for register-transfer synthesis of data paths

Numerous problems can be modeled as clique partitioning problems in digital design synthesis. In this paper, we present two new polynomial time heuristic algorithms for efficient clique partitioning with or without limiting the maximum clique size. The goal of clique partitioning is to partition a graph into a minimum number of cliques. The basic approach of the new algorithm is to find small cliques by first using a fast connectivity test. The performance comparisons of the new algorithm produces better results than other existing clique partitioning algorithms. The modified algorithm, which limits the maximum clique size, becomes more efficient and can be used to solve problems in register transfer synthesis of pipelined data paths, such as interconnection sharing. The results show that the new polynomial-time heuristic algorithm finds a near optimal solution quickly.