Edge separability based circuit clustering with application to circuit partitioning

In this paper, we introduce a new efficient O(n log n) graph search based bottom-up clustering algorithm named ESC (Edge Separability based Clustering). Unlike existing bottom-up algorithms that are based on local connectivity information of the netlist, ESC exploits more global connectivity information "edge separability" to guide clustering process while carefully monitoring cluster area balance. Computing the edge separability for a given edge e=(x, y) in an edge weighted undirected graph G(V, E, s, w) is equivalent to finding the x-y mincut. Then, we show that a simple and efficient algorithm CAPFOREST can be used to provide a good estimation of edge separability for all edges in G without using any network flow computation. Related experiments based on large scale ISPD98 benchmark circuits confirm that exploiting edge separability yields better quality partitioning solution compared to various bottom-up clustering algorithms proposed in the literature including Absorption, Density, Rent Parameter, Ratio Cut, Closeness, and Connectivity method. In addition, our ESC based multiway partitioning algorithm LR/ESC-PM provides comparable results to state-of-the-art hMetis and hMetis-Kway.