GTgraph: A Synthetic Graph Generator Suite

This document details the characteristics, input parameters and suggested usage of the three graph generators included in this suite. 1 SSCA#2 graph generator This generator produces graphs used in the DARPA HPCS SSCA#2 benchmark [1]. The SSCA#2 graph is directed with integer edge weights, and made up of random-sized cliques, with a hierarchical inter-clique distribution of edges based on a distance metric. An SSCA#2 clique is defined as a maximal set of vertices, where each pair of vertices is connected by directed edges in one or both directions. The following parameters can be specified by the user. The default values of each parameter are also given below. • SCALE: An integer value used to express the rest of the graph parameters. The problem size can be increased by increasing the value of SCALE. • TotVertices: The number of vertices (2) • MaxCliqueSize: The maximum number of vertices in a clique. Clique sizes are distributed uniformly on [1, MaxCliqueSize] ([2]) • ProbUnidirectional: Probability that the connections between two vertices will be unidirectional as opposed to bidirectional (0.2) • MaxParallelEdges: The maximum number of parallel edges from one vertex to another. The actual number is distributed uniformly on [1, MaxParallelEdges] (3) • ProbIntercliqueEdges: Initial probability of an interclique edges (0.5) • MaxIntWeight: Maximum value of integer edge weight. The weights are distributed uniformly on [1, MaxIntWeight] (2) For a detailed description of the intraand inter-clique edge generation algorithms, refer [1].