Integrating Efficient Partitioning Techniques For Graph Oriented Applications

. Load balancing and graph partitioning are areas of current research. Fractiling, a dynamic scheduling algorithm which balances loads by handling both predictable and unpredictable events, has been proven to improve application performance for N-body simulations. It operates by repeatedly subdividing the work into decreasing size chunks. For graph oriented applications, this means that a given graph would have to be partitioned many times throughout the execution of the application. Therefore to operate efficiently, when incorporated into graph oriented applications, Fractiling would require a highly efficient graph partitioning algorithm. This paper presents some of the current ongoing research being conducted by the NSF Engineering Research Center for Computational Field Simulation at Mississippi State University. This research seeks to improve the performance of graph oriented Computational Fluid Dynamic (CFD) applications by the integration of Fractiling and a highly efficient gr..