Better speedups using simpler parallel programming for graph connectivity and biconnectivity