Coarse Grained Parallel Maximum Matching In Convex Bipartite Graphs

We present a coarse grained parallel algorithm for computing a maximum matching in a convex bipartite graph G = A;B;E. For p processors with N=p memory per processor, N = jAj+jBj, N=p p, the algorithm requires Ologp communication rounds and OTsequn p; m p + n p logp local computation, where n = jAj, m = jBj and Tsequn;m is the sequential time complexity for the problem. For the BSP model, this implies Ologp supersteps with OgN + gn p logp communication cost and OTsequn p; m p + n p logp local computation.

• Type: Conference
• Year: 1999
• Where: IPPS
• Authors: Prosenjit Bose, Albert Chan, Frank K. H. A. Dehne, Markus Latzel
• Keywords: Distributed And Parallel Computing, IPPS 1999, Logp Communication Cost, Logp Local Computation, OTsequn P