Can Network-Offload based Non-Blocking Neighborhood MPI Collectives Improve Communication Overheads of Irregular Graph Algorithms? Int'l Workshop on Parallel Algorithm and Parallel Software (IWPAPS12)
K. Kandalla, H. Subramoni, K. Tomko, J. Vienne, L. Oliker, D. Panda
held in conjunction with IEEE Cluster (Cluster '12),
Sep 2012.