Packet size optimization for supporting coarse-grained pipelined parallelism | IEEE Conference Publication | IEEE Xplore