Efficient Inter-node MPI Communication Using GPUDirect RDMA for InfiniBand Clusters with NVIDIA GPUs2013-10-010 PRE AI DOI:10.1109/icpp.2013.17原文链接原文求助分享收藏摘要 En