Nov 20, 2014 · Abstract: An increasing number of MPI applications are being ported to take advantage of the compute power offered by GPUs.

In this paper, we first propose a set of optimized techniques to handle different MPI data types. Next, we propose a novel framework (HAND) that enables hybrid ...

Current MVAPICH2-GDR support for non-contiguous data movement processing uses an efficient latency oriented framework which combines generic and targeted GPU ...

This approach also enables the MPI runtimes to optimize data movement between GPUs using advanced features (such as. GPUDirect RDMA and CUDA IPC) and techniques ...

HAND: A Hybrid Approach to Accelerate Non-contiguous Data Movement Using MPI Datatypes on GPU Clusters. Authors: Rong Shi. Rong Shi. View Profile. , Xiaoyi Lu.

HAND: A Hybrid Approach to Accelerate Non-contiguous Data Movement Using MPI Datatypes on GPU Clusters. Shi, Rong, Lu, Xiaoyi, Potluri, Sreeram, Hamidouche ...

HAND: A Hybrid Approach to Accelerate Non-contiguous Data Movement Using MPI Datatypes on GPU Clusters. Shi, R., Lu, X., Potluri, S., Hamidouche, K., Zhang, ...

HAND: A Hybrid Approach to Accelerate Non-contiguous Data Movement Using MPI Datatypes on GPU Clusters. Proceedings of International Conference on Parallel ...

[RDF data]. HAND: A Hybrid Approach to Accelerate Non-contiguous Data Movement Using MPI Datatypes on GPU Clusters. Resource URI: https://dblp.l3s.de/d2r ...

MPI datatype processing algorithm for data in GPU memory ... 2D, 3D arrays with large, contiguous chunks. Use hand-coded packing kernels for small sized, simple ...