Accelerating LINPACK with MPI-OpenCL on Clusters of Multi-GPU Nodes | IEEE Journals & Magazine | IEEE Xplore