Parallelizing neural network training for cluster systems
PDCN '08 Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Networks
| Hi-index | 0.00 |
We discuss the current state of development for the key aspects of MPI programming on clusters. These aspects are the evolution of the MPI Standard itself, developments in cluster hardware and system software that directly affect MPI implementations, and supporting software that facilitates th use of MPI on scalable clusters. In each case we give a brief background and summarize the current status.