Search Results for author: Amith R Mamidala

Found 2 papers, 0 papers with code

Efficient Embedding of MPI Collectives in MXNET DAGs for scaling Deep Learning

no code implementations20 Feb 2018 Amith R Mamidala

In particular, we demonstrate three designs: Funneled, Concurrent communication and Dependency chaining of using MPI collectives with DAGs.

MXNET-MPI: Embedding MPI parallelism in Parameter Server Task Model for scaling Deep Learning

no code implementations11 Jan 2018 Amith R Mamidala, Georgios Kollias, Chris Ward, Fausto Artico

In this paper, we discuss the drawbacks of such approaches and propose a generic framework supporting both PS and MPI programming paradigms, co-existing at the same time.

Cannot find the paper you are looking for? You can Submit a new open access paper.