The Community for Technology Leaders
2013 IEEE International Conference on Cluster Computing (CLUSTER) (2003)
Hong Kong
Dec. 1, 2003 to Dec. 4, 2003
ISBN: 0-7695-2066-9
pp: 259
Jay Brockman , University of Notre Dame
Richard Murphy , University of Notre Dame
Keith Underwood , Sandia National Labs
Peter Kogge , University of Notre Dame
Ron Brightwell , Sandia National Labs
Arun Rodrigues , University of Notre Dame
Memory may be the only system component that is more commoditized than a microprocessor. To simultaneously exploit this and address the impending memory wall, processing in memory (PIM) research efforts are considering ways to move processing into memory without significantly increasing the cost of the memory. As such, PIM devices may become the basis for future commodity clusters. Although these PIM devices may leverage new computational paradigms such as hardware support for multi-threading and traveling threads, they must provide support for legacy programming models if they are to supplant commodity clusters. This paper presents a prototype implementation of MPI over a traveling thread mechanism called parcels. A performance analysis indicates that the direct hardware support of a traveling thread model can lead to an efficient, lightweight MPI implementation.
Jay Brockman, Richard Murphy, Keith Underwood, Peter Kogge, Ron Brightwell, Arun Rodrigues, "Implications of a PIM Architectural Model for MPI", 2013 IEEE International Conference on Cluster Computing (CLUSTER), vol. 00, no. , pp. 259, 2003, doi:10.1109/CLUSTR.2003.1253323
106 ms
(Ver 3.1 (10032016))