Symmetric data objects and remote memory access communication for fortran-95 applications

Authors:
Jarek Nieplocha;Doug Baxter;Vinod Tipparaju;Craig Rasmunssen;Robert W. Numrich
Affiliations:
Pacific Northwest National Laboratory, Richland, WA;Pacific Northwest National Laboratory, Richland, WA;Pacific Northwest National Laboratory, Richland, WA;Los Alamos National Laboratory, Los Alamos, NM;Minnesota Supercomputing Institute, University of Minnesota, Minneapolis, MN
Venue:
Euro-Par'05 Proceedings of the 11th international Euro-Par conference on Parallel Processing
Year:
2005

Citing 5
Cited 0

Co-array Fortran for parallel programming

ACM SIGPLAN Fortran Forum
Performance Evaluation of the Quadrics Interconnection Network

Cluster Computing
ARMCI: A Portable Remote Memory Copy Libray for Ditributed Array Libraries and Compiler Run-Time Systems

Proceedings of the 11 IPPS/SPDP'99 Workshops Held in Conjunction with the 13th International Parallel Processing Symposium and 10th Symposium on Parallel and Distributed Processing
Performance and Experience with LAPI -- A New High-Performance Communication Library for the IBM RS/6000 SP

IPPS '98 Proceedings of the 12th. International Parallel Processing Symposium on International Parallel Processing Symposium
Bridging the language gap in scientific computing: the Chasm approach

Concurrency and Computation: Practice & Experience - Computational Frameworks

Quantified Score

Hi-index	0.01

Visualization

Abstract

Symmetric data objects have been introduced by Cray Inc. in context of SHMEM remote memory access communication on Cray T3D/E systems and later adopted by SGI for their Origin servers. Symmetric data objects greatly simplify parallel programming by allowing programmers to reference remote instance of a data structure by specifying address of the local counterpart. The current paper describes how symmetric data objects and remote memory access communication could be implemented in Fortran-95 without requiring specialized hardware or compiler support. NAS Multi-Grid parallel benchmark was used as an application example and demonstrated competitive performance to the standard MPI implementation.