Shared memory NUMA programming on I-WAY

J. Nieplocha; R. J. Harrison

doi:10.1109/HPDC.1996.546214

Shared memory NUMA programming on I-WAY

J. Nieplocha
, R. J. Harrison

Applied Mathematics and Statistics

Pacific Northwest National Laboratory

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

15 Scopus citations

Abstract

The performance of the Global Array shared-memory non-uniform memory-access programming model is explored on the I-WAY, wide-area-network distributed supercomputer environment. The Global Array model is extended by introducing a concept of mirrored arrays. Latencies and bandwidths for remote memory access are studied, and the performance of a large application from computational chemistry is evaluated using both fully distributed and also mirrored arrays. Excellent performance can be obtained with mirroring if even modest (0.5 MB/s) network bandwidth is available.

Original language	English
Title of host publication	IEEE International Symposium on High Performance Distributed Computing, Proceedings
Pages	432-441
Number of pages	10
DOIs	https://doi.org/10.1109/HPDC.1996.546214
State	Published - 1996
Event	Proceedings of the 1996 5th IEEE International Symposium on High Performance Distributed Computing - Syracuse, NY, USA Duration: Aug 6 1996 → Aug 9 1996

Conference

Conference	Proceedings of the 1996 5th IEEE International Symposium on High Performance Distributed Computing
City	Syracuse, NY, USA
Period	08/6/96 → 08/9/96

Access to Document

10.1109/HPDC.1996.546214

Cite this

@inproceedings{57478ccf279d4c0ebf2a71ab664c10c5,

title = "Shared memory NUMA programming on I-WAY",

abstract = "The performance of the Global Array shared-memory non-uniform memory-access programming model is explored on the I-WAY, wide-area-network distributed supercomputer environment. The Global Array model is extended by introducing a concept of mirrored arrays. Latencies and bandwidths for remote memory access are studied, and the performance of a large application from computational chemistry is evaluated using both fully distributed and also mirrored arrays. Excellent performance can be obtained with mirroring if even modest (0.5 MB/s) network bandwidth is available.",

author = "J. Nieplocha and Harrison, \{R. J.\}",

year = "1996",

doi = "10.1109/HPDC.1996.546214",

language = "English",

isbn = "0818675829",

pages = "432--441",

booktitle = "IEEE International Symposium on High Performance Distributed Computing, Proceedings",

note = "Proceedings of the 1996 5th IEEE International Symposium on High Performance Distributed Computing ; Conference date: 06-08-1996 Through 09-08-1996",

}

TY - GEN

T1 - Shared memory NUMA programming on I-WAY

AU - Nieplocha, J.

AU - Harrison, R. J.

PY - 1996

Y1 - 1996

N2 - The performance of the Global Array shared-memory non-uniform memory-access programming model is explored on the I-WAY, wide-area-network distributed supercomputer environment. The Global Array model is extended by introducing a concept of mirrored arrays. Latencies and bandwidths for remote memory access are studied, and the performance of a large application from computational chemistry is evaluated using both fully distributed and also mirrored arrays. Excellent performance can be obtained with mirroring if even modest (0.5 MB/s) network bandwidth is available.

AB - The performance of the Global Array shared-memory non-uniform memory-access programming model is explored on the I-WAY, wide-area-network distributed supercomputer environment. The Global Array model is extended by introducing a concept of mirrored arrays. Latencies and bandwidths for remote memory access are studied, and the performance of a large application from computational chemistry is evaluated using both fully distributed and also mirrored arrays. Excellent performance can be obtained with mirroring if even modest (0.5 MB/s) network bandwidth is available.

UR - https://www.scopus.com/pages/publications/0029749270

U2 - 10.1109/HPDC.1996.546214

DO - 10.1109/HPDC.1996.546214

M3 - Conference contribution

AN - SCOPUS:0029749270

SN - 0818675829

SP - 432

EP - 441

BT - IEEE International Symposium on High Performance Distributed Computing, Proceedings

T2 - Proceedings of the 1996 5th IEEE International Symposium on High Performance Distributed Computing

Y2 - 6 August 1996 through 9 August 1996

ER -

Shared memory NUMA programming on I-WAY

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this