Performance characteristics of Graph500 on large-scale distributed environment

Toyotaro Suzumura, Koji Ueno, Hitoshi Sato, Katsuki Fujisawa, Satoshi Matsuoka

Research output: Chapter in Book/Report/Conference proceedingConference contribution

39 Citations (Scopus)

Abstract

Graph500 is a new benchmark for supercomputers based on large-scale graph analysis, which is becoming an important form of analysis in many real-world applications. Graph algorithms run well on supercomputers with shared memory. For the Linpack-based supercomputer rankings, TOP500 reports that heterogeneous and distributed-memory super-computers with large numbers of GPGPUs are becoming dominant. However, the performance characteristics of large-scale graph analysis benchmarks such as Graph500 on distributed-memory supercomputers have so far received little study. This is the first report of a performance evaluation and analysis for Graph500 on a commodity-processor-based distributed-memory supercomputer. We found that the reference implementation "replicated-csr" based on distributed level-synchronized breadth-first search solves a large free graph problem with 231 vertices and 235 edges (approximately 2.15 billon vertices and 34.3 billion edges) in 3.09 seconds with 128 nodes and 3,072 cores. This equates to 11 giga-edges traversed per second. We describe the algorithms and implementations of the reference implementations of Graph500, and analyze the performance characteristics with varying graph sizes and numbers of computer nodes and different implementations. Our results will also contribute to the development of optimized algorithms for the coming exascale machines.

Original languageEnglish
Title of host publicationProceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011
Pages149-158
Number of pages10
DOIs
Publication statusPublished - Dec 1 2011
Event2011 IEEE International Symposium on Workload Characterization, IISWC - 2011 - Austin, TX, United States
Duration: Nov 6 2011Nov 8 2011

Publication series

NameProceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011

Other

Other2011 IEEE International Symposium on Workload Characterization, IISWC - 2011
CountryUnited States
CityAustin, TX
Period11/6/1111/8/11

Fingerprint

Supercomputers
Data storage equipment

All Science Journal Classification (ASJC) codes

  • Computational Theory and Mathematics
  • Computer Science Applications

Cite this

Suzumura, T., Ueno, K., Sato, H., Fujisawa, K., & Matsuoka, S. (2011). Performance characteristics of Graph500 on large-scale distributed environment. In Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011 (pp. 149-158). [6114175] (Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011). https://doi.org/10.1109/IISWC.2011.6114175

Performance characteristics of Graph500 on large-scale distributed environment. / Suzumura, Toyotaro; Ueno, Koji; Sato, Hitoshi; Fujisawa, Katsuki; Matsuoka, Satoshi.

Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011. 2011. p. 149-158 6114175 (Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Suzumura, T, Ueno, K, Sato, H, Fujisawa, K & Matsuoka, S 2011, Performance characteristics of Graph500 on large-scale distributed environment. in Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011., 6114175, Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011, pp. 149-158, 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011, Austin, TX, United States, 11/6/11. https://doi.org/10.1109/IISWC.2011.6114175
Suzumura T, Ueno K, Sato H, Fujisawa K, Matsuoka S. Performance characteristics of Graph500 on large-scale distributed environment. In Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011. 2011. p. 149-158. 6114175. (Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011). https://doi.org/10.1109/IISWC.2011.6114175
Suzumura, Toyotaro ; Ueno, Koji ; Sato, Hitoshi ; Fujisawa, Katsuki ; Matsuoka, Satoshi. / Performance characteristics of Graph500 on large-scale distributed environment. Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011. 2011. pp. 149-158 (Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011).
@inproceedings{224f42e75f6a4a53988baea85a086645,
title = "Performance characteristics of Graph500 on large-scale distributed environment",
abstract = "Graph500 is a new benchmark for supercomputers based on large-scale graph analysis, which is becoming an important form of analysis in many real-world applications. Graph algorithms run well on supercomputers with shared memory. For the Linpack-based supercomputer rankings, TOP500 reports that heterogeneous and distributed-memory super-computers with large numbers of GPGPUs are becoming dominant. However, the performance characteristics of large-scale graph analysis benchmarks such as Graph500 on distributed-memory supercomputers have so far received little study. This is the first report of a performance evaluation and analysis for Graph500 on a commodity-processor-based distributed-memory supercomputer. We found that the reference implementation {"}replicated-csr{"} based on distributed level-synchronized breadth-first search solves a large free graph problem with 231 vertices and 235 edges (approximately 2.15 billon vertices and 34.3 billion edges) in 3.09 seconds with 128 nodes and 3,072 cores. This equates to 11 giga-edges traversed per second. We describe the algorithms and implementations of the reference implementations of Graph500, and analyze the performance characteristics with varying graph sizes and numbers of computer nodes and different implementations. Our results will also contribute to the development of optimized algorithms for the coming exascale machines.",
author = "Toyotaro Suzumura and Koji Ueno and Hitoshi Sato and Katsuki Fujisawa and Satoshi Matsuoka",
year = "2011",
month = "12",
day = "1",
doi = "10.1109/IISWC.2011.6114175",
language = "English",
isbn = "9781457720642",
series = "Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011",
pages = "149--158",
booktitle = "Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011",

}

TY - GEN

T1 - Performance characteristics of Graph500 on large-scale distributed environment

AU - Suzumura, Toyotaro

AU - Ueno, Koji

AU - Sato, Hitoshi

AU - Fujisawa, Katsuki

AU - Matsuoka, Satoshi

PY - 2011/12/1

Y1 - 2011/12/1

N2 - Graph500 is a new benchmark for supercomputers based on large-scale graph analysis, which is becoming an important form of analysis in many real-world applications. Graph algorithms run well on supercomputers with shared memory. For the Linpack-based supercomputer rankings, TOP500 reports that heterogeneous and distributed-memory super-computers with large numbers of GPGPUs are becoming dominant. However, the performance characteristics of large-scale graph analysis benchmarks such as Graph500 on distributed-memory supercomputers have so far received little study. This is the first report of a performance evaluation and analysis for Graph500 on a commodity-processor-based distributed-memory supercomputer. We found that the reference implementation "replicated-csr" based on distributed level-synchronized breadth-first search solves a large free graph problem with 231 vertices and 235 edges (approximately 2.15 billon vertices and 34.3 billion edges) in 3.09 seconds with 128 nodes and 3,072 cores. This equates to 11 giga-edges traversed per second. We describe the algorithms and implementations of the reference implementations of Graph500, and analyze the performance characteristics with varying graph sizes and numbers of computer nodes and different implementations. Our results will also contribute to the development of optimized algorithms for the coming exascale machines.

AB - Graph500 is a new benchmark for supercomputers based on large-scale graph analysis, which is becoming an important form of analysis in many real-world applications. Graph algorithms run well on supercomputers with shared memory. For the Linpack-based supercomputer rankings, TOP500 reports that heterogeneous and distributed-memory super-computers with large numbers of GPGPUs are becoming dominant. However, the performance characteristics of large-scale graph analysis benchmarks such as Graph500 on distributed-memory supercomputers have so far received little study. This is the first report of a performance evaluation and analysis for Graph500 on a commodity-processor-based distributed-memory supercomputer. We found that the reference implementation "replicated-csr" based on distributed level-synchronized breadth-first search solves a large free graph problem with 231 vertices and 235 edges (approximately 2.15 billon vertices and 34.3 billion edges) in 3.09 seconds with 128 nodes and 3,072 cores. This equates to 11 giga-edges traversed per second. We describe the algorithms and implementations of the reference implementations of Graph500, and analyze the performance characteristics with varying graph sizes and numbers of computer nodes and different implementations. Our results will also contribute to the development of optimized algorithms for the coming exascale machines.

UR - http://www.scopus.com/inward/record.url?scp=84856155189&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84856155189&partnerID=8YFLogxK

U2 - 10.1109/IISWC.2011.6114175

DO - 10.1109/IISWC.2011.6114175

M3 - Conference contribution

AN - SCOPUS:84856155189

SN - 9781457720642

T3 - Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011

SP - 149

EP - 158

BT - Proceedings - 2011 IEEE International Symposium on Workload Characterization, IISWC - 2011

ER -