A simple performance model for multithreaded applications executing on non-uniform memory access computers

R. Yang; J. Antony; A. P. Rendell

doi:10.1109/HPCC.2009.39

A simple performance model for multithreaded applications executing on non-uniform memory access computers

R. Yang, J. Antony, A. P. Rendell

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

12 Citations (Scopus)

Abstract

In this work, we extend and evaluate a simple performance model to account for NUMA and bandwidth effects for single and multi-threaded calculations within the Gaussian 03 computational chemistry code on a contemporary multi-core, NUMA platform. By using the thread and memory placement APIs in Solaris, we present results for a set of calculations from which we analyze on-chip interconnect and intra-core bandwidth contention and show the importance of load-balancing between threads. The extended model predicts single threaded performance to within 1% errors and most multi-threaded experiments within 15% errors. Our results and modeling shows that accounting for bandwidth constraints within user-space code is beneficial.

Original language	English
Title of host publication	2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009
Pages	79-86
Number of pages	8
DOIs	https://doi.org/10.1109/HPCC.2009.39
Publication status	Published - 20 Nov 2009
Externally published	Yes
Event	11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009 - Seoul, Korea, Republic of Duration: 25 Jun 2009 → 27 Jun 2009

Publication series

Name	2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009

Conference

Conference	11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009
Country/Territory	Korea, Republic of
City	Seoul
Period	25/06/09 → 27/06/09

Keywords

performance modelling
NUMA
multithreaded application

Access to Document

10.1109/HPCC.2009.39Licence: In Copyright

Cite this

Yang, R., Antony, J., & Rendell, A. P. (2009). A simple performance model for multithreaded applications executing on non-uniform memory access computers. In 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009 (pp. 79-86). Article 5166979 (2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009). https://doi.org/10.1109/HPCC.2009.39

@inproceedings{0e8a091571d746178f71c4577904b8dc,

title = "A simple performance model for multithreaded applications executing on non-uniform memory access computers",

abstract = "In this work, we extend and evaluate a simple performance model to account for NUMA and bandwidth effects for single and multi-threaded calculations within the Gaussian 03 computational chemistry code on a contemporary multi-core, NUMA platform. By using the thread and memory placement APIs in Solaris, we present results for a set of calculations from which we analyze on-chip interconnect and intra-core bandwidth contention and show the importance of load-balancing between threads. The extended model predicts single threaded performance to within 1% errors and most multi-threaded experiments within 15% errors. Our results and modeling shows that accounting for bandwidth constraints within user-space code is beneficial.",

keywords = "performance modelling, NUMA, multithreaded application",

author = "R. Yang and J. Antony and Rendell, {A. P.}",

year = "2009",

month = nov,

day = "20",

doi = "10.1109/HPCC.2009.39",

language = "English",

isbn = "9780769537382",

series = "2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009",

pages = "79--86",

booktitle = "2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009",

note = "11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009 ; Conference date: 25-06-2009 Through 27-06-2009",

}

Yang, R, Antony, J & Rendell, AP 2009, A simple performance model for multithreaded applications executing on non-uniform memory access computers. in 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009., 5166979, 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009, pp. 79-86, 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009, Seoul, Korea, Republic of, 25/06/09. https://doi.org/10.1109/HPCC.2009.39

A simple performance model for multithreaded applications executing on non-uniform memory access computers. / Yang, R.; Antony, J.; Rendell, A. P.
2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009. 2009. p. 79-86 5166979 (2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009).

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

TY - GEN

T1 - A simple performance model for multithreaded applications executing on non-uniform memory access computers

AU - Yang, R.

AU - Antony, J.

AU - Rendell, A. P.

PY - 2009/11/20

Y1 - 2009/11/20

N2 - In this work, we extend and evaluate a simple performance model to account for NUMA and bandwidth effects for single and multi-threaded calculations within the Gaussian 03 computational chemistry code on a contemporary multi-core, NUMA platform. By using the thread and memory placement APIs in Solaris, we present results for a set of calculations from which we analyze on-chip interconnect and intra-core bandwidth contention and show the importance of load-balancing between threads. The extended model predicts single threaded performance to within 1% errors and most multi-threaded experiments within 15% errors. Our results and modeling shows that accounting for bandwidth constraints within user-space code is beneficial.

AB - In this work, we extend and evaluate a simple performance model to account for NUMA and bandwidth effects for single and multi-threaded calculations within the Gaussian 03 computational chemistry code on a contemporary multi-core, NUMA platform. By using the thread and memory placement APIs in Solaris, we present results for a set of calculations from which we analyze on-chip interconnect and intra-core bandwidth contention and show the importance of load-balancing between threads. The extended model predicts single threaded performance to within 1% errors and most multi-threaded experiments within 15% errors. Our results and modeling shows that accounting for bandwidth constraints within user-space code is beneficial.

KW - performance modelling

KW - NUMA

KW - multithreaded application

UR - http://www.scopus.com/inward/record.url?scp=70449584580&partnerID=8YFLogxK

UR - http://purl.org/au-research/grants/ARC/LP0347178

UR - http://purl.org/au-research/grants/ARC/LP0774896

U2 - 10.1109/HPCC.2009.39

DO - 10.1109/HPCC.2009.39

M3 - Conference contribution

AN - SCOPUS:70449584580

SN - 9780769537382

T3 - 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009

SP - 79

EP - 86

BT - 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009

T2 - 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009

Y2 - 25 June 2009 through 27 June 2009

ER -

Yang R, Antony J, Rendell AP. A simple performance model for multithreaded applications executing on non-uniform memory access computers. In 2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009. 2009. p. 79-86. 5166979. (2009 11th IEEE International Conference on High Performance Computing and Communications, HPCC 2009). doi: 10.1109/HPCC.2009.39

A simple performance model for multithreaded applications executing on non-uniform memory access computers

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this