Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies

V. J. Leung; E. M. Arkin; M. A. Bender; D. Bunde; J. Johnston; Alok Lal; J. S.B. Mitchell; C. Phillips; S. S. Seiden

doi:10.1109/CLUSTR.2002.1137758

Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies

V. J. Leung
, E. M. Arkin
, M. A. Bender
, D. Bunde
, J. Johnston
, Alok Lal
, J. S.B. Mitchell
, C. Phillips
, S. S. Seiden

Sandia National Laboratories, New Mexico
University of Illinois at Urbana-Champaign
Tufts University
Louisiana State University

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

51 Scopus citations

Abstract

The Computational Plant or Cplant is a commodity-based supercomputer under development at Sandia National Laboratories. This paper describes resource-allocation strategies to achieve processor locality for parallel jobs in Cplant and other supercomputers. Users of Cplant and other Sandia supercomputers submit parallel jobs to a job queue. When a job is scheduled to run, it is assigned to a set of processors. To obtain maximum throughput, jobs should be allocated to localized clusters of processors to minimize communication costs and to avoid bandwidth contention caused by overlapping jobs. This paper introduces new allocation strategies and performance metrics based on space-filling curves and one dimensional allocation strategies. These algorithms are general and simple. Preliminary simulations and Cplant experiments indicate that both space-filling curves and one-dimensional packing improve processor locality compared to the sorted free list strategy previously used on Cplant. These new allocation strategies are implemented in the new release of the Cplant System Software, Version 2.0, phased into the Cplant systems at Sandia by May 2002.

Original language	English
Title of host publication	Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002
Editors	Bill Gropp, Rajkumar Buyya, Rob Pennington, Maxine Brown, Mark Baker, Dan Reed
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	296-304
Number of pages	9
ISBN (Electronic)	0769517455
DOIs	https://doi.org/10.1109/CLUSTR.2002.1137758
State	Published - 2002
Event	IEEE International Conference on Cluster Computing, CLUSTER 2002 - Chicago, United States Duration: Sep 23 2002 → Sep 26 2002

Publication series

Name	Proceedings - IEEE International Conference on Cluster Computing, ICCC
Volume	2002-January
ISSN (Print)	1552-5244

Conference

Conference	IEEE International Conference on Cluster Computing, CLUSTER 2002
Country/Territory	United States
City	Chicago
Period	09/23/02 → 09/26/02

Keywords

Bandwidth
Communication switching
Computer science
Costs
Laboratories
Processor scheduling
Resource management
Supercomputers
Switches
Throughput

Access to Document

10.1109/CLUSTR.2002.1137758

Cite this

Leung, V. J., Arkin, E. M., Bender, M. A., Bunde, D., Johnston, J., Lal, A., Mitchell, J. S. B., Phillips, C., & Seiden, S. S. (2002). Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies. In B. Gropp, R. Buyya, R. Pennington, M. Brown, M. Baker, & D. Reed (Eds.), Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002 (pp. 296-304). Article 1137758 (Proceedings - IEEE International Conference on Cluster Computing, ICCC; Vol. 2002-January). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CLUSTR.2002.1137758

Leung, V. J. ; Arkin, E. M. ; Bender, M. A. et al. / Processor allocation on Cplant : Achieving general processor locality using one-dimensional allocation strategies. Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002. editor / Bill Gropp ; Rajkumar Buyya ; Rob Pennington ; Maxine Brown ; Mark Baker ; Dan Reed. Institute of Electrical and Electronics Engineers Inc., 2002. pp. 296-304 (Proceedings - IEEE International Conference on Cluster Computing, ICCC).

@inproceedings{346fba15b6d44790a3da7b8c87a31057,

title = "Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies",

abstract = "The Computational Plant or Cplant is a commodity-based supercomputer under development at Sandia National Laboratories. This paper describes resource-allocation strategies to achieve processor locality for parallel jobs in Cplant and other supercomputers. Users of Cplant and other Sandia supercomputers submit parallel jobs to a job queue. When a job is scheduled to run, it is assigned to a set of processors. To obtain maximum throughput, jobs should be allocated to localized clusters of processors to minimize communication costs and to avoid bandwidth contention caused by overlapping jobs. This paper introduces new allocation strategies and performance metrics based on space-filling curves and one dimensional allocation strategies. These algorithms are general and simple. Preliminary simulations and Cplant experiments indicate that both space-filling curves and one-dimensional packing improve processor locality compared to the sorted free list strategy previously used on Cplant. These new allocation strategies are implemented in the new release of the Cplant System Software, Version 2.0, phased into the Cplant systems at Sandia by May 2002.",

keywords = "Bandwidth, Communication switching, Computer science, Costs, Laboratories, Processor scheduling, Resource management, Supercomputers, Switches, Throughput",

author = "Leung, \{V. J.\} and Arkin, \{E. M.\} and Bender, \{M. A.\} and D. Bunde and J. Johnston and Alok Lal and Mitchell, \{J. S.B.\} and C. Phillips and Seiden, \{S. S.\}",

note = "Publisher Copyright: {\textcopyright} 2002 IEEE.; IEEE International Conference on Cluster Computing, CLUSTER 2002 ; Conference date: 23-09-2002 Through 26-09-2002",

year = "2002",

doi = "10.1109/CLUSTR.2002.1137758",

language = "English",

series = "Proceedings - IEEE International Conference on Cluster Computing, ICCC",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "296--304",

editor = "Bill Gropp and Rajkumar Buyya and Rob Pennington and Maxine Brown and Mark Baker and Dan Reed",

booktitle = "Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002",

}

Leung, VJ, Arkin, EM , Bender, MA, Bunde, D, Johnston, J, Lal, A, Mitchell, JSB, Phillips, C & Seiden, SS 2002, Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies. in B Gropp, R Buyya, R Pennington, M Brown, M Baker & D Reed (eds), Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002., 1137758, Proceedings - IEEE International Conference on Cluster Computing, ICCC, vol. 2002-January, Institute of Electrical and Electronics Engineers Inc., pp. 296-304, IEEE International Conference on Cluster Computing, CLUSTER 2002, Chicago, United States, 09/23/02. https://doi.org/10.1109/CLUSTR.2002.1137758

Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies. / Leung, V. J.; Arkin, E. M.; Bender, M. A. et al.
Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002. ed. / Bill Gropp; Rajkumar Buyya; Rob Pennington; Maxine Brown; Mark Baker; Dan Reed. Institute of Electrical and Electronics Engineers Inc., 2002. p. 296-304 1137758 (Proceedings - IEEE International Conference on Cluster Computing, ICCC; Vol. 2002-January).

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

TY - GEN

T1 - Processor allocation on Cplant

T2 - IEEE International Conference on Cluster Computing, CLUSTER 2002

AU - Leung, V. J.

AU - Arkin, E. M.

AU - Bender, M. A.

AU - Bunde, D.

AU - Johnston, J.

AU - Lal, Alok

AU - Mitchell, J. S.B.

AU - Phillips, C.

AU - Seiden, S. S.

PY - 2002

Y1 - 2002

N2 - The Computational Plant or Cplant is a commodity-based supercomputer under development at Sandia National Laboratories. This paper describes resource-allocation strategies to achieve processor locality for parallel jobs in Cplant and other supercomputers. Users of Cplant and other Sandia supercomputers submit parallel jobs to a job queue. When a job is scheduled to run, it is assigned to a set of processors. To obtain maximum throughput, jobs should be allocated to localized clusters of processors to minimize communication costs and to avoid bandwidth contention caused by overlapping jobs. This paper introduces new allocation strategies and performance metrics based on space-filling curves and one dimensional allocation strategies. These algorithms are general and simple. Preliminary simulations and Cplant experiments indicate that both space-filling curves and one-dimensional packing improve processor locality compared to the sorted free list strategy previously used on Cplant. These new allocation strategies are implemented in the new release of the Cplant System Software, Version 2.0, phased into the Cplant systems at Sandia by May 2002.

AB - The Computational Plant or Cplant is a commodity-based supercomputer under development at Sandia National Laboratories. This paper describes resource-allocation strategies to achieve processor locality for parallel jobs in Cplant and other supercomputers. Users of Cplant and other Sandia supercomputers submit parallel jobs to a job queue. When a job is scheduled to run, it is assigned to a set of processors. To obtain maximum throughput, jobs should be allocated to localized clusters of processors to minimize communication costs and to avoid bandwidth contention caused by overlapping jobs. This paper introduces new allocation strategies and performance metrics based on space-filling curves and one dimensional allocation strategies. These algorithms are general and simple. Preliminary simulations and Cplant experiments indicate that both space-filling curves and one-dimensional packing improve processor locality compared to the sorted free list strategy previously used on Cplant. These new allocation strategies are implemented in the new release of the Cplant System Software, Version 2.0, phased into the Cplant systems at Sandia by May 2002.

KW - Bandwidth

KW - Communication switching

KW - Computer science

KW - Costs

KW - Laboratories

KW - Processor scheduling

KW - Resource management

KW - Supercomputers

KW - Switches

KW - Throughput

UR - https://www.scopus.com/pages/publications/84948970946

U2 - 10.1109/CLUSTR.2002.1137758

DO - 10.1109/CLUSTR.2002.1137758

M3 - Conference contribution

AN - SCOPUS:84948970946

T3 - Proceedings - IEEE International Conference on Cluster Computing, ICCC

SP - 296

EP - 304

BT - Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002

A2 - Gropp, Bill

A2 - Buyya, Rajkumar

A2 - Pennington, Rob

A2 - Brown, Maxine

A2 - Baker, Mark

A2 - Reed, Dan

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 23 September 2002 through 26 September 2002

ER -

Leung VJ, Arkin EM , Bender MA, Bunde D, Johnston J, Lal A et al. Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies. In Gropp B, Buyya R, Pennington R, Brown M, Baker M, Reed D, editors, Proceedings - IEEE International Conference on Cluster Computing, CLUSTER 2002. Institute of Electrical and Electronics Engineers Inc. 2002. p. 296-304. 1137758. (Proceedings - IEEE International Conference on Cluster Computing, ICCC). doi: 10.1109/CLUSTR.2002.1137758

Processor allocation on Cplant: Achieving general processor locality using one-dimensional allocation strategies

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this