The maximal utilization of processor co-allocation in multicluster systems

A.I.D. Bucur; D.H.J. Epema

doi:10.1109/IPDPS.2003.1213154

The maximal utilization of processor co-allocation in multicluster systems

A.I.D. Bucur, D.H.J. Epema

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

29 Citations (Scopus)

205 Downloads (Pure)

Abstract

In systems consisting of multiple clusters of processors which employ space sharing for scheduling jobs, such as our distributed ASCI supercomputer (DAS), co-allocation, i.e., the simultaneous allocation of processors to single jobs in multiple clusters, may be required. In studies of scheduling in single clusters it has been shown that the achievable (maximal) utilization may be much less than 100%, a problem that may be aggravated in multicluster systems. In this paper we study the maximal utilization when co-allocating jobs in multicluster systems, both with analytic means (we derive exact and approximate formulas when the service-time distribution is exponential), and with simulations with synthetic workloads and with workloads derived from the logs of actual systems.

Original language	English
Title of host publication	Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS 2003, Nice, France, April 22-26, 2003)
Publisher	Institute of Electrical and Electronics Engineers
Pages	1-10
ISBN (Print)	0-7695-1926-1
DOIs	https://doi.org/10.1109/IPDPS.2003.1213154
Publication status	Published - 2003

Access to Document

10.1109/IPDPS.2003.1213154

Metis255886Final published version, 364 KB

Cite this

@inproceedings{d88fa33d279d4a78b1d45758a2018c5f,

title = "The maximal utilization of processor co-allocation in multicluster systems",

abstract = "In systems consisting of multiple clusters of processors which employ space sharing for scheduling jobs, such as our distributed ASCI supercomputer (DAS), co-allocation, i.e., the simultaneous allocation of processors to single jobs in multiple clusters, may be required. In studies of scheduling in single clusters it has been shown that the achievable (maximal) utilization may be much less than 100%, a problem that may be aggravated in multicluster systems. In this paper we study the maximal utilization when co-allocating jobs in multicluster systems, both with analytic means (we derive exact and approximate formulas when the service-time distribution is exponential), and with simulations with synthetic workloads and with workloads derived from the logs of actual systems.",

author = "A.I.D. Bucur and D.H.J. Epema",

year = "2003",

doi = "10.1109/IPDPS.2003.1213154",

language = "English",

isbn = "0-7695-1926-1",

pages = "1--10",

booktitle = "Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS 2003, Nice, France, April 22-26, 2003)",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

}

The maximal utilization of processor co-allocation in multicluster systems. / Bucur, A.I.D.; Epema, D.H.J.
Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS 2003, Nice, France, April 22-26, 2003). Institute of Electrical and Electronics Engineers, 2003. p. 1-10.

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

TY - GEN

T1 - The maximal utilization of processor co-allocation in multicluster systems

AU - Bucur, A.I.D.

AU - Epema, D.H.J.

PY - 2003

Y1 - 2003

N2 - In systems consisting of multiple clusters of processors which employ space sharing for scheduling jobs, such as our distributed ASCI supercomputer (DAS), co-allocation, i.e., the simultaneous allocation of processors to single jobs in multiple clusters, may be required. In studies of scheduling in single clusters it has been shown that the achievable (maximal) utilization may be much less than 100%, a problem that may be aggravated in multicluster systems. In this paper we study the maximal utilization when co-allocating jobs in multicluster systems, both with analytic means (we derive exact and approximate formulas when the service-time distribution is exponential), and with simulations with synthetic workloads and with workloads derived from the logs of actual systems.

AB - In systems consisting of multiple clusters of processors which employ space sharing for scheduling jobs, such as our distributed ASCI supercomputer (DAS), co-allocation, i.e., the simultaneous allocation of processors to single jobs in multiple clusters, may be required. In studies of scheduling in single clusters it has been shown that the achievable (maximal) utilization may be much less than 100%, a problem that may be aggravated in multicluster systems. In this paper we study the maximal utilization when co-allocating jobs in multicluster systems, both with analytic means (we derive exact and approximate formulas when the service-time distribution is exponential), and with simulations with synthetic workloads and with workloads derived from the logs of actual systems.

U2 - 10.1109/IPDPS.2003.1213154

DO - 10.1109/IPDPS.2003.1213154

M3 - Conference contribution

SN - 0-7695-1926-1

SP - 1

EP - 10

BT - Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS 2003, Nice, France, April 22-26, 2003)

PB - Institute of Electrical and Electronics Engineers

ER -

The maximal utilization of processor co-allocation in multicluster systems

Abstract

Access to Document

Fingerprint

Cite this