Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era

Bharathwaj Raghunathan, Siddharth Garg

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

Abstract

The rate at which jobs arrive for processing at servers in a data-center (i.e., the job arrival rate) can vary significantly with time. Each server in a data-center is a multi-core processor, allowing jobs to be processed with different degrees of parallelism (DoPs) (i.e., number of threads per job). In this paper, we show both analytically and empirically that the optimal DoP that minimizes mean service time varies with job arrival rate. In addition, we show that for asymmetric multi-core server processors (i.e., processors with multiple clusters, each consisting of cores of a different type, and assuming that only one cluster is active at any given time while the others are dark), the best cluster to select is also dependent on job arrival rate. Based on these observations, we propose a run-time scheduler that determines the optimal DoP and performs inter-cluster migration to minimize mean service time within a power budget. Experimental results demonstrate significant reduction in mean service time compared to job arrival rate unaware schedulers.

Original languageEnglish (US)
Title of host publication2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014
PublisherAssociation for Computing Machinery, Inc
Pages1-6
Number of pages6
ISBN (Print)9781450330510
DOIs
StatePublished - Oct 12 2014
Event2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014 - New Delhi, India
Duration: Oct 12 2014Oct 17 2014

Other

Other2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014
CountryIndia
CityNew Delhi
Period10/12/1410/17/14

Fingerprint

Servers
Scheduling
Silicon
Processing

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture

Cite this

Raghunathan, B., & Garg, S. (2014). Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era. In 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014 (pp. 1-6). [a14] Association for Computing Machinery, Inc. https://doi.org/10.1145/2656075.2656091

Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era. / Raghunathan, Bharathwaj; Garg, Siddharth.

2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014. Association for Computing Machinery, Inc, 2014. p. 1-6 a14.

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

Raghunathan, B & Garg, S 2014, Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era. in 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014., a14, Association for Computing Machinery, Inc, pp. 1-6, 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014, New Delhi, India, 10/12/14. https://doi.org/10.1145/2656075.2656091
Raghunathan B, Garg S. Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era. In 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014. Association for Computing Machinery, Inc. 2014. p. 1-6. a14 https://doi.org/10.1145/2656075.2656091
Raghunathan, Bharathwaj ; Garg, Siddharth. / Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era. 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014. Association for Computing Machinery, Inc, 2014. pp. 1-6
@inproceedings{c090ba871cb940ccb7dd1096a4bcddbc,
title = "Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era",
abstract = "The rate at which jobs arrive for processing at servers in a data-center (i.e., the job arrival rate) can vary significantly with time. Each server in a data-center is a multi-core processor, allowing jobs to be processed with different degrees of parallelism (DoPs) (i.e., number of threads per job). In this paper, we show both analytically and empirically that the optimal DoP that minimizes mean service time varies with job arrival rate. In addition, we show that for asymmetric multi-core server processors (i.e., processors with multiple clusters, each consisting of cores of a different type, and assuming that only one cluster is active at any given time while the others are dark), the best cluster to select is also dependent on job arrival rate. Based on these observations, we propose a run-time scheduler that determines the optimal DoP and performs inter-cluster migration to minimize mean service time within a power budget. Experimental results demonstrate significant reduction in mean service time compared to job arrival rate unaware schedulers.",
author = "Bharathwaj Raghunathan and Siddharth Garg",
year = "2014",
month = "10",
day = "12",
doi = "10.1145/2656075.2656091",
language = "English (US)",
isbn = "9781450330510",
pages = "1--6",
booktitle = "2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - Job arrival rate aware scheduling for asymmetric multi-core servers in the dark silicon era

AU - Raghunathan, Bharathwaj

AU - Garg, Siddharth

PY - 2014/10/12

Y1 - 2014/10/12

N2 - The rate at which jobs arrive for processing at servers in a data-center (i.e., the job arrival rate) can vary significantly with time. Each server in a data-center is a multi-core processor, allowing jobs to be processed with different degrees of parallelism (DoPs) (i.e., number of threads per job). In this paper, we show both analytically and empirically that the optimal DoP that minimizes mean service time varies with job arrival rate. In addition, we show that for asymmetric multi-core server processors (i.e., processors with multiple clusters, each consisting of cores of a different type, and assuming that only one cluster is active at any given time while the others are dark), the best cluster to select is also dependent on job arrival rate. Based on these observations, we propose a run-time scheduler that determines the optimal DoP and performs inter-cluster migration to minimize mean service time within a power budget. Experimental results demonstrate significant reduction in mean service time compared to job arrival rate unaware schedulers.

AB - The rate at which jobs arrive for processing at servers in a data-center (i.e., the job arrival rate) can vary significantly with time. Each server in a data-center is a multi-core processor, allowing jobs to be processed with different degrees of parallelism (DoPs) (i.e., number of threads per job). In this paper, we show both analytically and empirically that the optimal DoP that minimizes mean service time varies with job arrival rate. In addition, we show that for asymmetric multi-core server processors (i.e., processors with multiple clusters, each consisting of cores of a different type, and assuming that only one cluster is active at any given time while the others are dark), the best cluster to select is also dependent on job arrival rate. Based on these observations, we propose a run-time scheduler that determines the optimal DoP and performs inter-cluster migration to minimize mean service time within a power budget. Experimental results demonstrate significant reduction in mean service time compared to job arrival rate unaware schedulers.

UR - http://www.scopus.com/inward/record.url?scp=84910607196&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84910607196&partnerID=8YFLogxK

U2 - 10.1145/2656075.2656091

DO - 10.1145/2656075.2656091

M3 - Conference contribution

AN - SCOPUS:84910607196

SN - 9781450330510

SP - 1

EP - 6

BT - 2014 International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2014

PB - Association for Computing Machinery, Inc

ER -