Autoscaling of cores in multicore processors using power and thermal workload signatures

Rupesh Raj Karn; Ibrahim M. Elfadel

doi:10.1109/MWSCAS.2016.7870120

Autoscaling of cores in multicore processors using power and thermal workload signatures

Rupesh Raj Karn
, Ibrahim M. Elfadel

Electrical Engineering

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

1 Scopus citations

Abstract

Autoscaling of cloud clusters based on dynamic estimation of workloads is a well-known practice in data center management. However, this practice has not been widely adopted in the multicore processor area due to the lack of a real-Time workload classification front end. In this paper, we present a novel methodology for core autoscaling in multicore processors. The methodology is based on the identification of workload signatures at both the core and the thread level. In particular, correlations between thread-based performance counters are used to decide thread migration policies to maximize per-core utilization and reduce the number of active cores. Power, thermal, and performance-Aware autoscaling policies are presented, and extensive numerical experiments are used to illustrate the advantages of our algorithm for real-Time multicore power and performance management.

Original language	British English
Title of host publication	2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781509009169
DOIs	https://doi.org/10.1109/MWSCAS.2016.7870120
State	Published - 2 Jul 2016
Event	59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016 - Abu Dhabi, United Arab Emirates Duration: 16 Oct 2016 → 19 Oct 2016

Publication series

Name	Midwest Symposium on Circuits and Systems
Volume	0
ISSN (Print)	1548-3746

Conference

Conference	59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016
Country/Territory	United Arab Emirates
City	Abu Dhabi
Period	16/10/16 → 19/10/16

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 10 Reduced Inequalities

Keywords

Core
Correlation
Performance
Power
Signature
Temperature
Workload

Access to Document

10.1109/MWSCAS.2016.7870120

OpenUrl availability

Full text

Cite this

Karn, R. R., & Elfadel, I. M. (2016). Autoscaling of cores in multicore processors using power and thermal workload signatures. In 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016 Article 7870120 (Midwest Symposium on Circuits and Systems; Vol. 0). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/MWSCAS.2016.7870120

@inproceedings{1323721a6c094689be3661b5a81bd4fa,

title = "Autoscaling of cores in multicore processors using power and thermal workload signatures",

abstract = "Autoscaling of cloud clusters based on dynamic estimation of workloads is a well-known practice in data center management. However, this practice has not been widely adopted in the multicore processor area due to the lack of a real-Time workload classification front end. In this paper, we present a novel methodology for core autoscaling in multicore processors. The methodology is based on the identification of workload signatures at both the core and the thread level. In particular, correlations between thread-based performance counters are used to decide thread migration policies to maximize per-core utilization and reduce the number of active cores. Power, thermal, and performance-Aware autoscaling policies are presented, and extensive numerical experiments are used to illustrate the advantages of our algorithm for real-Time multicore power and performance management.",

keywords = "Core, Correlation, Performance, Power, Signature, Temperature, Workload",

author = "Karn, \{Rupesh Raj\} and Elfadel, \{Ibrahim M.\}",

note = "Funding Information: The authors would like to acknowledge M. A. Silva, H. Franke, G. Kandiraju and P. Kudva from IBM Research for useful discussions. They also acknowledge the inputs of A. Henroid (Intel), and P. Bose, A. Buyuktosunoglu and C. Isci (IBM research) in the early stages of this research. This work is supported by the SRC with customised funding from Mubadala, Abu Dhabi, UAE. Publisher Copyright: {\textcopyright} 2016 IEEE.; 59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016 ; Conference date: 16-10-2016 Through 19-10-2016",

year = "2016",

month = jul,

day = "2",

doi = "10.1109/MWSCAS.2016.7870120",

language = "British English",

series = "Midwest Symposium on Circuits and Systems",

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

booktitle = "2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016",

address = "United States",

}

Karn, RR & Elfadel, IM 2016, Autoscaling of cores in multicore processors using power and thermal workload signatures. in 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016., 7870120, Midwest Symposium on Circuits and Systems, vol. 0, Institute of Electrical and Electronics Engineers Inc., 59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016, Abu Dhabi, United Arab Emirates, 16/10/16. https://doi.org/10.1109/MWSCAS.2016.7870120

Autoscaling of cores in multicore processors using power and thermal workload signatures. / Karn, Rupesh Raj; Elfadel, Ibrahim M.
2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016. Institute of Electrical and Electronics Engineers Inc., 2016. 7870120 (Midwest Symposium on Circuits and Systems; Vol. 0).

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

TY - GEN

T1 - Autoscaling of cores in multicore processors using power and thermal workload signatures

AU - Karn, Rupesh Raj

AU - Elfadel, Ibrahim M.

N1 - Funding Information: The authors would like to acknowledge M. A. Silva, H. Franke, G. Kandiraju and P. Kudva from IBM Research for useful discussions. They also acknowledge the inputs of A. Henroid (Intel), and P. Bose, A. Buyuktosunoglu and C. Isci (IBM research) in the early stages of this research. This work is supported by the SRC with customised funding from Mubadala, Abu Dhabi, UAE. Publisher Copyright: © 2016 IEEE.

PY - 2016/7/2

Y1 - 2016/7/2

N2 - Autoscaling of cloud clusters based on dynamic estimation of workloads is a well-known practice in data center management. However, this practice has not been widely adopted in the multicore processor area due to the lack of a real-Time workload classification front end. In this paper, we present a novel methodology for core autoscaling in multicore processors. The methodology is based on the identification of workload signatures at both the core and the thread level. In particular, correlations between thread-based performance counters are used to decide thread migration policies to maximize per-core utilization and reduce the number of active cores. Power, thermal, and performance-Aware autoscaling policies are presented, and extensive numerical experiments are used to illustrate the advantages of our algorithm for real-Time multicore power and performance management.

AB - Autoscaling of cloud clusters based on dynamic estimation of workloads is a well-known practice in data center management. However, this practice has not been widely adopted in the multicore processor area due to the lack of a real-Time workload classification front end. In this paper, we present a novel methodology for core autoscaling in multicore processors. The methodology is based on the identification of workload signatures at both the core and the thread level. In particular, correlations between thread-based performance counters are used to decide thread migration policies to maximize per-core utilization and reduce the number of active cores. Power, thermal, and performance-Aware autoscaling policies are presented, and extensive numerical experiments are used to illustrate the advantages of our algorithm for real-Time multicore power and performance management.

KW - Core

KW - Correlation

KW - Performance

KW - Power

KW - Signature

KW - Temperature

KW - Workload

UR - https://www.scopus.com/pages/publications/85015976704

U2 - 10.1109/MWSCAS.2016.7870120

DO - 10.1109/MWSCAS.2016.7870120

M3 - Conference contribution

AN - SCOPUS:85015976704

T3 - Midwest Symposium on Circuits and Systems

BT - 2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016

Y2 - 16 October 2016 through 19 October 2016

ER -

Autoscaling of cores in multicore processors using power and thermal workload signatures

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

OpenUrl availability

Other files and links

Fingerprint

Cite this