Efficient dynamic pinning of parallelized applications by reinforcement learning with applications

Georgios Chasparis; Michael Rossbory; Vladimir Janjic

doi:10.1007/978-3-319-64203-1_12

Efficient dynamic pinning of parallelized applications by reinforcement learning with applications

Georgios Chasparis, Michael Rossbory, Vladimir Janjic

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

Abstract

This paper describes a dynamic framework for mapping the threads of parallel applications to the computation cores of parallel systems. We propose a feedback-based mechanism where the performance of each thread is collected and used to drive the reinforcement-learning policy of assigning affinities of threads to CPU cores. The proposed framework is flexible enough to address different optimization criteria, such as maximum processing speed and minimum speed variance among threads. We evaluate the framework on the Ant Colony optimization parallel benchmark from the heuristic optimization application domain, and demonstrate that we can achieve an improvement of 12% in the execution time compared to the default operating system scheduling/mapping of threads under varying availability of resources (e.g. when multiple applications are running on the same system).

Original language	English
Title of host publication	Euro-Par 2017: Parallel Processing
Subtitle of host publication	23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings
Editors	Francisco F. Rivera, Tomás F. Pena, José C. Cabaleiro
Place of Publication	Cham
Publisher	Springer
Pages	164-176
Number of pages	13
ISBN (Electronic)	9783319642031
ISBN (Print)	9783319642024
DOIs	https://doi.org/10.1007/978-3-319-64203-1_12
Publication status	Published - 1 Aug 2017
Event	23rd International Conference on Parallel and Distributed Computing (Euro-Par) - Santiago de Compostela, Spain Duration: 28 Aug 2017 → 1 Sept 2017 Conference number: 23 https://europar2017.usc.es/index.html#euro-par-2017

Publication series

Name	Lecture Notes in Computer Science (Theoretical Computer Science and General Issues)
Publisher	Springer
Volume	10417
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	23rd International Conference on Parallel and Distributed Computing (Euro-Par)
Abbreviated title	Euro-Par 2017
Country/Territory	Spain
City	Santiago de Compostela
Period	28/08/17 → 1/09/17
Internet address	https://europar2017.usc.es/index.html#euro-par-2017

Access to Document

10.1007/978-3-319-64203-1_12

Chasparis-2017-EuroPar17-AAM
© 2017, Springer International Publishing AG. This work has been made available online in accordance with the publisher's policies. This is the author created accepted version manuscript following peer review and as such may differ slightly from the final published version. The final published version of this work is available at https://doi.org/10.1007/978-3-319-64203-1_12
Accepted author manuscript, 459 KB

https://zenodo.org/record/1186628Licence: CC BY

H2020 Collaboration REPHRASE: H2020 Collaboration 2014 - RePhrase
Hammond, K. (PI)
European Commission
1/04/15 → 31/03/18
Project: Standard

Cite this

Chasparis, G., Rossbory, M., & Janjic, V. (2017). Efficient dynamic pinning of parallelized applications by reinforcement learning with applications. In F. F. Rivera, T. F. Pena, & J. C. Cabaleiro (Eds.), Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings (pp. 164-176). (Lecture Notes in Computer Science (Theoretical Computer Science and General Issues); Vol. 10417). Springer. https://doi.org/10.1007/978-3-319-64203-1_12

Chasparis, Georgios ; Rossbory, Michael ; Janjic, Vladimir. / Efficient dynamic pinning of parallelized applications by reinforcement learning with applications. Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings. editor / Francisco F. Rivera ; Tomás F. Pena ; José C. Cabaleiro. Cham : Springer, 2017. pp. 164-176 (Lecture Notes in Computer Science (Theoretical Computer Science and General Issues)).

@inproceedings{9fdbba0823d3450e98cd5d1cf881607c,

title = "Efficient dynamic pinning of parallelized applications by reinforcement learning with applications",

abstract = "This paper describes a dynamic framework for mapping the threads of parallel applications to the computation cores of parallel systems. We propose a feedback-based mechanism where the performance of each thread is collected and used to drive the reinforcement-learning policy of assigning affinities of threads to CPU cores. The proposed framework is flexible enough to address different optimization criteria, such as maximum processing speed and minimum speed variance among threads. We evaluate the framework on the Ant Colony optimization parallel benchmark from the heuristic optimization application domain, and demonstrate that we can achieve an improvement of 12% in the execution time compared to the default operating system scheduling/mapping of threads under varying availability of resources (e.g. when multiple applications are running on the same system).",

author = "Georgios Chasparis and Michael Rossbory and Vladimir Janjic",

note = "Funding: This work has been partially supported by the European Union grant EU H2020-ICT-2014-1 project RePhrase (No. 644235).; 23rd International Conference on Parallel and Distributed Computing (Euro-Par), Euro-Par 2017 ; Conference date: 28-08-2017 Through 01-09-2017",

year = "2017",

month = aug,

day = "1",

doi = "10.1007/978-3-319-64203-1_12",

language = "English",

isbn = "9783319642024",

series = "Lecture Notes in Computer Science (Theoretical Computer Science and General Issues)",

publisher = "Springer",

pages = "164--176",

editor = "Rivera, {Francisco F.} and Pena, {Tom{\'a}s F.} and Cabaleiro, {Jos{\'e} C.}",

booktitle = "Euro-Par 2017: Parallel Processing",

address = "Netherlands",

url = "https://europar2017.usc.es/index.html#euro-par-2017",

}

Chasparis, G, Rossbory, M & Janjic, V 2017, Efficient dynamic pinning of parallelized applications by reinforcement learning with applications. in FF Rivera, TF Pena & JC Cabaleiro (eds), Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings. Lecture Notes in Computer Science (Theoretical Computer Science and General Issues), vol. 10417, Springer, Cham, pp. 164-176, 23rd International Conference on Parallel and Distributed Computing (Euro-Par), Santiago de Compostela, Spain, 28/08/17. https://doi.org/10.1007/978-3-319-64203-1_12

Efficient dynamic pinning of parallelized applications by reinforcement learning with applications. / Chasparis, Georgios; Rossbory, Michael; Janjic, Vladimir.
Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings. ed. / Francisco F. Rivera; Tomás F. Pena; José C. Cabaleiro. Cham: Springer, 2017. p. 164-176 (Lecture Notes in Computer Science (Theoretical Computer Science and General Issues); Vol. 10417).

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

TY - GEN

T1 - Efficient dynamic pinning of parallelized applications by reinforcement learning with applications

AU - Chasparis, Georgios

AU - Rossbory, Michael

AU - Janjic, Vladimir

N1 - Conference code: 23

PY - 2017/8/1

Y1 - 2017/8/1

N2 - This paper describes a dynamic framework for mapping the threads of parallel applications to the computation cores of parallel systems. We propose a feedback-based mechanism where the performance of each thread is collected and used to drive the reinforcement-learning policy of assigning affinities of threads to CPU cores. The proposed framework is flexible enough to address different optimization criteria, such as maximum processing speed and minimum speed variance among threads. We evaluate the framework on the Ant Colony optimization parallel benchmark from the heuristic optimization application domain, and demonstrate that we can achieve an improvement of 12% in the execution time compared to the default operating system scheduling/mapping of threads under varying availability of resources (e.g. when multiple applications are running on the same system).

AB - This paper describes a dynamic framework for mapping the threads of parallel applications to the computation cores of parallel systems. We propose a feedback-based mechanism where the performance of each thread is collected and used to drive the reinforcement-learning policy of assigning affinities of threads to CPU cores. The proposed framework is flexible enough to address different optimization criteria, such as maximum processing speed and minimum speed variance among threads. We evaluate the framework on the Ant Colony optimization parallel benchmark from the heuristic optimization application domain, and demonstrate that we can achieve an improvement of 12% in the execution time compared to the default operating system scheduling/mapping of threads under varying availability of resources (e.g. when multiple applications are running on the same system).

U2 - 10.1007/978-3-319-64203-1_12

DO - 10.1007/978-3-319-64203-1_12

M3 - Conference contribution

SN - 9783319642024

T3 - Lecture Notes in Computer Science (Theoretical Computer Science and General Issues)

SP - 164

EP - 176

BT - Euro-Par 2017: Parallel Processing

A2 - Rivera, Francisco F.

A2 - Pena, Tomás F.

A2 - Cabaleiro, José C.

PB - Springer

CY - Cham

T2 - 23rd International Conference on Parallel and Distributed Computing (Euro-Par)

Y2 - 28 August 2017 through 1 September 2017

ER -

Chasparis G, Rossbory M, Janjic V. Efficient dynamic pinning of parallelized applications by reinforcement learning with applications. In Rivera FF, Pena TF, Cabaleiro JC, editors, Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings. Cham: Springer. 2017. p. 164-176. (Lecture Notes in Computer Science (Theoretical Computer Science and General Issues)). doi: 10.1007/978-3-319-64203-1_12

Efficient dynamic pinning of parallelized applications by reinforcement learning with applications

Abstract

Publication series

Conference

Access to Document

Fingerprint

Projects

H2020 Collaboration REPHRASE: H2020 Collaboration 2014 - RePhrase

Cite this