Massively parallel automated software tuning

Jakub Kurzak; Yaohung M. Tsai; Mark Gates; Ahmad Abdelfattah; Jack Dongarra

doi:10.1145/3337821.3337908

Massively parallel automated software tuning

Jakub Kurzak, Yaohung M. Tsai, Mark Gates, Ahmad Abdelfattah, Jack Dongarra

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

3 Scopus citations

Abstract

This article presents an implementation of a distributed autotuning engine developed as part of the Bench-testing OpenN Software Autotuning Infrastructure project. The system is geared towards performance optimization of computational kernels for graphics processing units, and allows for the deployment of vast autotuning sweeps to massively parallel machines. The software implements dynamic work scheduling to distributed-memory resources and takes advantage of multithreading for parallel compilation and dispatches kernel launches to multiple accelerators. This paper lays out the main design principles of the system and discusses the basic mechanics of the initial implementation. Preliminary performance results are presented, encountered challenges are discussed, and the future directions are outlined.

Original language	English
Title of host publication	Proceedings of the 48th International Conference on Parallel Processing, ICPP 2019
Publisher	Association for Computing Machinery
ISBN (Electronic)	9781450362955
DOIs	https://doi.org/10.1145/3337821.3337908
State	Published - Aug 5 2019
Externally published	Yes
Event	48th International Conference on Parallel Processing, ICPP 2019 - Kyoto, Japan Duration: Aug 5 2019 → Aug 8 2019

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	48th International Conference on Parallel Processing, ICPP 2019
Country/Territory	Japan
City	Kyoto
Period	08/5/19 → 08/8/19

Keywords

Automated software tuning
Graphics processing unit

Access to Document

10.1145/3337821.3337908

Cite this

@inproceedings{4273b8b909d04a29b3ea430b03dd4fd1,

title = "Massively parallel automated software tuning",

abstract = "This article presents an implementation of a distributed autotuning engine developed as part of the Bench-testing OpenN Software Autotuning Infrastructure project. The system is geared towards performance optimization of computational kernels for graphics processing units, and allows for the deployment of vast autotuning sweeps to massively parallel machines. The software implements dynamic work scheduling to distributed-memory resources and takes advantage of multithreading for parallel compilation and dispatches kernel launches to multiple accelerators. This paper lays out the main design principles of the system and discusses the basic mechanics of the initial implementation. Preliminary performance results are presented, encountered challenges are discussed, and the future directions are outlined.",

keywords = "Automated software tuning, Graphics processing unit",

author = "Jakub Kurzak and Tsai, {Yaohung M.} and Mark Gates and Ahmad Abdelfattah and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} 2019 ACM.; 48th International Conference on Parallel Processing, ICPP 2019 ; Conference date: 05-08-2019 Through 08-08-2019",

year = "2019",

month = aug,

day = "5",

doi = "10.1145/3337821.3337908",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

booktitle = "Proceedings of the 48th International Conference on Parallel Processing, ICPP 2019",

}

Kurzak, J, Tsai, YM, Gates, M, Abdelfattah, A & Dongarra, J 2019, Massively parallel automated software tuning. in Proceedings of the 48th International Conference on Parallel Processing, ICPP 2019., a92, ACM International Conference Proceeding Series, Association for Computing Machinery, 48th International Conference on Parallel Processing, ICPP 2019, Kyoto, Japan, 08/5/19. https://doi.org/10.1145/3337821.3337908

Massively parallel automated software tuning. / Kurzak, Jakub; Tsai, Yaohung M.; Gates, Mark et al.
Proceedings of the 48th International Conference on Parallel Processing, ICPP 2019. Association for Computing Machinery, 2019. a92 (ACM International Conference Proceeding Series).

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

TY - GEN

T1 - Massively parallel automated software tuning

AU - Kurzak, Jakub

AU - Tsai, Yaohung M.

AU - Gates, Mark

AU - Abdelfattah, Ahmad

AU - Dongarra, Jack

PY - 2019/8/5

Y1 - 2019/8/5

N2 - This article presents an implementation of a distributed autotuning engine developed as part of the Bench-testing OpenN Software Autotuning Infrastructure project. The system is geared towards performance optimization of computational kernels for graphics processing units, and allows for the deployment of vast autotuning sweeps to massively parallel machines. The software implements dynamic work scheduling to distributed-memory resources and takes advantage of multithreading for parallel compilation and dispatches kernel launches to multiple accelerators. This paper lays out the main design principles of the system and discusses the basic mechanics of the initial implementation. Preliminary performance results are presented, encountered challenges are discussed, and the future directions are outlined.

AB - This article presents an implementation of a distributed autotuning engine developed as part of the Bench-testing OpenN Software Autotuning Infrastructure project. The system is geared towards performance optimization of computational kernels for graphics processing units, and allows for the deployment of vast autotuning sweeps to massively parallel machines. The software implements dynamic work scheduling to distributed-memory resources and takes advantage of multithreading for parallel compilation and dispatches kernel launches to multiple accelerators. This paper lays out the main design principles of the system and discusses the basic mechanics of the initial implementation. Preliminary performance results are presented, encountered challenges are discussed, and the future directions are outlined.

KW - Automated software tuning

KW - Graphics processing unit

UR - http://www.scopus.com/inward/record.url?scp=85071072188&partnerID=8YFLogxK

U2 - 10.1145/3337821.3337908

DO - 10.1145/3337821.3337908

M3 - Conference contribution

AN - SCOPUS:85071072188

T3 - ACM International Conference Proceeding Series

BT - Proceedings of the 48th International Conference on Parallel Processing, ICPP 2019

PB - Association for Computing Machinery

T2 - 48th International Conference on Parallel Processing, ICPP 2019

Y2 - 5 August 2019 through 8 August 2019

ER -

Massively parallel automated software tuning

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this