An application framework for migrating GPGPU Cloud applications

Sho Yuhara, Yusuke Suzuki, Kenji Kono

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

Abstract

Graphics Processing Units (GPUS) have become a common computing resource for general-purpose computing (GPGPU). GPU usage has also spread to high-Throughput server applications, taking advantage of its massively parallel nature and wide availability at various cloud platforms. Although various methods currently exist to share a single GPU among multiple applications, migrating GPGPU server applications across different machines is challenging due to lack of hardware mechanisms, such as programmable preemption and access to GPU context. This paper presents an event-driven framework for GPGPU server applications, which enables us to implement a software based approach for migration which overcomes current hardware limitations.

Original languageEnglish
Title of host publicationProceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018
PublisherIEEE Computer Society
Pages62-66
Number of pages5
Volume2018-December
ISBN (Electronic)9781538678992
DOIs
Publication statusPublished - 2018 Dec 26
Event10th International Conference on Cloud Computing Technology and Science, CloudCom 2018 - Nicosia, Cyprus
Duration: 2018 Dec 102018 Dec 13

Other

Other10th International Conference on Cloud Computing Technology and Science, CloudCom 2018
CountryCyprus
CityNicosia
Period18/12/1018/12/13

Fingerprint

GPGPU
Servers
Server
Hardware
Preemption
Computing
Event-driven
Graphics Processing Unit
High Throughput
Migration
Availability
Throughput
Resources
Software
Framework
Graphics processing unit

Keywords

  • Cloud Computing
  • GPGPU
  • Migration

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Networks and Communications
  • Software
  • Theoretical Computer Science

Cite this

Yuhara, S., Suzuki, Y., & Kono, K. (2018). An application framework for migrating GPGPU Cloud applications. In Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018 (Vol. 2018-December, pp. 62-66). [8590995] IEEE Computer Society. https://doi.org/10.1109/CloudCom2018.2018.00026

An application framework for migrating GPGPU Cloud applications. / Yuhara, Sho; Suzuki, Yusuke; Kono, Kenji.

Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018. Vol. 2018-December IEEE Computer Society, 2018. p. 62-66 8590995.

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

Yuhara, S, Suzuki, Y & Kono, K 2018, An application framework for migrating GPGPU Cloud applications. in Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018. vol. 2018-December, 8590995, IEEE Computer Society, pp. 62-66, 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018, Nicosia, Cyprus, 18/12/10. https://doi.org/10.1109/CloudCom2018.2018.00026
Yuhara S, Suzuki Y, Kono K. An application framework for migrating GPGPU Cloud applications. In Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018. Vol. 2018-December. IEEE Computer Society. 2018. p. 62-66. 8590995 https://doi.org/10.1109/CloudCom2018.2018.00026
Yuhara, Sho ; Suzuki, Yusuke ; Kono, Kenji. / An application framework for migrating GPGPU Cloud applications. Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018. Vol. 2018-December IEEE Computer Society, 2018. pp. 62-66
@inproceedings{761ba134dfb6440d90a4088ff2f72035,
title = "An application framework for migrating GPGPU Cloud applications",
abstract = "Graphics Processing Units (GPUS) have become a common computing resource for general-purpose computing (GPGPU). GPU usage has also spread to high-Throughput server applications, taking advantage of its massively parallel nature and wide availability at various cloud platforms. Although various methods currently exist to share a single GPU among multiple applications, migrating GPGPU server applications across different machines is challenging due to lack of hardware mechanisms, such as programmable preemption and access to GPU context. This paper presents an event-driven framework for GPGPU server applications, which enables us to implement a software based approach for migration which overcomes current hardware limitations.",
keywords = "Cloud Computing, GPGPU, Migration",
author = "Sho Yuhara and Yusuke Suzuki and Kenji Kono",
year = "2018",
month = "12",
day = "26",
doi = "10.1109/CloudCom2018.2018.00026",
language = "English",
volume = "2018-December",
pages = "62--66",
booktitle = "Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - An application framework for migrating GPGPU Cloud applications

AU - Yuhara, Sho

AU - Suzuki, Yusuke

AU - Kono, Kenji

PY - 2018/12/26

Y1 - 2018/12/26

N2 - Graphics Processing Units (GPUS) have become a common computing resource for general-purpose computing (GPGPU). GPU usage has also spread to high-Throughput server applications, taking advantage of its massively parallel nature and wide availability at various cloud platforms. Although various methods currently exist to share a single GPU among multiple applications, migrating GPGPU server applications across different machines is challenging due to lack of hardware mechanisms, such as programmable preemption and access to GPU context. This paper presents an event-driven framework for GPGPU server applications, which enables us to implement a software based approach for migration which overcomes current hardware limitations.

AB - Graphics Processing Units (GPUS) have become a common computing resource for general-purpose computing (GPGPU). GPU usage has also spread to high-Throughput server applications, taking advantage of its massively parallel nature and wide availability at various cloud platforms. Although various methods currently exist to share a single GPU among multiple applications, migrating GPGPU server applications across different machines is challenging due to lack of hardware mechanisms, such as programmable preemption and access to GPU context. This paper presents an event-driven framework for GPGPU server applications, which enables us to implement a software based approach for migration which overcomes current hardware limitations.

KW - Cloud Computing

KW - GPGPU

KW - Migration

UR - http://www.scopus.com/inward/record.url?scp=85061137365&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061137365&partnerID=8YFLogxK

U2 - 10.1109/CloudCom2018.2018.00026

DO - 10.1109/CloudCom2018.2018.00026

M3 - Conference contribution

AN - SCOPUS:85061137365

VL - 2018-December

SP - 62

EP - 66

BT - Proceedings - IEEE 10th International Conference on Cloud Computing Technology and Science, CloudCom 2018

PB - IEEE Computer Society

ER -