Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application

Blesson Varghese; Javier Prades; Carlos Reano; Federico Silla

doi:10.1109/eScience.2015.15

Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application

Blesson Varghese, Javier Prades, Carlos Reano, Federico Silla

School of Electronics, Electrical Engineering and Computer Science

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

8 Citations (Scopus)

312 Downloads (Pure)

Abstract

How can GPU acceleration be obtained as a service in a cluster? This question has become increasingly significant due to the inefficiency of installing GPUs on all nodes of a cluster. The research reported in this paper is motivated to address the above question by employing rCUDA (remote CUDA), a framework that facilitates Acceleration-as-a-Service (AaaS), such that the nodes of a cluster can request the acceleration of a set of remote GPUs on demand. The rCUDA framework exploits virtualisation and ensures that multiple nodes can share the same GPU. In this paper we test the feasibility of the rCUDA framework on a real-world application employed in the financial risk industry that can benefit from AaaS in the production setting. The results confirm the feasibility of rCUDA and highlight that rCUDA achieves similar performance compared to CUDA, provides consistent results, and more importantly, allows for a single application to benefit from all the GPUs available in the cluster without loosing efficiency.

Original language	English
Title of host publication	2015 11th IEEE International Conference on e-Science (e-Science)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	47-56
Number of pages	10
ISBN (Print)	978-1-4673-9325-6
DOIs	https://doi.org/10.1109/eScience.2015.15
Publication status	Published - Aug 2015
Event	2015 IEEE 11th International Conference on e-Science (e-Science) - Munich, Germany Duration: 31 Aug 2015 → 04 Sep 2015

Conference

Conference	2015 IEEE 11th International Conference on e-Science (e-Science)
Country/Territory	Germany
City	Munich
Period	31/08/2015 → 04/09/2015

Bibliographical note

Acceptance rate: 25%

Keywords

cloud computing
financial data processing
graphics processing units
parallel architectures
virtualisation
Acceleration-as-a-Service
GPU acceleration
financial application
financial risk industry
remote CUDA
virtualised GPU
Acceleration
Graphics processing units
Hardware
Kernel
Memory management
Servers
CUDA
GPU computing
rCUDA

Access to Document

10.1109/eScience.2015.15

Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application
© 2015 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript, 2.21 MB

Cite this

@inproceedings{9562599b7b654297b2c2cbdf1fd1a98e,

title = "Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application",

abstract = "How can GPU acceleration be obtained as a service in a cluster? This question has become increasingly significant due to the inefficiency of installing GPUs on all nodes of a cluster. The research reported in this paper is motivated to address the above question by employing rCUDA (remote CUDA), a framework that facilitates Acceleration-as-a-Service (AaaS), such that the nodes of a cluster can request the acceleration of a set of remote GPUs on demand. The rCUDA framework exploits virtualisation and ensures that multiple nodes can share the same GPU. In this paper we test the feasibility of the rCUDA framework on a real-world application employed in the financial risk industry that can benefit from AaaS in the production setting. The results confirm the feasibility of rCUDA and highlight that rCUDA achieves similar performance compared to CUDA, provides consistent results, and more importantly, allows for a single application to benefit from all the GPUs available in the cluster without loosing efficiency.",

keywords = "cloud computing, financial data processing, graphics processing units, parallel architectures, virtualisation, Acceleration-as-a-Service, GPU acceleration, financial application, financial risk industry, remote CUDA, virtualised GPU, Acceleration, Graphics processing units, Hardware, Kernel, Memory management, Servers, CUDA, GPU computing, rCUDA",

author = "Blesson Varghese and Javier Prades and Carlos Reano and Federico Silla",

note = "Acceptance rate: 25%; 2015 IEEE 11th International Conference on e-Science (e-Science) ; Conference date: 31-08-2015 Through 04-09-2015",

year = "2015",

month = aug,

doi = "10.1109/eScience.2015.15",

language = "English",

isbn = "978-1-4673-9325-6 ",

pages = "47--56",

booktitle = "2015 11th IEEE International Conference on e-Science (e-Science)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

}

Varghese, B, Prades, J, Reano, C & Silla, F 2015, Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application. in 2015 11th IEEE International Conference on e-Science (e-Science). Institute of Electrical and Electronics Engineers (IEEE), pp. 47-56, 2015 IEEE 11th International Conference on e-Science (e-Science), Munich, Germany, 31/08/2015. https://doi.org/10.1109/eScience.2015.15

Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application. / Varghese, Blesson; Prades, Javier; Reano, Carlos; Silla, Federico.

2015 11th IEEE International Conference on e-Science (e-Science). Institute of Electrical and Electronics Engineers (IEEE), 2015. p. 47-56.

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

TY - GEN

T1 - Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application

AU - Varghese, Blesson

AU - Prades, Javier

AU - Reano, Carlos

AU - Silla, Federico

N1 - Acceptance rate: 25%

PY - 2015/8

Y1 - 2015/8

N2 - How can GPU acceleration be obtained as a service in a cluster? This question has become increasingly significant due to the inefficiency of installing GPUs on all nodes of a cluster. The research reported in this paper is motivated to address the above question by employing rCUDA (remote CUDA), a framework that facilitates Acceleration-as-a-Service (AaaS), such that the nodes of a cluster can request the acceleration of a set of remote GPUs on demand. The rCUDA framework exploits virtualisation and ensures that multiple nodes can share the same GPU. In this paper we test the feasibility of the rCUDA framework on a real-world application employed in the financial risk industry that can benefit from AaaS in the production setting. The results confirm the feasibility of rCUDA and highlight that rCUDA achieves similar performance compared to CUDA, provides consistent results, and more importantly, allows for a single application to benefit from all the GPUs available in the cluster without loosing efficiency.

AB - How can GPU acceleration be obtained as a service in a cluster? This question has become increasingly significant due to the inefficiency of installing GPUs on all nodes of a cluster. The research reported in this paper is motivated to address the above question by employing rCUDA (remote CUDA), a framework that facilitates Acceleration-as-a-Service (AaaS), such that the nodes of a cluster can request the acceleration of a set of remote GPUs on demand. The rCUDA framework exploits virtualisation and ensures that multiple nodes can share the same GPU. In this paper we test the feasibility of the rCUDA framework on a real-world application employed in the financial risk industry that can benefit from AaaS in the production setting. The results confirm the feasibility of rCUDA and highlight that rCUDA achieves similar performance compared to CUDA, provides consistent results, and more importantly, allows for a single application to benefit from all the GPUs available in the cluster without loosing efficiency.

KW - cloud computing

KW - financial data processing

KW - graphics processing units

KW - parallel architectures

KW - virtualisation

KW - Acceleration-as-a-Service

KW - GPU acceleration

KW - financial application

KW - financial risk industry

KW - remote CUDA

KW - virtualised GPU

KW - Acceleration

KW - Graphics processing units

KW - Hardware

KW - Kernel

KW - Memory management

KW - Servers

KW - CUDA

KW - GPU computing

KW - rCUDA

U2 - 10.1109/eScience.2015.15

DO - 10.1109/eScience.2015.15

M3 - Conference contribution

SN - 978-1-4673-9325-6

SP - 47

EP - 56

BT - 2015 11th IEEE International Conference on e-Science (e-Science)

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - 2015 IEEE 11th International Conference on e-Science (e-Science)

Y2 - 31 August 2015 through 4 September 2015

ER -

Acceleration-as-a-Service: Exploiting Virtualised GPUs for a Financial Application

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this