Resource Elastic Virtualization for FPGAs using OpenCL

Anuj Vaishnav; Khoa Pham; Dirk Koch; James Garside

doi:10.1109/FPL.2018.00028

Resource Elastic Virtualization for FPGAs using OpenCL

Anuj Vaishnav, Khoa Pham, Dirk Koch, James Garside

Advanced Processor Technology

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

414 Downloads (Pure)

Abstract

FPGAs are rising in popularity for acceleration in all kinds of systems. However, even in cloud environments, FPGA devices are typically still used exclusively by one application only. To overcome this, and as an approach to manage FPGA resources with OS functionality, this paper introduces the concept of resource elastic virtualization which allows shrinking and growing of accelerators in the spatial domain with the help of partial reconfiguration. With this, we can serve multiple applications simultaneously on the same FPGA and optimize the resource utilization and consequently the overall system performance. We demonstrate how an implementation of resource elasticity can be realized for OpenCL accelerators along with how it can achieve 2.3x better FPGA utilization and 49% better performance on average while simultaneously lowering waiting time for tasks.

Original language	English
Title of host publication	28th International Conference on Field Programmable Logic and Application (FPL)
Place of Publication	Dublin, Ireland
Number of pages	8
DOIs	https://doi.org/10.1109/FPL.2018.00028
Publication status	Published - 6 Dec 2018
Event	28th International Conference on Field Programmable Logic and Application - TRINITY COLLEGE, Dublin, Ireland Duration: 27 Aug 2018 → 31 Aug 2018 Conference number: 28 https://fpl2018.org/

Publication series

Name	International Conference on Field Programmable Logic and Applications
Publisher	IEEE
ISSN (Print)	1946-147X
ISSN (Electronic)	1946-1488

Conference

Conference	28th International Conference on Field Programmable Logic and Application
Abbreviated title	FPL
Country/Territory	Ireland
City	Dublin
Period	27/08/18 → 31/08/18
Internet address	https://fpl2018.org/

Keywords

FPGA
Virtualization
Resource Elasticity
Space-time Domain Scheduling
OpenCL
HLS

Access to Document

10.1109/FPL.2018.00028

Resource Elastic Virtualization for FPGAs using OpenCL for FPLAccepted author manuscript, 1.06 MB

Cite this

@inproceedings{ab8417c8f76d44f8bf8608f1985cdf8e,

title = "Resource Elastic Virtualization for FPGAs using OpenCL",

abstract = "FPGAs are rising in popularity for acceleration in all kinds of systems. However, even in cloud environments, FPGA devices are typically still used exclusively by one application only. To overcome this, and as an approach to manage FPGA resources with OS functionality, this paper introduces the concept of resource elastic virtualization which allows shrinking and growing of accelerators in the spatial domain with the help of partial reconfiguration. With this, we can serve multiple applications simultaneously on the same FPGA and optimize the resource utilization and consequently the overall system performance. We demonstrate how an implementation of resource elasticity can be realized for OpenCL accelerators along with how it can achieve 2.3x better FPGA utilization and 49% better performance on average while simultaneously lowering waiting time for tasks.",

keywords = "FPGA, Virtualization, Resource Elasticity, Space-time Domain Scheduling, OpenCL, HLS",

author = "Anuj Vaishnav and Khoa Pham and Dirk Koch and James Garside",

year = "2018",

month = dec,

day = "6",

doi = "10.1109/FPL.2018.00028",

language = "English",

series = "International Conference on Field Programmable Logic and Applications",

publisher = "IEEE",

booktitle = "28th International Conference on Field Programmable Logic and Application (FPL)",

note = "28th International Conference on Field Programmable Logic and Application, FPL ; Conference date: 27-08-2018 Through 31-08-2018",

url = "https://fpl2018.org/",

}

Vaishnav, A, Pham, K, Koch, D & Garside, J 2018, Resource Elastic Virtualization for FPGAs using OpenCL. in 28th International Conference on Field Programmable Logic and Application (FPL). International Conference on Field Programmable Logic and Applications, Dublin, Ireland, 28th International Conference on Field Programmable Logic and Application, Dublin, Ireland, 27/08/18. https://doi.org/10.1109/FPL.2018.00028

Resource Elastic Virtualization for FPGAs using OpenCL. / Vaishnav, Anuj; Pham, Khoa; Koch, Dirk et al.
28th International Conference on Field Programmable Logic and Application (FPL). Dublin, Ireland, 2018. (International Conference on Field Programmable Logic and Applications).

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

TY - GEN

T1 - Resource Elastic Virtualization for FPGAs using OpenCL

AU - Vaishnav, Anuj

AU - Pham, Khoa

AU - Koch, Dirk

AU - Garside, James

N1 - Conference code: 28

PY - 2018/12/6

Y1 - 2018/12/6

N2 - FPGAs are rising in popularity for acceleration in all kinds of systems. However, even in cloud environments, FPGA devices are typically still used exclusively by one application only. To overcome this, and as an approach to manage FPGA resources with OS functionality, this paper introduces the concept of resource elastic virtualization which allows shrinking and growing of accelerators in the spatial domain with the help of partial reconfiguration. With this, we can serve multiple applications simultaneously on the same FPGA and optimize the resource utilization and consequently the overall system performance. We demonstrate how an implementation of resource elasticity can be realized for OpenCL accelerators along with how it can achieve 2.3x better FPGA utilization and 49% better performance on average while simultaneously lowering waiting time for tasks.

AB - FPGAs are rising in popularity for acceleration in all kinds of systems. However, even in cloud environments, FPGA devices are typically still used exclusively by one application only. To overcome this, and as an approach to manage FPGA resources with OS functionality, this paper introduces the concept of resource elastic virtualization which allows shrinking and growing of accelerators in the spatial domain with the help of partial reconfiguration. With this, we can serve multiple applications simultaneously on the same FPGA and optimize the resource utilization and consequently the overall system performance. We demonstrate how an implementation of resource elasticity can be realized for OpenCL accelerators along with how it can achieve 2.3x better FPGA utilization and 49% better performance on average while simultaneously lowering waiting time for tasks.

KW - FPGA

KW - Virtualization

KW - Resource Elasticity

KW - Space-time Domain Scheduling

KW - OpenCL

KW - HLS

U2 - 10.1109/FPL.2018.00028

DO - 10.1109/FPL.2018.00028

M3 - Conference contribution

T3 - International Conference on Field Programmable Logic and Applications

BT - 28th International Conference on Field Programmable Logic and Application (FPL)

CY - Dublin, Ireland

T2 - 28th International Conference on Field Programmable Logic and Application

Y2 - 27 August 2018 through 31 August 2018

ER -

Resource Elastic Virtualization for FPGAs using OpenCL

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this