Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers

Adarsh Mishra; K. J. Ajith; Kislay Bhatt; Kumar Vaibhav; Vibhuti Duggal

doi:10.1007/978-981-19-1018-0_44

Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers

Adarsh Mishra, K. J. Ajith, Kislay Bhatt, Kumar Vaibhav, Vibhuti Duggal

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

Abstract

Hardware accelerators have evolved into an important tool for meeting the ever-increasing performance demands of modern computation systems. In the modern high-performance computing domain, widely available hardware accelerators are PCIe-attached co-processors to which the host CPU can offload compute-intensive tasks. The goal of this paper is to determine whether FPGAs are a viable option as a hardware accelerator for software programmers, and if so, how their performance compares to existing processors/co-processors such as GPGPUs and CPUs in various types of HPC workloads. We can take advantage of recent advancements in high-level synthesis (HLS) tools, which enable simple programming and debugging for FPGAs. We chose OpenCL for programming because it supports a wide range of devices such as GPUs, FPGAs, DSPs, CPUs, and so on. We are using the Intel Devcloud setup for our experiments because it gives us access to modern Intel FPGAs, which can be used as hardware accelerators in conjunction with other resources such as GPGPUs and multicore processors.

Original language	English (US)
Title of host publication	Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022
Editors	Rashmi Ranjan Rout, Soumya Kanti Ghosh, Prasanta K. Jana, Asis Kumar Tripathy, Jyoti Prakash Sahoo, Kuan-Ching Li
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	511-520
Number of pages	10
ISBN (Print)	9789811910173
DOIs	https://doi.org/10.1007/978-981-19-1018-0_44
State	Published - 2022
Externally published	Yes
Event	3rd International Conference on Advances in Distributed Computing and Machine Learning, ICADCML 2022 - Warangal, India Duration: Jan 15 2022 → Jan 16 2022

Publication series

Name	Lecture Notes in Networks and Systems
Volume	427
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	3rd International Conference on Advances in Distributed Computing and Machine Learning, ICADCML 2022
Country/Territory	India
City	Warangal
Period	1/15/22 → 1/16/22

Keywords

FPGAs
GPUs
HPC
Hardware abstraction
Hardware acceleration
OpenCL
Re-configurability

ASJC Scopus subject areas

Control and Systems Engineering
Signal Processing
Computer Networks and Communications

Access to Document

10.1007/978-981-19-1018-0_44

Cite this

Mishra, A., Ajith, K. J., Bhatt, K., Vaibhav, K., & Duggal, V. (2022). Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers. In R. R. Rout, S. K. Ghosh, P. K. Jana, A. K. Tripathy, J. P. Sahoo, & K-C. Li (Eds.), Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022 (pp. 511-520). (Lecture Notes in Networks and Systems; Vol. 427). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-19-1018-0_44

Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers. / Mishra, Adarsh; Ajith, K. J.; Bhatt, Kislay et al.

Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022. ed. / Rashmi Ranjan Rout; Soumya Kanti Ghosh; Prasanta K. Jana; Asis Kumar Tripathy; Jyoti Prakash Sahoo; Kuan-Ching Li. Springer Science and Business Media Deutschland GmbH, 2022. p. 511-520 (Lecture Notes in Networks and Systems; Vol. 427).

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

Mishra, A, Ajith, KJ, Bhatt, K, Vaibhav, K & Duggal, V 2022, Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers. in RR Rout, SK Ghosh, PK Jana, AK Tripathy, JP Sahoo & K-C Li (eds), Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022. Lecture Notes in Networks and Systems, vol. 427, Springer Science and Business Media Deutschland GmbH, pp. 511-520, 3rd International Conference on Advances in Distributed Computing and Machine Learning, ICADCML 2022, Warangal, India, 1/15/22. https://doi.org/10.1007/978-981-19-1018-0_44

Mishra A, Ajith KJ, Bhatt K, Vaibhav K, Duggal V. Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers. In Rout RR, Ghosh SK, Jana PK, Tripathy AK, Sahoo JP, Li K-C, editors, Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022. Springer Science and Business Media Deutschland GmbH. 2022. p. 511-520. (Lecture Notes in Networks and Systems). doi: 10.1007/978-981-19-1018-0_44

Mishra, Adarsh ; Ajith, K. J. ; Bhatt, Kislay et al. / Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers. Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022. editor / Rashmi Ranjan Rout ; Soumya Kanti Ghosh ; Prasanta K. Jana ; Asis Kumar Tripathy ; Jyoti Prakash Sahoo ; Kuan-Ching Li. Springer Science and Business Media Deutschland GmbH, 2022. pp. 511-520 (Lecture Notes in Networks and Systems).

@inproceedings{3eb8fc6bf60a414f875e61fbe271a900,

title = "Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers",

abstract = "Hardware accelerators have evolved into an important tool for meeting the ever-increasing performance demands of modern computation systems. In the modern high-performance computing domain, widely available hardware accelerators are PCIe-attached co-processors to which the host CPU can offload compute-intensive tasks. The goal of this paper is to determine whether FPGAs are a viable option as a hardware accelerator for software programmers, and if so, how their performance compares to existing processors/co-processors such as GPGPUs and CPUs in various types of HPC workloads. We can take advantage of recent advancements in high-level synthesis (HLS) tools, which enable simple programming and debugging for FPGAs. We chose OpenCL for programming because it supports a wide range of devices such as GPUs, FPGAs, DSPs, CPUs, and so on. We are using the Intel Devcloud setup for our experiments because it gives us access to modern Intel FPGAs, which can be used as hardware accelerators in conjunction with other resources such as GPGPUs and multicore processors.",

keywords = "FPGAs, GPUs, HPC, Hardware abstraction, Hardware acceleration, OpenCL, Re-configurability",

author = "Adarsh Mishra and Ajith, {K. J.} and Kislay Bhatt and Kumar Vaibhav and Vibhuti Duggal",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.; 3rd International Conference on Advances in Distributed Computing and Machine Learning, ICADCML 2022 ; Conference date: 15-01-2022 Through 16-01-2022",

year = "2022",

doi = "10.1007/978-981-19-1018-0_44",

language = "English (US)",

isbn = "9789811910173",

series = "Lecture Notes in Networks and Systems",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "511--520",

editor = "Rout, {Rashmi Ranjan} and Ghosh, {Soumya Kanti} and Jana, {Prasanta K.} and Tripathy, {Asis Kumar} and Sahoo, {Jyoti Prakash} and Kuan-Ching Li",

booktitle = "Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022",

address = "Germany",

}

TY - GEN

T1 - Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers

AU - Mishra, Adarsh

AU - Ajith, K. J.

AU - Bhatt, Kislay

AU - Vaibhav, Kumar

AU - Duggal, Vibhuti

PY - 2022

Y1 - 2022

N2 - Hardware accelerators have evolved into an important tool for meeting the ever-increasing performance demands of modern computation systems. In the modern high-performance computing domain, widely available hardware accelerators are PCIe-attached co-processors to which the host CPU can offload compute-intensive tasks. The goal of this paper is to determine whether FPGAs are a viable option as a hardware accelerator for software programmers, and if so, how their performance compares to existing processors/co-processors such as GPGPUs and CPUs in various types of HPC workloads. We can take advantage of recent advancements in high-level synthesis (HLS) tools, which enable simple programming and debugging for FPGAs. We chose OpenCL for programming because it supports a wide range of devices such as GPUs, FPGAs, DSPs, CPUs, and so on. We are using the Intel Devcloud setup for our experiments because it gives us access to modern Intel FPGAs, which can be used as hardware accelerators in conjunction with other resources such as GPGPUs and multicore processors.

AB - Hardware accelerators have evolved into an important tool for meeting the ever-increasing performance demands of modern computation systems. In the modern high-performance computing domain, widely available hardware accelerators are PCIe-attached co-processors to which the host CPU can offload compute-intensive tasks. The goal of this paper is to determine whether FPGAs are a viable option as a hardware accelerator for software programmers, and if so, how their performance compares to existing processors/co-processors such as GPGPUs and CPUs in various types of HPC workloads. We can take advantage of recent advancements in high-level synthesis (HLS) tools, which enable simple programming and debugging for FPGAs. We chose OpenCL for programming because it supports a wide range of devices such as GPUs, FPGAs, DSPs, CPUs, and so on. We are using the Intel Devcloud setup for our experiments because it gives us access to modern Intel FPGAs, which can be used as hardware accelerators in conjunction with other resources such as GPGPUs and multicore processors.

KW - FPGAs

KW - GPUs

KW - HPC

KW - Hardware abstraction

KW - Hardware acceleration

KW - OpenCL

KW - Re-configurability

UR - http://www.scopus.com/inward/record.url?scp=85136135425&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85136135425&partnerID=8YFLogxK

U2 - 10.1007/978-981-19-1018-0_44

DO - 10.1007/978-981-19-1018-0_44

M3 - Conference contribution

AN - SCOPUS:85136135425

SN - 9789811910173

T3 - Lecture Notes in Networks and Systems

SP - 511

EP - 520

BT - Advances in Distributed Computing and Machine Learning - Proceedings of ICADCML 2022

A2 - Rout, Rashmi Ranjan

A2 - Ghosh, Soumya Kanti

A2 - Jana, Prasanta K.

A2 - Tripathy, Asis Kumar

A2 - Sahoo, Jyoti Prakash

A2 - Li, Kuan-Ching

PB - Springer Science and Business Media Deutschland GmbH

T2 - 3rd International Conference on Advances in Distributed Computing and Machine Learning, ICADCML 2022

Y2 - 15 January 2022 through 16 January 2022

ER -

Assessing the Performance and Suitability of FPGAs as Hardware Accelerator for Software Programmers

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this