
The Library
Lightweight programmable DSP block overlay for streaming neural network acceleration
Tools
Ioannou, Lenos and Fahmy, Suhaib A. (2020) Lightweight programmable DSP block overlay for streaming neural network acceleration. In: International Conference on Field Programmable Technology, Tianjin, China, 9–13 Dec 2019. Published in: 2019 International Conference on Field-Programmable Technology (ICFPT) ISBN 9781728129440. doi:10.1109/ICFPT47387.2019.00066
|
PDF
WRAP-lightweight-programmable-DSP-block-overlay-streaming-neural-network-acceleration-Fahmy-2019.pdf - Accepted Version - Requires a PDF viewer. Download (695Kb) | Preview |
Official URL: https://doi.org/10.1109/ICFPT47387.2019.00066
Abstract
Implementations of hardware accelerators for neu- ral networks are increasingly popular on FPGAs, due to flex- ibility, achievable performance and efficiency gains resulting from network optimisations. The long compilation time required by the backend toolflow, however, makes rapid deployment and prototyping of such accelerators on FPGAs more difficult. Moreover, achieving high frequency of operation requires sig- nificant low-level design effort. We present a neural network overlay for FPGAs that exploits DSP blocks, operating at near their theoretical maximum frequency, while minimizing resource utilization. The proposed architecture is flexible, enabling rapid runtime configuration of network parameters according to the desired network topology. It is tailored for lightweight edge implementations requiring acceleration, rather than the highest throughput achieved by more complex architectures in the datacenter.
Item Type: | Conference Item (Paper) | ||||||
---|---|---|---|---|---|---|---|
Subjects: | Q Science > QA Mathematics > QA76 Electronic computers. Computer science. Computer software T Technology > TK Electrical engineering. Electronics Nuclear engineering |
||||||
Divisions: | Faculty of Science, Engineering and Medicine > Engineering > Engineering | ||||||
Library of Congress Subject Headings (LCSH): | Field programmable gate arrays, Neural networks (Computer science) , Computer architecture | ||||||
Journal or Publication Title: | 2019 International Conference on Field-Programmable Technology (ICFPT) | ||||||
Publisher: | IEEE | ||||||
ISBN: | 9781728129440 | ||||||
Official Date: | 3 February 2020 | ||||||
Dates: |
|
||||||
DOI: | 10.1109/ICFPT47387.2019.00066 | ||||||
Status: | Peer Reviewed | ||||||
Publication Status: | Published | ||||||
Reuse Statement (publisher, data, author rights): | © 2019 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. | ||||||
Access rights to Published version: | Restricted or Subscription Access | ||||||
Date of first compliant deposit: | 1 November 2019 | ||||||
Date of first compliant Open Access: | 4 November 2019 | ||||||
RIOXX Funder/Project Grant: |
|
||||||
Conference Paper Type: | Paper | ||||||
Title of Event: | International Conference on Field Programmable Technology | ||||||
Type of Event: | Conference | ||||||
Location of Event: | Tianjin, China | ||||||
Date(s) of Event: | 9–13 Dec 2019 | ||||||
Related URLs: |
Request changes or add full text files to a record
Repository staff actions (login required)
![]() |
View Item |
Downloads
Downloads per month over past year