The University of Southampton
University of Southampton Institutional Repository

A method for swift selection of appropriate approximate multipliers for CNN hardware accelerators

A method for swift selection of appropriate approximate multipliers for CNN hardware accelerators
A method for swift selection of appropriate approximate multipliers for CNN hardware accelerators
As convolutional neural networks (CNNs) gain traction for embedded device implementation, there's a burgeoning interest in approximate computing technologies for increasing hardware efficiency. Most of the works in this field focus on proposing novel approximate hardware units and structures, but structured guidance for selecting optimal approximate calculation techniques for CNN accelerators remains scant. This paper introduces a novel error injection technique, leveraging the error rate matrix of approximate multipliers (AxMs), called Error Matrix Based Error Injected (EMEI). This facilitates the swift selection of appropriate AxMs for each PE in the CNN hardware accelerator. In addition, this approach is applied to a MobileNetV2-based CNN model on the CIFAR-10 dataset to demonstrate the performance. Experimental results show that our method adeptly optimises hardware resources by combining AxMs with different accuracy levels while ensuring accuracy. This innovation paves the way for streamlined CNN accelerator designs in embedded systems.
Approximate computing, Approximate multiplier, CNN, CNN hardware accelerator
Sun, Peiyao
e517faec-75c2-43e4-a45e-90f47e80d195
Yu, Haosen
71ded974-69df-4edd-8c1a-ad192b83a27e
Halak, Basel
8221f839-0dfd-4f81-9865-37def5f79f33
Kazmierski, Tomasz
a97d7958-40c3-413f-924d-84545216092a
Sun, Peiyao
e517faec-75c2-43e4-a45e-90f47e80d195
Yu, Haosen
71ded974-69df-4edd-8c1a-ad192b83a27e
Halak, Basel
8221f839-0dfd-4f81-9865-37def5f79f33
Kazmierski, Tomasz
a97d7958-40c3-413f-924d-84545216092a

Sun, Peiyao, Yu, Haosen, Halak, Basel and Kazmierski, Tomasz (2024) A method for swift selection of appropriate approximate multipliers for CNN hardware accelerators. 2024 International Symposium on Circuits and Systems, Resorts World Convention Centre, Singapore. 19 - 22 May 2024. 5 pp .

Record type: Conference or Workshop Item (Poster)

Abstract

As convolutional neural networks (CNNs) gain traction for embedded device implementation, there's a burgeoning interest in approximate computing technologies for increasing hardware efficiency. Most of the works in this field focus on proposing novel approximate hardware units and structures, but structured guidance for selecting optimal approximate calculation techniques for CNN accelerators remains scant. This paper introduces a novel error injection technique, leveraging the error rate matrix of approximate multipliers (AxMs), called Error Matrix Based Error Injected (EMEI). This facilitates the swift selection of appropriate AxMs for each PE in the CNN hardware accelerator. In addition, this approach is applied to a MobileNetV2-based CNN model on the CIFAR-10 dataset to demonstrate the performance. Experimental results show that our method adeptly optimises hardware resources by combining AxMs with different accuracy levels while ensuring accuracy. This innovation paves the way for streamlined CNN accelerator designs in embedded systems.

Text
ISCAS_final_submit_2024018714 - Accepted Manuscript
Restricted to Repository staff only
Request a copy

More information

Accepted/In Press date: 2024
Published date: 20 May 2024
Venue - Dates: 2024 International Symposium on Circuits and Systems, Resorts World Convention Centre, Singapore, 2024-05-19 - 2024-05-22
Keywords: Approximate computing, Approximate multiplier, CNN, CNN hardware accelerator

Identifiers

Local EPrints ID: 488054
URI: http://eprints.soton.ac.uk/id/eprint/488054
PURE UUID: 37ac455a-fafe-4265-86da-4879640f41cf
ORCID for Peiyao Sun: ORCID iD orcid.org/0009-0009-3641-7039
ORCID for Haosen Yu: ORCID iD orcid.org/0000-0002-6174-8579
ORCID for Basel Halak: ORCID iD orcid.org/0000-0003-3470-7226

Catalogue record

Date deposited: 14 Mar 2024 17:31
Last modified: 21 May 2024 02:05

Export record

Contributors

Author: Peiyao Sun ORCID iD
Author: Haosen Yu ORCID iD
Author: Basel Halak ORCID iD
Author: Tomasz Kazmierski

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×