Publication Details

Libraries of Approximate Circuits: Automated Design and Application in CNN Accelerators

MRÁZEK Vojtěch, SEKANINA Lukáš and VAŠÍČEK Zdeněk. Libraries of Approximate Circuits: Automated Design and Application in CNN Accelerators. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 10, no. 4, 2020, pp. 406-418. ISSN 2156-3357.
Czech title
Knihovny přibližně počítajících obvodů: automatický návrh a aplikace v CNN
Type
journal article
Language
english
Authors
Keywords

approximate circuit, genetic programming, convolutional neural network, hardware accelerator, optimization

Abstract

Libraries of approximate circuits are composed of fully characterized digital circuits that can be used as building blocks of energy-efficient implementations of hardware accelerators. They can be employed not only to speed up the accelerator development but also to analyze how an accelerator responds to introducing various approximate operations. In this paper, we present a methodology that automatically builds comprehensive libraries of approximate circuits with desired properties. Target approximate circuits are generated using Cartesian genetic programming. In addition to extending the EvoApprox8b library that contains common approximate arithmetic circuits, we show how to generate more specific approximate circuits; in particular, MxN-bit approximate multipliers that exhibit promising results when deployed in convolutional neural networks. By means of the evolved approximate multipliers, we perform a detailed error resilience analysis of five different ResNet networks. We identify the convolutional layers that are good candidates for adopting the approximate multipliers and suggest particular approximate multipliers whose application can lead to the best trade-offs between the classification accuracy and energy requirements. Experiments are reported for CIFAR-10 and CIFAR-100 data sets.

Published
2020
Pages
406-418
Journal
IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 10, no. 4, ISSN 2156-3357
Book
IEEE Journal on Emerging and Selected Topics in Circuits and Systems
Publisher
IEEE Circuits and Systems Society
DOI
UT WoS
000598110700002
EID Scopus
BibTeX
@ARTICLE{FITPUB12372,
   author = "Vojt\v{e}ch Mr\'{a}zek and Luk\'{a}\v{s} Sekanina and Zden\v{e}k Va\v{s}\'{i}\v{c}ek",
   title = "Libraries of Approximate Circuits: Automated Design and Application in CNN Accelerators",
   pages = "406--418",
   booktitle = "IEEE Journal on Emerging and Selected Topics in Circuits and Systems",
   journal = "IEEE Journal on Emerging and Selected Topics in Circuits and Systems",
   volume = 10,
   number = 4,
   year = 2020,
   ISSN = "2156-3357",
   doi = "10.1109/JETCAS.2020.3032495",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/12372"
}
Back to top