Publication detail

Multiobjective evolution of approximate multiple constant multipliers

PETRLÍK, J. SEKANINA, L.

Original Title

Multiobjective evolution of approximate multiple constant multipliers

English Title

Multiobjective evolution of approximate multiple constant multipliers

Type

conference paper

Language

en

Original Abstract

Multiple constant multiplier (MCM) is a digital circuit which multiplies its single input by N constants. As MCMs are composed of adders and shifters, their implementation cost is relatively low. In this paper, we propose a method for design of approximate multiple constant multipliers where the requirement on functional equivalence between the specification and implementation is relaxed in order to further reduce the area on a chip or minimize delay. The proposed method is based on multiobjective Cartesian Genetic Programming. It provides many trade-off solutions among accuracy, area and delay. 

English abstract

Multiple constant multiplier (MCM) is a digital circuit which multiplies its single input by N constants. As MCMs are composed of adders and shifters, their implementation cost is relatively low. In this paper, we propose a method for design of approximate multiple constant multipliers where the requirement on functional equivalence between the specification and implementation is relaxed in order to further reduce the area on a chip or minimize delay. The proposed method is based on multiobjective Cartesian Genetic Programming. It provides many trade-off solutions among accuracy, area and delay. 

Keywords

multiple constant multiplier, approximate computing

RIV year

2013

Released

08.04.2013

Publisher

IEEE Computer Society

Location

Brno

ISBN

978-1-4673-6133-0

Book

IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems 2013

Edition

NEUVEDEN

Edition number

NEUVEDEN

Pages from

116

Pages to

119

Pages count

4

URL

BibTex


@inproceedings{BUT103446,
  author="Jiří {Petrlík} and Lukáš {Sekanina}",
  title="Multiobjective evolution of approximate multiple constant multipliers",
  annote="Multiple constant multiplier (MCM) is a digital circuit which multiplies its
single input by N constants. As MCMs are composed of adders and shifters, their
implementation cost is relatively low. In this paper, we propose a method for
design of approximate multiple constant multipliers where the requirement on
functional equivalence between the specification and implementation is relaxed in
order to further reduce the area on a chip or minimize delay. The proposed method
is based on multiobjective Cartesian Genetic Programming. It provides many
trade-off solutions among accuracy, area and delay. ",
  address="IEEE Computer Society",
  booktitle="IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems 2013",
  chapter="103446",
  doi="10.1109/DDECS.2013.6549800",
  edition="NEUVEDEN",
  howpublished="print",
  institution="IEEE Computer Society",
  year="2013",
  month="april",
  pages="116--119",
  publisher="IEEE Computer Society",
  type="conference paper"
}