Emulated digital CNN-UM solution of partial differential equations

Zoltán Nagy, Zsolt Vöröshazi, P. Szolgay

Research output: Contribution to journalArticle

38 Citations (Scopus)

Abstract

We present here new cellular neural/non-linear networks (CNN)-based emulated digital architectures specifically designed for the solution of different partial differential equations (PDE). The array structure and local connectivity of the CNN paradigm make it a natural framework to describe the behaviour of locally interconnected dynamical systems. Solution of the PDE is carried out by a spatio-temporal dynamics, which can be computed in real-time on analogue CNN-UM chips, but the accuracy of the solution is low. Additionally, solution of PDEs on a CNN-UM architecture often requires a multi-layer structure and non-linear templates which is partially or not supported on the current analogue VLSI CNN-UM chips. To overcome these obstacles while preserving high computing performance a configurable emulated digital CNN-UM can be used where the main parameters (accuracy, template size and number of layers) are configurable. Additionally, the symmetry of the finite difference operators makes it possible to specialize the emulated digital CNN-UM architecture to solve a specific type of PDE, which results in higher performance. Emulated digital CNN-UM processors use fixed-point numbers to carry out computations, and by decreasing the precision the speed of the computations can be improved. Hence, a simple algorithm is introduced to determine the optimal fixed-point precision and maximize computing performance.

Original languageEnglish
Pages (from-to)445-470
Number of pages26
JournalInternational Journal of Circuit Theory and Applications
Volume34
Issue number4
DOIs
Publication statusPublished - Jul 2006

Fingerprint

Nonlinear networks
Partial differential equations
Partial differential equation
Network architecture
Network Architecture
Template
Chip
Fixed point
Local Connectivity
Analogue
Network Processor
Computing
Difference Operator
Dynamical systems
Multilayer
Finite Difference
High Performance
Dynamical system
Maximise
Paradigm

Keywords

  • Cellular neural networks
  • Emulated digital CNN-UM
  • Ocean modelling
  • Partial differential equations
  • Reconfigurable architectures
  • Retina modelling

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Emulated digital CNN-UM solution of partial differential equations. / Nagy, Zoltán; Vöröshazi, Zsolt; Szolgay, P.

In: International Journal of Circuit Theory and Applications, Vol. 34, No. 4, 07.2006, p. 445-470.

Research output: Contribution to journalArticle

@article{34b8e21ff3bc46f1accea803105c4964,
title = "Emulated digital CNN-UM solution of partial differential equations",
abstract = "We present here new cellular neural/non-linear networks (CNN)-based emulated digital architectures specifically designed for the solution of different partial differential equations (PDE). The array structure and local connectivity of the CNN paradigm make it a natural framework to describe the behaviour of locally interconnected dynamical systems. Solution of the PDE is carried out by a spatio-temporal dynamics, which can be computed in real-time on analogue CNN-UM chips, but the accuracy of the solution is low. Additionally, solution of PDEs on a CNN-UM architecture often requires a multi-layer structure and non-linear templates which is partially or not supported on the current analogue VLSI CNN-UM chips. To overcome these obstacles while preserving high computing performance a configurable emulated digital CNN-UM can be used where the main parameters (accuracy, template size and number of layers) are configurable. Additionally, the symmetry of the finite difference operators makes it possible to specialize the emulated digital CNN-UM architecture to solve a specific type of PDE, which results in higher performance. Emulated digital CNN-UM processors use fixed-point numbers to carry out computations, and by decreasing the precision the speed of the computations can be improved. Hence, a simple algorithm is introduced to determine the optimal fixed-point precision and maximize computing performance.",
keywords = "Cellular neural networks, Emulated digital CNN-UM, Ocean modelling, Partial differential equations, Reconfigurable architectures, Retina modelling",
author = "Zolt{\'a}n Nagy and Zsolt V{\"o}r{\"o}shazi and P. Szolgay",
year = "2006",
month = "7",
doi = "10.1002/cta.363",
language = "English",
volume = "34",
pages = "445--470",
journal = "International Journal of Circuit Theory and Applications",
issn = "0098-9886",
publisher = "John Wiley and Sons Ltd",
number = "4",

}

TY - JOUR

T1 - Emulated digital CNN-UM solution of partial differential equations

AU - Nagy, Zoltán

AU - Vöröshazi, Zsolt

AU - Szolgay, P.

PY - 2006/7

Y1 - 2006/7

N2 - We present here new cellular neural/non-linear networks (CNN)-based emulated digital architectures specifically designed for the solution of different partial differential equations (PDE). The array structure and local connectivity of the CNN paradigm make it a natural framework to describe the behaviour of locally interconnected dynamical systems. Solution of the PDE is carried out by a spatio-temporal dynamics, which can be computed in real-time on analogue CNN-UM chips, but the accuracy of the solution is low. Additionally, solution of PDEs on a CNN-UM architecture often requires a multi-layer structure and non-linear templates which is partially or not supported on the current analogue VLSI CNN-UM chips. To overcome these obstacles while preserving high computing performance a configurable emulated digital CNN-UM can be used where the main parameters (accuracy, template size and number of layers) are configurable. Additionally, the symmetry of the finite difference operators makes it possible to specialize the emulated digital CNN-UM architecture to solve a specific type of PDE, which results in higher performance. Emulated digital CNN-UM processors use fixed-point numbers to carry out computations, and by decreasing the precision the speed of the computations can be improved. Hence, a simple algorithm is introduced to determine the optimal fixed-point precision and maximize computing performance.

AB - We present here new cellular neural/non-linear networks (CNN)-based emulated digital architectures specifically designed for the solution of different partial differential equations (PDE). The array structure and local connectivity of the CNN paradigm make it a natural framework to describe the behaviour of locally interconnected dynamical systems. Solution of the PDE is carried out by a spatio-temporal dynamics, which can be computed in real-time on analogue CNN-UM chips, but the accuracy of the solution is low. Additionally, solution of PDEs on a CNN-UM architecture often requires a multi-layer structure and non-linear templates which is partially or not supported on the current analogue VLSI CNN-UM chips. To overcome these obstacles while preserving high computing performance a configurable emulated digital CNN-UM can be used where the main parameters (accuracy, template size and number of layers) are configurable. Additionally, the symmetry of the finite difference operators makes it possible to specialize the emulated digital CNN-UM architecture to solve a specific type of PDE, which results in higher performance. Emulated digital CNN-UM processors use fixed-point numbers to carry out computations, and by decreasing the precision the speed of the computations can be improved. Hence, a simple algorithm is introduced to determine the optimal fixed-point precision and maximize computing performance.

KW - Cellular neural networks

KW - Emulated digital CNN-UM

KW - Ocean modelling

KW - Partial differential equations

KW - Reconfigurable architectures

KW - Retina modelling

UR - http://www.scopus.com/inward/record.url?scp=33746382549&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33746382549&partnerID=8YFLogxK

U2 - 10.1002/cta.363

DO - 10.1002/cta.363

M3 - Article

AN - SCOPUS:33746382549

VL - 34

SP - 445

EP - 470

JO - International Journal of Circuit Theory and Applications

JF - International Journal of Circuit Theory and Applications

SN - 0098-9886

IS - 4

ER -