Design, Implementation, and Test of a Multi-Model Systolic Neural-Network Accelerator

Issue title: Parallel Computing Projects of the Swiss Priority Programme

Article type: Research Article

Authors: CORNU, THIERRY | IENNE, PAOLO | NIEBUR, DAGMAR | THIRAN, PATRICK | VIREDAZ, MARC A.

Affiliations: Swiss Federal Institute of Technology, Centre for Neuro-Mimetic Systems, IN-J Ecublens, CH-1015 Lausanne, Switzerland; e-mail: {cornu.ienne.thiran}@epfl.ch | Now with Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA; e-mail: niebur@telerobotics.jpl.nasa.gov | Now with NEC Research Institute, 4 Independence Way, Princeton, NJ 08540; e-mail: viredaz@research.nj.nec.com

Abstract: A multi-model neural-network computer has been designed and built. A compute-intensive application in the field of power-system monitoring, using the Kohonen neural network, has then been ported onto this machine. After a short description of the system, this article focuses on the programming paradigm adopted. The performance of the machine is also evaluated and discussed.

Journal: Scientific Programming, vol. 5, no. 1, pp. 47-61, 1996