Fast neural net simulation with a DSP processor array

Author

Müller, Urs A. ; Gunzinger, Anton ; Guggenbühl, Walter

Author_Institution

Electron. Lab., Swiss Federal Inst. of Technol., Zurich, Switzerland

Volume

6

Issue

1

fYear

1995

fDate

1/1/1995 12:00:00 AM

Firstpage

203

Lastpage

213

Abstract

This paper describes the implementation of a fast neural net simulator on a novel parallel distributed-memory computer. A 60-processor system, named MUSIC (multiprocessor system with intelligent communication), is operational and runs the backpropagation algorithm at a speed of 330 million connection updates per second (continuous weight update) using 32-b floating-point precision. This is equal to 1.4 Gflops sustained performance. The complete system with 3.8 Gflops peak performance consumes less than 800 W of electrical power and fits into a 19-in rack. While reaching the speed of modern supercomputers, MUSIC still can be used as a personal desktop computer at a researcher´s own disposal. In neural net simulation, this gives a computing performance to a single user which was unthinkable before. The system´s real-time interfaces make it especially useful for embedded applications

Keywords

backpropagation; digital signal processing chips; distributed memory systems; neural nets; 1.4 to 3.8 GFLOPS; 32-b floating-point precision; DSP processor array; MUSIC; backpropagation algorithm; embedded applications; fast neural net simulation; multiprocessor system with intelligent communication; parallel distributed-memory computer; personal desktop computer; real-time interfaces; Backpropagation algorithms; Computational modeling; Computer simulation; Concurrent computing; Digital signal processing; Distributed computing; Intelligent systems; Multiple signal classification; Multiprocessing systems; Neural networks;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/72.363436

Filename

363436