Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Programmable Neurocomputing Krste Asanovic
 

Summary: Programmable Neurocomputing
Krste Asanovi´c
MIT Laboratory for Computer Science
200 Technology Square
Cambridge, MA 02139
krste@mit.edu
Appears in The Handbook of Brain Theory and Neural Networks, 2nd edition,
(M.A. Arbib, Ed.), Cambridge, MA: The MIT Press, 2002. (c) The MIT Press
http://mitpress.mit.edu
1 Introduction
General-purpose personal computers and workstations are the most popular computing
platforms used by researchers to simulate artificial neural network (ANN) algorithms.
They provide a convenient and flexible programming environment and technology ad-
vances have been rapidly increasing their performance and reducing their cost. But
large ANN simulations can still overwhelm the capabilities of even the most power-
ful workstations. For example, computations may require greater than 1015 arithmetic
operations and operate on data sets containing several gigabytes of data [5].
Many neural net algorithms are highly parallelizable, allowing simulation speed to
be improved by employing a network of workstations (NOW) [2]. Compared with more
specialized hardware, a NOW can be an expensive solution. Fast network hardware

  

Source: Asanovic, Krste - Computer Science and Artificial Intelligence Laboratory & Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology (MIT)

 

Collections: Computer Technologies and Information Sciences