A coprocessor card for fast neural network emulation
In this article we present a coprocessor card for PC, designed for the fast emulation of neural network-based systems. The card is composed of 6 custom neural processors arranged using a special parallel architecture. The processor is briefly presented and the structure of the card discussed, along with the architecture's basics. Also presented are the software tools developed around the card, the graphical interface used for drawing the net and the compiler used to translate the drawing into a series of instructions understood by the processor. Last, we present some results regarding a sample application in which the card was tried: in the control of an inverted pendulum.
Unable to display preview. Download preview PDF.
- [Adap91]Adaptive Solutions, Inc., “CNAPs Neurocomputing”, 1991.Google Scholar
- [Cast91]F.Castillo, J.Cabestany, J.M.Moreno, “An Integrated Circuit for Artificial Neural Networks”, in Lecture Notes in Computer Science: Artificial Neural Networks, Ed. A.Prieto, Springer-Verlag, 1991.Google Scholar
- [Cast92]F.Castillo, j.Cabestany, J.M.Moreno, “The Dynamic Ring Architecture”, in Artificial Neural Networks,2, Eds.I.Aleksander, J.Taylor, North-Holland, 1992.Google Scholar
- [Haye88]J.P.Hayes, “Computer Architecture and Organization”, 2nd Ed., McGraw-Hill, 1988.Google Scholar
- [Hech90]R.Hecht-Nielsen, “Neurocomputing”, Addison-Wesley, 1990, p. 342.Google Scholar
- [Inte91]Intel Corporation, “80170NX Electrically Trainable Analog Neural Network”, Experimental Data Sheet, June 1991.Google Scholar
- [More94]J.M.Moreno, “VLSI Architectures for Evolutive Neural Network Models”, PhD.Thesis, Univ. Politècnica de Catalunya, December 1994.Google Scholar
- [Nest94]Nestor Inc., “Ni1000 Development System”, May 1994.Google Scholar