The TInMANN VLSI chip
Authors:
- T. PhanTPT. Phan
- D.S. ReevesDRD.S. Reeves
- D.E. Van Den BoutDBD.E. Van Den Bout
- 1 May 1992
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks
- Vol. 3 (3) , 375-384
Abstract
A massively parallel, all-digital, stochastic architecture-TInMANN-is described which acts as a Kohonen self-organizing feature map. A VLSI design is shown for a TInMANN neuron which fits within a small, inexpensive MOSIS TinyChip frame, yet which can be configured to build networks of arbitrary size. The neuron operates at a speed of 15 MHz, making it capable of processing 195 000 three-dimensional training examples per second. Three man-months were required to synthesize the neuron and its associated level-sensitive scan logic using the OASIS silicon compiler. The ease of synthesis allowed many performance trade-offs to be examined, while the automatic testability features of the compiler helped the designers achieve 100% fault coverage of the chip. These factors served to create a fast, dense, and reliable neural chip.Keywords
This publication has 15 references indexed in Scilit:
- An 11-million Transistor Neural Network Execution Engine1991
- A digital architecture employing stochasticism for the simulation of Hopfield neural netsIEEE Transactions on Circuits and Systems, 1989
- An electrically trainable artificial neural network (ETANN) with 10240 'floating gate' synapses1989
- Adding a conscience to competitive learning1988
- A stochastic architecture for neural nets1988
- VLSI implementation of a neural network memory with several hundreds of neuronsAIP Conference Proceedings, 1986
- Vector quantizationIEEE ASSP Magazine, 1984
- CAzM: A circuit analyzer with macromodelingIEEE Transactions on Electron Devices, 1983
- Self-organized formation of topologically correct feature mapsBiological Cybernetics, 1982
- An Algorithm for Vector Quantizer DesignIEEE Transactions on Communications, 1980