Neural network simulation at Warp speed: How we got 17 million connections per second - Robotics Institute Carnegie Mellon University

Neural network simulation at Warp speed: How we got 17 million connections per second

Dean Pomerleau, G. L. Gusciora, David S. Touretzky, and H. T. Kung
Conference Paper, Proceedings of IEEE International Joint Conference on Neural Networks, Vol. 2, pp. 143 - 150, July, 1988

Abstract

A fast back-propagation algorithm for a linear array of processors is described. Results of an implementation of this algorithm on Warp, a ten-processor, programmable systolic array computer, are reviewed and compared with back-propagation implementations on other machines. The current Warp simulator is about eight times faster at simulating the NETtalk text-to-speech network than the fastest back-propagation simulator previously reported in the literature. This fast simulator on Warp is being used routinely in a road-recognition experiment for robot navigation. Results indicate that linear systolic array machines can be efficient neural network simulators. Planned extensions and improvements to the current algorithm are discussed.

BibTeX

@conference{Pomerleau-1988-15414,
author = {Dean Pomerleau and G. L. Gusciora and David S. Touretzky and H. T. Kung},
title = {Neural network simulation at Warp speed: How we got 17 million connections per second},
booktitle = {Proceedings of IEEE International Joint Conference on Neural Networks},
year = {1988},
month = {July},
volume = {2},
pages = {143 - 150},
}