Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!snorkelwacker!bloom-beacon!eru!hagbard!sunic!mcsun!cernvax!chx400!unizh!almassy
From: almassy@ifi.unizh.ch (Nikolaus Almassy)
Newsgroups: comp.ai.neural-nets
Subject: NN simulator benchmarks wanted
Keywords: C, Back-prop, Neural Nets
Message-ID: <1990Sep11.134628.23507@ifi.unizh.ch>
Date: 11 Sep 90 13:46:28 GMT
Organization: University of Zurich, Department of Computer Science
Lines: 20


Hi Neural Experts,

I am looking for a
	STRATEGY FOR DEFINING BENCHMARKS
on the performance of neural network simulators/implementations, i.e.
updates/per second or just a update of a simple connection vs the whole
net (using the learning rule Back Propagation). I think there are some
well known strategies. Especially I would like to know how fast some
common (RCS, ...) simulators on a SPARCststion 1 running SunOS 4.0.3 are
and what exactly is measured. 

Thanks to every reply,
		Nick.

-- 
Nikolaus Almassy
Department for Informatik         |       Tel + 41 - 1 - 257 43 15
University Zurich-Irchel          |       Fax + 41 - 1 - 363 00 35
Winterthurerstrasse 190           |       almassy@ifi.unizh.ch