Dear ptolemyers,
I'm trying to simulate a large system of ODEs, namely a
continuous-timed, continuous-valued Hopfield network with a number of
nodes N of the order of 100. Pratically this requires the creation of N
integrators and N other actors calculating the first derivative of the
node's state to be fed into the integrators.
For a systems with 100 nodes (AKA neurons), the number of ports which
have to created is about 10,000, which needs quite a few Mbytes of RAM.
Is there a way to get around this problem?
I've tried to keep the state of the network in an array common to all
actors, and having it updated in the fire() methods of the actors, but
that's really a poor man's solution, as it disrupts ptolemy's model of
computation. Am I right?
Thank you,
Alessandro
_/ _/ _/ _/ Alessandro Usseglio Viretta / Tel +41 1
6353066
_/ _/_/ _/ _/ Institute of Neuroinformatics / Fax +41 1
6353053
_/ _/ _/ _/ _/ University/ETH Zurich /
_/ _/ _/_/ _/ Winterthurerstr. 190 /
_/ _/ _/ _/ CH-8057 Zurich /
www.ini.unizh.ch/~ale
----------------------------------------------------------------------------
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
