androula@ecn.purdue.edu (Ioannis Androulakis) (03/16/91)
S. Aslam and C. Gear, in their Tech. Report UIUCDCS-R-89-1525,
do not seem to encounter expicitely the effect of using
utdated information in their asynchronous integration of ODE's.
It is known that in the case of non-linear optimization, there
is a relation between the step size used by each processor and
the amount of asynchronism that can be tolerated by the system.
The fact that Aslam and Gear, do not impose any similar restrictions
on their integration means that asynchronous integration can tolerate
any amount of delays, as long as they are finite, without affecting
any of the integration papameters (i.e. step size ) ?
Thank you,
ioannis
androula@ecn.purdue.edu
--
=========================== MODERATOR ==============================
Steve Stevenson {steve,fpst}@hubcap.clemson.edu
Department of Computer Science, comp.parallel
Clemson University, Clemson, SC 29634-1906 (803)656-5880.mabell