[kepler-dev] More performance improvements

Edward A. Lee eal at eecs.berkeley.edu
Sat Sep 30 05:48:34 PDT 2006


I'm checking in changes that improve performance dramatically
and second and subsequent runs of a model.  In particular, unless
you have actors that alter the structure of the model during a
run or otherwise obtain write access on the workspace, then
on the second and subsequent runs, there is no need to re-do
type resolution and scheduling.

To accomplish this, I changed the Manager to no longer
increment the workspace version in wrapup(), and I modified
AtomicActor and CompositeActor so that they no longer create
new Receivers in preinitialize(). This then required
fixing the Receivers in CT and DT to correctly implement reset().

Ray:

Note that you have to use our version of the Publisher/Subscriber
to get this improvement, since your version reconstructs connections
in preinitialize(), which forces all preinitialize() activity to
recur.

Edward


------------
Edward A. Lee
Professor, Chair of EECS
231 Cory Hall, UC Berkeley, Berkeley, CA 94720-1770
phone: 510-642-0253 or 510-642-0455, fax: 510-642-2845
eal at eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal  



More information about the Kepler-dev mailing list