HI Andreas,

Why is the value of MaxNumberOfGraphcisContexts so high?  How many graphics contexts do you ever expect to need?  I wouldn't set the value to lower than this otherwise you could end up with threading issues.  If you arn't using multi-threading of the scene graph then this isn't something to worry about, and you can set the MaxNumberOfGraphcisContexts to 1 quite safely as the structure themseles should automatically resize.

Robert.


On 10/6/06, Andreas Ekstrand <[EMAIL PROTECTED]> wrote:
Hi,

I'm trying to minimize the memory consumption of my 12 000 osgText::Text
objects. I noticed that some of the members of the class (and the
classes it uses) are buffered, i.e. there is one copy of the member per
available graphic context. This means that at least on my system there
are 32 copies of each of these members, causing the memory consumption
to rise substantially, even if I only use one context.

To avoid this, I tried to call
osg::DisplaySettings::instance()->setMaxNumberOfGraphicsContexts(1)
temporarily when creating my osgText::Text objects and then set it back
to its original value. As expected it cut away a lot of the allocated
memory and everything worked fine. I have implemented a subclass of
osgText::Text which overrides setPosition and only calls
computePositions with the graphic context IDs currently used, to gain
some (or actually a lot) performance when creating the text objects.
Thanks to this (I guess), it even works fine now if I have several
viewports with different graphic contexts.

My question is, do I risk something doing it this way? It does feel a
bit like cheating... I mean, the text positions are calculated for each
graphic context, but there is only one copy each of the buffered members
of my text objects. Is it a good or bad idea to do this to save memory?

Regards,
Andreas Ekstrand

--
________________________________________________________________________
Andreas Ekstrand
Remograph
Rekrytgatan 10
SE-582 14 Linköping
SWEDEN
Website: http://www.remograph.com
E-mail: [EMAIL PROTECTED]

_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/

_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/

Reply via email to