Hi Guys,

On the topic of double support, in general the OSG should load vertex
data as floats as that's what the vast majority of OpenGL cards in
existence support in a hardware accelerated form.  Even with the
advent of cards that support double maths performance will still be
less than for vertex counter-parts due to bandwidth and memory
concerns.   For a long while yet floats will remain the best way to
manage vertex data for scene graph based applications.

However, driving OpenGL isn't the only thing we want to do with our
apps - often users will be loading data that is geospatial or requires
very high precision and this requires one to use double vertex, and in
some cases it may even make sense to load the data by default as
doubles - shape files are one example where the data is typically
geospatial data that requires high precision for any subsequent maths.
  This type of data you don't want to throw at the graphics card as it
- you'll need to pre-process it first for best performance, you do
have to have the ability to get the data in as doubles first then
process accordingly.

This all points to us needing a standard osgDB::Options hint to tell
plugins how we want double vertex data loaded.  We could just use the
standard Options string, but we could potentially use a dedicated
option.

Robert.
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to