Hello Black,
Black wrote:
> I am developing a 2D app that consists of somewhere in the
> neighborhood of 2000 "tiles", where a tile is essentially a quad with
> a border with some text or and image on it. The tiles will need to be
> individually selectable and movable. Performance is an issue, so I
> thought I would run this past the list to see if anyone had an opinion.
>
> As I see it currently, there are two ways that I could go about this:
> a) each tile is in its own branch of the scene graph, complete with
> transformation node, or b) all of the quads are stored in a single
> geometry node and all of the movement is handled externally. The
> advantages of option a are pretty clear from a coding perspective -
> encapsulation, easier selection, easier grouping of tiles and
> contents, etc. However, if my understanding is correct, what I lose is
> the intelligent grouping of like primitives (i.e., 2000 glBegin/glEnd
> pairs instead of 1).
>
> So, is my analysis correct? And if you were in my shoes, which path
> would you take?
there is one more thing to consider: with one geometry you are
restricted to one material and that might cause problems depending on
what your tiles actually show.
Other than that it is true that you essentially get many begin/end
pairs, but I'd expect the larger overhead coming from the traversal of
the 2000 nodes and the potentially many material changes (OpenSG will
minimize those, if you use the same material for more than one geometry).
Hope it helps,
Carsten
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
