Horde3D

I have a scene that consists of several hundred adjoining pieces of procedurally generated
geometry and the whole structure can be pretty stretched/long. Each piece is added
to the Horde scene graph as a separate Model node.
To improve performance I wanted to group some of these mesh-pieces into a
surrounding bounding box to let Horde perform a visibility check with this
surrounding bounding box first instead of checking each piece of geometry anyway.
So I tried to add sequences of 20 adjoining mesh pieces to a Horde groupnode.
But as far as I have seen this groupnode has no separate bounding box...
What's the trick to "combine" several model nodes in a surrounding bounding box?

Right now, I don't think that there is one. The new spatial culling and LOD systems are still a work in progress, but it should eventually support multibounds for this purpose.

_________________
Tristam MacDonald - [swiftcoding]

With the new spatial graph the group node is no more a culling primitive.

Instead of creating many models you could create meshes and attach them to a single model. The model has a bounding box for all its meshes.

Thanks. I will give this a try.

[edit]
Uhm... (how) can I attach my geometry resource to a meshnode instead of a modelnode?
As far as I have seen I can attach a GeometryResource only to a modelnode.

Only models can have a geometry resource but meshes can use a section of that geometry resource. So you need to put the vertex and triangle data of your small objects into a single geometry resource.