I love the SetMaxTexturePoolSize() and setDoPreCompile() options for the DatabasePager, but how does the osgDB::DatabasePager work for large files? (e.g. a proxy Node)? Is the Object automatically split up into a acceleration structure and the according meshes are loaded in and out of the RAM depending on the Frustum? Or is there some kind of mechanism to reduce cache misses? Or just several threads loading the object, presenting it when ready? My C++ is not good enough to understand the full functionality from the source code.
Have a great day,