Some drivers (at least the Intel drivers on Windows) are slow at
incrementally updating a VBO with hundreds of calls to glBufferSubData
every frame. Performance is significantly better if you use
glBufferData(NULL) to tell it to discard all the previous contents, and
then re-upload all the data at once.
Update CVertexBuffer so that GL_DYNAMIC_DRAW/GL_STREAM_DRAW buffers are
handled with the new mechanism. This requires the caller to hold onto
the backing store so it can be re-uploaded when necessary, and needs a
bit more signalling to indicate exactly what needs uploading.
I see an improvement from roughly 60 to 75 fps on Intel HD Graphics
3000, Windows, 1024x768, Siwa Oasis.
This was SVN commit r16241.
Add textured line overlay rendering.
Change terrain height calculations to be triangulation-dependent for
improved accuracy.
Add triangulation-dependent terrain normal function.
Support separate S/T wrap modes for textures.
Rename CVector2D_Maths since it no longer conflicts with simulation
CVector2D.
Coalesce freed chunks in vertex buffers, to avoid excessive
fragmentation.
Add some things to help debug vertex buffer allocation a little.
This was SVN commit r9929.
Use index VBOs in model renderers, for performance.
Be more explicit about static/dynamic VBOs.
Add VBO usage to renderer stats.
Clean up some obsolete unused code.
This was SVN commit r9052.
which is just "class C { void C::f(); }" code, which other compilers
seem to have ignored.
Also removed HAVE_PCH from GCC because our build system doesn't have
PCH. Added necessary headers to make it compile again.
This was SVN commit r4437.