as the subject suggest, i'm looking for some hints about md2 interpolation optimizations.
I'm using the exellent OpenGL MD2Load from LoneRunner.

With my celeron 416 (OC) + geForce II Mx, i currently have 40 fps for 36 models (578 triangles per models) with a small landscape (4096 tris)....

This may seem cool but i expect to lose frames because the ai and move/collision code are not implemented yet

If anyone has good kick-ass piece of info about this i would be glad to share