I’m new to openGL programming, but have had a little experience doing 3D graphics without an API. I was always told that fixed point numbers should be used instead of floating point numbers, for performance reasons. I notice that OpenGL uses floats for most things(e.g describing vertexes). Why is this?
Hi there!
Well, that was true in the past, but if we are talking about current system with build in floating point processor it is not a issue anymore.
If you take an Intel P1,P2,P3 for example, it performs the integer multiply/divide inside the FPU.
If you are not sure if you want to take float’s or double’s take float’s for speed (caching, bursting) and double’s only if you need the accuracy.
Kind reagrds,
LG 
Hi there!
Well, that was true in the past, but if we are talking about current system with build in floating point processor it is not a issue anymore.
If you take an Intel P1,P2,P3 for example, it performs the integer multiply/divide inside the FPU.
If you are not sure if you want to take float’s or double’s take float’s for speed (caching, bursting) and double’s only if you need the accuracy.
Kind reagrds,
LG