tutorials about using t&l efficiently?

Hi guys,

does anyone of you know a good tutorial about how to use T&L? I know, that if i use some gl functions like MultMatrix i do use T&L. I also know, that there is no extension or other mechanism i have to use to activate T&L.
However i really don´t know, what i can improve my programs with T&L and how exactly i do this. So i would like to know which functions are hardware-accelerated and how i can use them as efficiently as possible.

Thanks in advance,

If you’re talking about HARDWARE t&l then there are three (four) ways to optimize for it:

  1. Use DrawRangeElements() to specify geometry.

  2. Use LockArraysEXT() if your arrays are sufficiently small (say, < 2000 verts)

  3. Use vendor-specific geometry transfer functionality, such as VertexArrayRange for nVIDIA and VertexArrayObject for ATI.

(4) Use display lists)

I know, that if i use some gl functions like MultMatrix i do use T&L.

No, those functions are not using hw T&L.
Only the transformations for every vertex itself are done in hardware (AFAIK)

With glMult/LoadMatrix (and implicit with glRotatef, glTranslatef, etc) you’re only giving the GPU a matrix that should be used for those transformations.

I think I read that MultMatrix is executed on the CPU and the result is transfered to the GPU.

This is probably because MultMatrix is not used often enough and is not worthy of being in hardware.

PS: visit ati.com and nvidia.com for documents about optimizing (not just T&L optimizing!)