Multipass rendering or sthg else ?

Hello everyone,

here is a small question for you:
I need to render the same points in two different ways, one is using a cg program
and the other the fixed function pipeline.
I need to read back some info from the first
rendering pass. I’m wondering:
do you know if I have to send the same points
twice to the graphic card to achieve this ?
Would this “two-pass rendering” (I’m not sure the term applies in this case) could benefit from display lists or VAR (though the points are changing at each iterations) ?
I hope I made it clear…

Thx for all your contributions,


Hi !

Yes you have to send it down the OpenGL pipe line twice and if the points are changing it would be difficult do use display lists or any other optimization to improve performance.


Thx for your answer mikael,

though a little doubt subsist. I read
on a presentation of NV_VERTEX_ARRAY_RANGE
that this functionality was made for a lot
of vertices that are varying in time and
that may be used for multi pass rendering
(hardware T&L). I must admit that I’m surely
missing some points here or misunderstanding
stuff. If anyone could make it clear to me
about the use of VAR…

thx for your contributions !


[EDIT] I browse the forum and I found many stuff on VAR (and also nvidia site). There
still remains smthg I don’t undertsand: when
vertices are modified by hardware T&L, their
values in the VA are not modified ? No ?

[This message has been edited by Escondida (edited 07-04-2003).]