I’m wondering if I’m allowed to use translation, scaling and rotation when I start a new display list with glNewList? Also, would multiple display lists of quads render faster than one single display list containing a long list of quads? Thanks in advance.
Of course! That’s the purpose of the DL.
No! A single DL should be faster, because of lower number of referring to driver and better opportunity for a driver to optimize DL.
This is old (GeForce 6) but this is based on benchmarks/measurement:
slide 29/35 “What about display lists?” shows that for GeForce 6, it is better to have batches of roughly 50000 triangles or less in a display list.
slide/32/35 recommends to have batches of VBO about the size of the pre-T&L (texture and lighting) vertex cache (~64000 vertices).
Again this is old, this is pre unified shaders but the lesson is:
test/benchmark/measure yourself to figure out what is the best solution for your case and your generation of GPU.
There are other results on the same page: