From NVidia OpenGL performance FAQ:
From fastest to slowest:
Can encapsulate data in the most efficient manner for hardware, though they are immutable (i.e. once created, you can’t alter them in any way)
DrawElements/DrawArrays Using wglAllocateMemoryNV(size,0,0,1)
Saves data in video memory, eliminating any bus bottleneck. Very poor read/write access.
DrawElements/DrawArrays Using wglAllocateMemoryNV(size,0,0,.5)
Saves data in AGP (uncached) memory, and allows hardware to pull it directly. Very poor read access, must write sequentially.
DrawElements using Compiled Vertex Arrays (glLockArraysEXT)
Copies locked vertices to AGP memory, so that the hardware can pull it directly. Only certain modes are supported. (t2f/t2f/c4ub/v3f)
Multiple function calls required per primitive results in relatively poor performance compared to other options above.
All Other Vertex Arrays
Must be copied from application memory to AGP memory before the hardware can pull it. Since data can change between data calls, data must be copied every time, which is expensive.
I was pretty surprised to hear that all the other vertex arrays are even slower than immediate mode, but it really makes sense. For static objects use display lists. For dynamic objects check for extensions and try to use them, otherwise use immediate mode.
[This message has been edited by Hude (edited 03-23-2000).]