I’ve made a little prog which loads a model which has about 2500 triangles, and render it using OpenGL (HW accelerated). I render them untextured and unlit. The problem is that i’m only getting around 5 fps. why? I thought OpenGL was fast. At least on my machine (celeron 550, Matrox G400). Where’s the bottleneck here?
If it is that slow, it sounds like it is running in software (i.e. unaccelerated). Check the gl strings to see which OpenGL implementation is being used. If it indicates Matrox’s OpenGL implementation, then you’ll need to examine the pixel format requested, and adjust it until you find a pixel format that supports hardware acceleration.
Heard Matrox cards doesn’t have the best OpenGL drivers. Tried new drivers?
There can also be something in your code. For example, you recalculate sine/cosine for each vertex instead of just once and reuse the value from a variable. Look through your for/while loops and look for heavy calculations that occur once per loop, and see if you can precalculate them as much as possible.
More unnessecary things that can slow down a program is recalculating the projection matrix every frame.
2.5k polygons is no big deal. So unless you are in software, there is most certainly something in your code.
now it runs fine… the problem was the drivers. I tried with the latest drivers, and they didn’t fasten things up, BUT I found some old drivers on a cd (the one I got with the card) and now things run SMOOTH!