landscape optimisation....

Shrink the window to a really tiny size to make sure it’s not fill rate.

  • Matt

Shrink the window to a really tiny size to make sure it’s not fill rate

I did what you said and i’ve seen no improvement :frowning: It seem my code is crap (but i think there’s something else i don’t see)

Originally posted by holocaust:
Well i know that i should/could use CVA/VAR/DL and such, but i hoped that, without these techs, even a geforce 2 mx could push more than 24000 tris @ 30 FPS (0.72 Mtris/sec)

Heck no. If you’re just rendering everything in immediate mode, your performance will always be dreadful. To get the most out of your card, you need to batch everything up into vertex arrays.

  • Tom

Have you disabled vsync?

Immediate mode shouldn’t be sub-1M triangles/sec.

  • Matt

holocaust:
I think its pretty clear that your celeron, even overclocked, could be your bottle neck. Remember the celeron has a small cache, 66 mhz bus (not sure if it affects if you have agp), no SSE instructions, and a relativelly low speed.
After upgrading from a P2 350 to a p3 600, I noticed a very big improvement in performance. T&L is great, but you have to feed it fast enough so it can really shine.
In such a case, display lists and/or glDrawElements will help even more.

Try glDrawElements. Its very simple to implement. Some tips:
-minimize glBegin/glEnd or glDrawElements calls (try to put as much as posible in a single call).
-share as much vertices as you can when using vertex arrays.
-Test with display lists, even if you wont use then later, just to see if the driver optimizes the scene better (it does in 90% of the cases).
-Triangle strips is the fastes way to send data to the card. In cases such as spheres, you easilly get 3x the performance of GL_TRIANGLES.

sorry about this post but is there any way i can see the document mentioned in the first post. i dont have excel

Use QuickView