• Advertisement

Archived

This topic is now archived and is closed to further replies.

#of terrain vertices at which displaylists will speed

This topic is 5503 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hi, I have a 256x256 vertices textured heightmap. Nothing else. And the exact same framerate with or without using a displaylist. (am doing no culling yet). At what point would a display list speed things up? I've tried both single texture and 2-textured (using the arb extension). Does 15 & 10 fps for single & 2 texture sound reasonable for this number of vetices? I'm on a 1.6 P4, ATI Radeon 8500, 512 MB DDR Thanx Also, when you do culling, how many rendered vertices do you generally aim for? [edited by - Cat_B on February 1, 2003 2:14:43 AM]

Share this post


Link to post
Share on other sites
Advertisement
quote:
Original post by Cat_B
Hi,

I have a 256x256 vertices textured heightmap. Nothing else. And the exact same framerate with or without using a displaylist. (am doing no culling yet). At what point would a display list speed things up? I''ve tried both single texture and 2-textured (using the arb extension). Does 15 & 10 fps for single & 2 texture sound reasonable for this number of vetices? I''m on a 1.6 P4, ATI Radeon 8500, 512 MB DDR

Thanx


Also, when you do culling, how many rendered vertices do you generally aim for?

[edited by - Cat_B on February 1, 2003 2:14:43 AM]


definetly not.. i''m using a 1800++ 256 ddr s3 32 mb video card and with a heightmap of 1024 by 1024 .. using glVertex to draw i get about 40 fps

Share this post


Link to post
Share on other sites
It seems to me that you are doing a lot of equations every frame.The idea is to store everything and just call it every frame instead of calculating it!!!Try to store the vertices in float array(float vertices[3 * num_vertices]) and then use a mixture of OpenGL vertex and texcoord pointers and display lists.The 3DS loader of my engine works this way and i promise that the results will be outstanding!!!

The PAIN is coming...this summer!!!In cinemas everywhere.

Share this post


Link to post
Share on other sites
I get ~40 FPS on a XP1700 and a GF2 GTS. A heightmap of 256*256, yet no texturing or any form of LOD (as I haven''t implemented that yet)

Share this post


Link to post
Share on other sites
with DLs (as well as with vertex arrays) u dont wanna go over a limit (usually about 4000 vertices) going over this limit will severly affect performance.
thus u need to break the terrain intio smaller chunks

http://uk.geocities.com/sloppyturds/kea/kea.html
http://uk.geocities.com/sloppyturds/gotterdammerung.html

Share this post


Link to post
Share on other sites
edit: Holy Chit- I just heard about the Columbia!!!! (and here I'm worried about vertices..)


Thank you for the tips...

Mihail121:
I'm putting them into an array, thanx- not finished. BUT, I really wasn't doing that many calculations: my poly's are only quads so I just have a double for loop and I add on the map spacing each time- that's it. Well, I was also calculating the normals, but I've since pre-calculated them and put them in an array and I see no difference in frame rate.

Structural:
with no texturing, I get 24 fps. But I have lighting-if I omit glNormal(), I get 30 fps. Are you using lighting? Also I am just generating quads, not quad strips. And each frame I transform (for cruising around):
glTranslatef(-xpos,-ypos,-zpos);
glRotatef(rotx,1.0f,0.0f,0.0f);
glRotatef(-roty,0.0f,1.0f,0.0f);
glRotatef(-rotz,0.0f,0.0f,1.0f);


zedzeek:
Thank you for the tip on max vertices in a display list and vertex array. I was thinking an array of 3x256x256=196,608 (.8MB)floats was a little large..., not sure- I've never used arrays this big.



[edited by - Cat_B on February 1, 2003 4:44:41 PM]

Share this post


Link to post
Share on other sites

  • Advertisement