Archived

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

Arch@on

Adaptivity BinTriTrees and vertexarrays

Recommended Posts

Arch@on    100
I''m using a tight quadtree(produces no cracks) to partition my terrain into smaller peaces for frustum culling, however as you can imagine bare frustum culling doesn''t help much if the view distances are long. My next step is to introduce adaptivity with binary triangle trees and I have not yet decided how. I have been thinking something along the lines where triangles in each one of the "patches" in the quadtree gets reduced using binary triangle tree. Doing this run-time would be extremely CPU heavy and if there were a lot of nodes the speedgain would be almost non-existent or even negative. Right now it seems to be that I have to pick between memory and run-time calculations. Should I just forget about BinTriTrees and go with clod? Can anyone throw any pointers? My current quadtree algorithm is really good, since it is impossible to have cracks and it is decent for collision detection. Also, does anyone have any vertex array experience on bintritrees? To me it seems to be that they would work perfectly with my main method, since I could have on big array and then just use indices to get the what I want to get out of it?

Share this post


Link to post
Share on other sites
Arch@on    100
Also, with BTTs can you get the recurciveness out of the code? With quadtrees your doomned to recursion, which in my opinion isn''t very good. Although I''m not worried about the inlining problems, recursiveness limits me to and makes the code kind of "dull".

It would be much nicer, that instead of recurion i''d have a simple straightforward way to check the structure, pehaps inline assembly loop.

Share this post


Link to post
Share on other sites