Archived

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

Lantz

VAR and vertex format

Recommended Posts

Hey guys. I could use some recommendations. For optimal triangle throughput, what techniques should I use? I''ve looked at VAR and it looks interesting. What do you think? And what is best, separate arrays or an interleaved format? Is there any difference in performance? Any tips or suggestions is welcome. Regards, Lantz

Share this post


Link to post
Share on other sites
try them both, ive found seperate arrays works better.

btw personally ive been able to actually get the theoritcal max performance of my card 20million tris/sec (gf2mx200) with them the downside is the window is about 100x30 pixels big

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

Share this post


Link to post
Share on other sites
VAR is definitely the way to go on nVidia cards. About separate vs. interleaved arrays: nVidia says that they should perform equally well. In reality, it changes with driver versions, sometimes interleaved are faster, sometimes separate are... I''d say, go with the format that suits your engine structure best.

/ Yann

Share this post


Link to post
Share on other sites