well, 3dmark2000 uses triple buffering as default and 3dmark2001 uses double buffering as default. how it affects frame rate, you just have to experiment with your system. In theory, using triple buffering will lower your frame rate if that additional back buffer created in the video memory forces texture to be stored in system memory(AGP texturing).
The reason they use triple buffering in benchmark(with V-Sync off) is that they think it gives more reliable benchmark result since redering process is less frequently stalled cuz the 3D chip can just write to the 2nd back buffer when the 1st back buffer is already written and can not be flipped to the front buffer fast enuf(swapping of front and back buffer).
In real world situation, triple buffering would be used when you enabled V-Sync(cuz your frame rate is much higer than refresh rate so as to give you image tearing and flickering). Since front buffer waits for the V-Sync interval even when back buffer is already written and needs to be flipped. so 3D chip just writes to the 2nd back buffer so as not to stall redering process.
Using triple buffering with V-Sync off may slow the card down(as i mentioned earlier) but very unlikely to give you higer frame rate. With V-Sync off, frame buffer is flipped as soon as back buffer is finished so 3D chip can use what is previously front buffer as back buffer to write the next scene. With double buffering when frame rate is very high, there may be instances when flipping doesn't happen on right time so the 3D chip is not able to start writing the next scene to the back buffer immediately. Triple buffering (Usinn 2 back buffers) may alleviate this. But i higly doubt this translates into higer frame rates in real world situation(although it may be theoretically possible)
You just have to experiment with your system.
<P ID="edit"><FONT SIZE=-1><EM>Edited by DDR64 on 04/21/02 04:18 AM.</EM></FONT></P>
Facebook
Twitter
Instagram