VERY IMPORTANT, please !!!

Hello, i’m using a geforce 2 on a PIII 800 and i just made a test.

I’m using the glcopytexsubimage2d opengl function and my test is this:

I do at each cycle:

glCopyTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, 0, 0, 1024, 768);

and it’s faster than

glCopyTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, 0, 0, 512, 256);

Is that coherent???
The 1024, 768 size is my all buffer. I thougth that if i copy a bigger area, the frame rate should be less than if i copy a smaller area isn’t it?

But it’s faster to copy the 1024, 768 area than the 512, 256…

Is that normal??
Please tell me if it’s normal and u’ll save my life !!!

Thanks in advance


Apparently, I’ve heard that some cards/drivers dont get sub parts of the frame buffer, instead they get the whole thing, then crop it later. This could be the case on yours. IF you get the whole thing, then it doesn’t need to crop, so it works out faster.

possibly anyway.

What is the size of your destination texture?

Relic had the good answer…
Just after posting this message, i saw that i asked ogl to copy a big area in a small texture…

And as anyone know it, when u want to do that, nothing is copied so it’s very fast

Sorry for that stupid question relic, but maybe someone saw it and it helped…

Anyways, have a good day.