QUOTE(dwgregg3d @ 06/30/08, 09:07 PM) [snapback]287677[/snapback]
Just seems like dividing up textures to optimum size is something that the software should be smart enough to do.
It is, your your opengl viewport does not do larger than that. The software renderers dont have any problems with bigger texures. The viewport is a preview after all so no idea in choking it up unnesseserily. (remeber textures use up memory in squares, so a 8K texture uses up 4 times the memory of a 4k texture. And as a result it uses without alpha EXACTLY 201,326,592 bytes of video memory, no ifile size does not count for lookup it must be loaded. This essentially means theres room for exactly 1-2 8k textures on a quadro fx 3700 for example, ofcourse the gfx card is a bit smarter then that and it never loads your full size texture entirely since it KNOWS you don't use it all)
It is however a totaly different thing do you actually need the full resolution for anything. I mean its not like you will be using every single resulting textel anyway.
So you should turn it to map or bot (for gfx card dds) so the texel lookup can be saner. And thus the rendrer can load tiles on demand. Normal formats arent ordered for this kind of look up so they must be in memory
PS: so to recap the limitation is there for a reason, id suspect thya will pump it to 16k in a while but still.