Archived

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

Lash

Texture sizes

Recommended Posts

What exactly is the advantege to requiring texure sizes of the powers of two? I know those numbers multiply faster, but most older video cards can only do powers of two, so there has to be something else, or am i wrong?

Share this post


Link to post
Share on other sites
1. faster multiplies.

2. hardware alignment. Same deal as with aligning structures for things like SSE. Data reads of natural word sizes are faster.

3. mipmapping - powers-of-2 are always mip-mappable.

4. related to #3, tileability - some chips break textures into tiles to optimise texture accesses (if the scene only uses a quarter of a texture, then just that quarter is uploaded into video memory/texel cache).

5. texture/VRAM management schemes - it''s much easier to write an allocator/manager if you know everything is power-of-2. Less likely to have say 3x17 pixel "unusable" areas of texture memory.

--
Simon O''Connor
Creative Asylum Ltd
www.creative-asylum.com

Share this post


Link to post
Share on other sites