> There are a number of different factors that come into play: image
> size, filters being used, etc.
> The one big difference that you can track is GPU vs. CPU by
> checking for the ARB fragment programing capability. But to state
> again, just because a filter runs in software, doesn't mean that it
> is slow.
> So you would have to check with your use cases on different
> hardware. You might be surprised how well filters run in software.
It would be nice to test, but I don't know if I can. Can you give me
pointers on testing for ARB fragment programming capability?
Yahoo! Mac Engineer
ballard@yaho... Phone: (408) 349-6693
Cell: (408) 660-5028
Yahoo! ID: k_r_ballard