I pushed your commit, thanks. On Sat, Jan 19, 2013 at 12:18:41PM +0000, Micael wrote: > I'm also thinking that it could be improved by passing "1.0 / ratio" > to the "calculate_rr_antialiased" function and multiplying instead of > dividing
Given how complex the rest of calculate_rr_antialiased() is, I doubt it will make much difference, but it may be worth to try and measure if it actually improves the performance. > Wouldn't it be better to fill a struct inside "render_dab_mask" and > pass a pointer to the "calculate_*" functions instead of passing so > many arguments per iteration? Better for what? Readability or performance? If you mean performance, my guess is it would change nothing or make things slightly slower. But only a benchmark will tell for sure. > It would also make it easier to eventually pass more info if needed. Maybe... or it may scatter the definition and use of variables even more around different places in the source. Not sure. As long as the compiler realizes that this stuff still needs to go into registers, I am in favour of whatever is easier to read for someone who doesn't know the code. -- Martin Renold _______________________________________________ Mypaint-discuss mailing list [email protected] https://mail.gna.org/listinfo/mypaint-discuss
