On Sun, Oct 25, 2015 at 11:12 AM, Nicolas George <geo...@nsup.org> wrote: > Le quartidi 4 brumaire, an CCXXIV, Ganesh Ajjanagadde a écrit : >> Seems like the (x > y) - (x < y) trick works very well: GCC optimizes >> it to branch-less code, and it is overflow safe: >> https://stackoverflow.com/questions/10996418/efficient-integer-compare-function/10997428#10997428 >> >> Will submit patch to convert all qsort comparators (the ones for >> integers) to this form, yielding two benefits: > > Nice. Maybe make it a macro: AV_DIFFSIGN(x, y).
Maybe more like FFDIFFSIGN, I don't think the AV_ prefix is ideal. FF also has the advantage of being consistent with existing FFMIN, FFMAX, FFABS, etc. Nevertheless, this is off-topic for this thread. We can discuss further when I submit the patches. > > Regards, > > -- > Nicolas George > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
