On Tue, 2006-08-08 at 10:46 -0700, Carl Worth wrote: > On Tue, 08 Aug 2006 18:01:27 +0300, Jorn Baayen wrote: > > > > > Has this approach been evaluated at all yet? It seems you should be > > > able to re-purpose a bunch of the work you've been doing already to > > > apply to an approach like this. > > > > cairo_fixed_from_double() already accounts for 5% of the cairo > > profile :( > > But that should be easy to get rid of---and we've had various patches > for that in the past. > > I did say when I originally floated this plan that obviously we > wouldn't use the slow conversion code we have in place now. > > I'm planning to put new fixed<->floating-point conversion code into > cairo this week.
Ok, that sounds great! If this works out nicely, the main remaining problem for at least font rendering are the transformations using cairo-matrix. What would you suggest here? Using fixed point matrices internally (for which we'd still need two different cairo-matrix implementations), or having cairo-matrix convert doubles to ints before performing any calculations? Jorn > > -Carl -- OpenedHand Ltd. http://o-hand.com/ _______________________________________________ Performance-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/performance-list
