On Wed, Mar 11, 2009 at 11:59 AM, Christian Packmann <Christian.Packmann@xxxxxx> wrote: > I /think/ this limit to the x-loop should be no performance problem in most > cases. IIUC the length of the loop will depend on the currently processed > clipping rectangle, this should normally be several pixels wide. But if the > clipping rectangles get very small horizontally, the vector routines will > perform worse than the C++ code due to the overhead of the function call and > internal setup. If you think this will be a problem, I'll change the scope > of the routine to include the y-loop. > Just a minor nitpick that I'm sure you're probably already aware of, be sure to leave the C++ versions around for non-x86 platforms :) Regards, Rene