Jens Grunert wrote:
> Thanks to all feedback to my previous questions wich helped me to get started 
> with my little application. It has simple zoom and each rectangle is 
> selectable. The little program below works good for up to 100.000 rectangles 
> but trying to display about 1 Mio components it's by far too slow.
> 
> I would be very glad if somebody could give me some hints on howto make the 
> programm a lot faster when I want to display many rectangles.
> 
> Here is the program (below is the output from the profiling).
> 

If you are talking about just squares, use opengl.  It will be 
accelerated in hardware on most gfx cards of the past 10 years.

As the code goes:
        a) Handle your selection of a square in a single handle() function in a 
global container like a derived group that contains all squares and 
always return 1.  That will avoid fltk from having to dispatch 100000+ 
handle calls to each square.  On a pick, that function should select and 
set a redraw on the proper square directly.
        b) Iterate through your std::vector using an iterator not an
           at() loop.
         c) precalculate 1/gzoom and use a multiply instead.
        d) Remove all printing commands.


-- 
Gonzalo Garramuño
[EMAIL PROTECTED]

AMD4400 - ASUS48N-E
GeForce7300GT
Xubuntu Gutsy

_______________________________________________
fltk mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk

Reply via email to