Jens Grunert wrote:
> Thank you very much for you answers they are more or less what I feared.
> 
> Unfortunately here stops my knowledge of good c++ design.

        In the case of my recommendations, it's not so much good c++ design
        as it is optimization, and actually veering away from good c++ design
        in favor of speed.

        There may be a way to use good c++ design and still get good speed,
        but it's hard to say cause one has to try several things to see what
        really improves speed.

        My guess is that it's an issue of having the program only consider
        the cells that are viewable, depending on the zoom and scroll.

> Maybe if Greg has the time (and it's not too difficult)
> could you send me some code snippets of the suggestions by you?

        There are probably several ways to do this, but pre-building
        arrays of just the viewable data whenever zoom/scroll is changed
        is probably what it comes down to.

        It's not trivial, and does involve some careful programming,
        something I don't have time to really offer this particular month
        which is particularly busy for me, having just started a corp.

        Possibly someone else can assist you with an example, but I think
        what I posed should work as described. The details are just doing
        a pass at all your data to find boxes that are within the scroll/zoom
        area defined.. you would call this code whenever either are changed,
        and this would update the arrays I mentioned that only contain the
        cells (or pointers to cells) that are within the viewing area, and
        have your draw() routine only use that array to draw the data, instead
        of the master array.

> I'm a hardware ASIC design engineer who wants to write a small and fast 
> viewer to display graphically the cell placement of ASIC standard cells.

        ASIC sounds neat. I've been exposed to FPGA stuff, but it sounds like
        ASIC is the grownup version where one is really designing a dedicated
        chip that will be mass produced at high costs.

        Are you basically writing something that converts HDL into a
        2D chip mask?

> In simply gluing together already existing widgets I got the program working 
> for smaller designs for to 100.000 cells (a surprisingly short time thanks to 
> FLTK and the support).

        Is it possible you can break up your original data into pieces,
        so that the app can focus on separate regions, instead of all the
        data together?

        I think if you can even just quarter your data, it might improve
        viewing times enough to be usable with the app you have without having
        to cull on the fly.

        Possibly you can make a separate view that lets the user initially
        view all the data, where they can drag out a selection area of the
        data they want to look at in detail, so you can cull the data for
        easier/faster viewing in zoomed detail.

        Similar to how the mandelbrot FLTK example program works, where you
        drag out a region, and zoom down..
_______________________________________________
fltk mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk

Reply via email to