[Kst] Interesting performance results

[George Staikos]

On Wednesday 15 December 2004 17:41, Andrew Walker wrote:
> You don't mention how many points are in each of the curves,
> which would seem to be rather important. How do these numbers
> compare if you points, points and lines, dashed line, etc?
>
> Also, I think to get more meaningful results it might be worth doing
> some of these tests in loops of 1000 (for example) to get a good average.
> Assuming that caching doesn't play a major role it will give a more
> accurate idea of where the time is going.

  Ah sorry, ~15000 points.  Caching is an important factor for us though.  
Performing well only on the very first plot isn't of much use in general, and 
we should try to keep the loops tight enough to be fully cached.

  I'll do more tests with different configurations now.

-- 
George Staikos
KDE Developer			http://www.kde.org/
Staikos Computing Services Inc.	http://www.staikos.net/