[Kstars-devel] Bugfixes in branch

[James Bowlin]

On Mon June 16 2008, Akarsh Simha wrote:
> However, these fixes seem to have added some overhead on slew times
> at low zoom factors. I was unable to reproduce the freezing that
> James experienced at high zoom, but find it jerky at low zoom.

I can not reproduce the freezing I had encountered earlier.  I have not
seen it since I updated my kdelibs.

> As indicated by the timings (obtained by pressing 'B'), it takes
> about 225ms for a single redraw at intermediate zoom levels. This is
> bad, but I see quite a few places where I can try improving it, and
> will try to do the same tomorrow.

I can reproduce the jerkyness at intermediate zoom levels (19 visible 
trixels) when I have the "faint limit zoomed out" set to 6.0.  But when 
I reduce it to 5.0, things become smooth again.  The 'B' key reports 
draw times of:

     515 ms at mag 9.0  (maglim = 12.9)
     189 ms at mag 6.0  (maglim =  9.9)
      35 ms at mag 5.0  (maglim =  8.9)

But at mag 9.0 there is almost more star than background.  To properly
measure performance I think we really need to count the number of stars
being drawn and look at the stars/ms rate.

The zoom level can be deduced from the difference between the "faint
limit zoomed out" setting and the actually maglim.

I have no performance problems at any zoom level if the "faint limit 
zoomed out" is set to 5.0.  Also, the star field seems to be reasonably 
dense at this setting.  This makes me suspect that there may be no 
performance problems with your code and the performace problems you 
have encountered are only due to trying to draw "too many" stars.


-- 
Peace, James