On Wed, Apr 06, 2011 at 02:44:12AM +1200, Ralph Versteegen wrote:
> On 5 April 2011 03:51, James Paige <[email protected]> wrote:
> > On Mon, Apr 04, 2011 at 08:29:14AM -0700, 
> > [email protected] wrote:
> >> http://rpg.hamsterrepublic.com/bugzilla/show_bug.cgi?id=886
> >>
> >> Ralph Versteegen <[email protected]> changed:
> >>
> >>            What    |Removed                     |Added
> >> ----------------------------------------------------------------------------
> >>              Status|NEW                         |RESOLVED
> >>          Resolution|                            |FIXED
> >>             Summary|Random high CPU usage which |Nightmare on Elmo Street 
> >> is
> >>                    |persists until you quit the |slooooooow
> >>                    |game                        |
> >>
> >> --- Comment #1 from Ralph Versteegen <[email protected]> 2011-04-04 
> >> 08:29:14 PDT ---
> >> Opps, looks like I imagined seeing this in Dragon Bustier. It is a NoES 
> >> bug: on
> >> closer inspection, the script running on a timer is actually loading an 
> >> enemy
> >> sprite over and over! It's really bizarre how the game runs fine and then
> >> suddenly the frame rate plummets. And how J_Taylor never noticed this.
> >>
> >
> > It makes sense to me that it takes some time for this to slow things
> > down. Every time a new copy of the sprite is loaded, the old slice
> > is effectively leaked, so sooner or later there are too many of them,
> > and the swapping starts.
> 
> No swapping takes place; the memory usage wasn't increasing
> significantly. The lag must have been just due to drawing thousands of
> identical sprites on the screen, at the same spot.
> 
> > It seems to be a fairly common mistake for new scripters who are
> > cycle of a
> > loop.
> >
> > I think some documentation improvements could help this... But it is
> > hard for me to know exactly what to say.
> 
> Maybe it could also be automatically detected. Such as: at a couple
> check points, say when 500 and when 2000 slices are loaded for the
> first time, run a check to see whether  large numbers of them have
> identical position, parent and extra data, and if so raise a warning.
> We can add a bit to disable the warning, which it might be wise to
> default to on.

That is a good idea. I like that!

---
James
_______________________________________________
Ohrrpgce mailing list
[email protected]
http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org

Reply via email to