Rubicon freezes occasionally

[jf]

On my machine, Rubicon occasionally freezes. The window either does not repaint at all, or shows a smeared painting mess. Has anybody else run into this? Usually, if I have a level I don't want to lose, I leave it alone and most of the time it comes back minutes later -- I can tell when the small bit of window exposed on the desktop finally repaints. Sometimes, I don't want to wait for it (or I've waited a long, long time) and I kill Firefox from Task Manager and start over. I'm running Windows XP, Sun Java 1.5.0, and Firefox 1.5.0.9. I've launched Internet Explorer to try to avoid it, but it strikes there also.

[jnz]

Yes, it happens to me too.  I think this has something to do with Processing and the way it limits the framerate combined with some sort of clock drift.  In the case of Rubicon the framerate is set to 10fps, so Processing tries to enforce at least 1/10sec delay between frames.  It does this by waiting until the clock is at least 1/10 of a second greater than when the last frame was drawn before it will draw the current frame.

In my case I'm running this on a laptop and I think that SpeedStep or perhaps some weird SMI bug is causing the clock to drift.  Say the clock is running a bit fast for a while and thinks it is time T.  Then something notices this and sets it back to the proper time (probably by reading the hardware clock), say T-120.  Now processing is waiting until time T+0.1 to draw the next frame, which is 2 minutes away, and the whole applet seems locked up until then.

I've also seen a strange parallel of this problem when coming out of standby with Rubicon still running.  The framerate limiter doesn't seem to be doing anything at all and the applet runs as fast as it can draw itself.

I've recently adjusted my Power settings in Control Panel to try to see if that makes any difference.  I haven't had enough time to draw any conclusions yet though.

Are you running a system with SpeedStep or Cool'n'Quiet that ramps the system clock up and down?

[jf]

I have an AMD dual core processor, and if you google "AMD dual core clock drift" you can find all sorts of information on clock drift caused by the TSC on the two cores to drift apart, which looks like it can cause time to go backwards sometimes. I guess that can be an advantage -- the time travel aspect.