[Bf-committers] Sunday meeting, nov 7 2004
Nick Winters
nickrwinters at gmail.com
Sat Nov 13 20:54:52 CET 2004
it appears that attempting to use a fixed time for physics can cause
physics to take more an more time each frame.
see:
http://mywebpage.netscape.com/YinYangEvilSane/BlenderStuff/boxstack2.blend
[it may depend somewhat on my system, I have an athlon 2000+ and run
windows xp, with a radeon 9600se]
so, what I'd guess is happening is that on one frame the swap is, say,
half a second
the next frame uses this to determine how many physics iterations to
do, so that would be 0.5*30 = 15. If this frame takes more than half
a second to do [the physics alone could cause that] then the next
frame will do even more physics iterations.
and so on, and so forth
perhaps at each frame there should be a maximum number of physics
iterations done [say, 3] so that the framerate doesn't get slower and
slower and slower, but just gets slow [3 iterations/frame would be
expected at 10fps]
[it wouldn't surprize me if this is too late to fix now, I probably
should file a bug report anyway]
On Mon, 8 Nov 2004 23:25:36 +1300, Kester Maddock
<christopher.maddock.1 at uni.massey.ac.nz> wrote:
> On Monday 08 November 2004 05:35, Ton Roosendaal wrote:
> > 1) Game engine
> > - Kester; don't forget a nice release log for this release!
>
> 1. Fixed time system. Logic & physics no longer depend on frame rate. Logic
> happens at 30Hz and physics at 60Hz, and this is configurable with Python.
> Some logic bricks (eg IPO, Action, camera, track to actuators) still happen
> at framerate for nice smooth motion. If there are new bugs, this is probably
> why.
>
> 2. New stereo options: anaglyph (red-cyan goggles,) side by side (defocussed
> eyes, stereo displays), vinterlace (autostereo displays)
>
> 3. Physics should be faster. (Collisions were doing the penetration test
> twice)
>
> 3. Bugs fixed:
> 3.1 setObject take KX_GameObject & not string.
> 3.2 Allow actuators with missing targets (ie for setting them from python)
> 3.3 Blenderplayer command line options
> 3.4 Fixed Python Matrix parsing
>
> Cheers
>
> Kester
>
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at projects.blender.org
> http://projects.blender.org/mailman/listinfo/bf-committers
>
More information about the Bf-committers
mailing list