|
1 registered members (TipmyPip),
18,449
guests, and 6
spiders. |
|
Key:
Admin,
Global Mod,
Mod
|
|
|
Re: Fake Fizziks with commercial edition (A6)
[Re: ventilator]
#27771
05/23/04 13:26
05/23/04 13:26
|
Joined: Mar 2003
Posts: 5,377
USofA
fastlane69
Senior Expert
|
Senior Expert
Joined: Mar 2003
Posts: 5,377
USofA
|
Ahhh, that looks meaty. Thanks!
Again, I have no reason to doubt that the 3DGS, ODE are slower than Newton; just want to know why.
Thanks again! 
EDIT: Read it and interestingly enough, it's apperantly headed in the same O(n) direction Newton is supposedly already on...did I read that right?
Last edited by fastlane69; 05/23/04 13:48.
|
|
|
Re: Fake Fizziks with commercial edition (A6)
[Re: fastlane69]
#27772
05/24/04 06:42
05/24/04 06:42
|
Joined: Sep 2003
Posts: 3,236
San Diego, CA
Marco_Grubert
Expert
|
Expert
Joined: Sep 2003
Posts: 3,236
San Diego, CA
|
Quote:
EDIT: Read it and interestingly enough, it's apperantly headed in the same O(n) direction Newton is supposedly already on...did I read that right?
Hard to say, since Mr Jerez does not explain how his algorithm works, which IMO makes stability and accuracy claims somewhat dubious.
I am assuming that the O(n*log(n)) is number of total constraints thrown in and not number of constraints per island or something else. The other interesting thing to look at (and again, no documentation or source) would be behavior with near-singular matrices.
Here are some black box box stacking speed comparisons (please note that "Stepfast1" is currently in GameStudio Beta, Quickstep is scheduled for Fall 2004): http://q12.org/pipermail/ode/2004-May/012851.html
Some general information about ODE: http://ode.org/ode-0.039-userguide.html and the Stepfast1 explanation: http://ode.org/ode-0.039-userguide.html#ref58
Quote:
I get the picture, the user would think that the game ACTUALLY has a physics engine, when it's really one object per microsecond. This is because the human brain cant pick out what happens every microsecond of time, thus giving the illusion of a real physics engine. Sorry if I'm being too repetitive.
That was the idea behind George's code. However, it does not work thar way since the object's velocity is reset to 0 whenever you switch objects. That prevents continuous motion. A better way would be to switch objects once it no longer moves or at least have longer switching cycles.
Last edited by Marco_Grubert; 05/24/04 06:46.
|
|
|
Re: Fake Fizziks with commercial edition (A6)
[Re: fastlane69]
#27774
05/24/04 17:32
05/24/04 17:32
|
Joined: May 2002
Posts: 7,441
ventilator
Senior Expert
|
Senior Expert
Joined: May 2002
Posts: 7,441
|
if i remember correctly i read a post by julio in which he stated that ngd has no problems with singular matrices. we don't know much about newton's internals but fact is that while working with it you will curse less than with other physics engines.
btw. -> http://www.physicsengine.com/forum/viewtopic.php?t=376&highlight=
...the other physics engines shouldn't call themselves rigid body dynamics but spongy body dynamics or something like that!
|
|
|
Re: Fake Fizziks with commercial edition (A6)
[Re: ventilator]
#27775
05/25/04 04:26
05/25/04 04:26
|
Joined: Sep 2003
Posts: 3,236
San Diego, CA
Marco_Grubert
Expert
|
Expert
Joined: Sep 2003
Posts: 3,236
San Diego, CA
|
Quote:
other physics engines shouldn't call themselves rigid body dynamics but spongy body dynamics or something like that!
So I take it the inital Newton demo has been fixed so that the alien no longer keeps bouncing on the floor and the fan doesn't explode when shooting at it repeatedly?
I am thinking about moving Newton Game Dynamics related posts to General Tools since technical questions are better answered in your forum ventilator and ideological statements don't aid in finding answers to GameStudio's physics engine here.
|
|
|
|
