New Bullet jenga demo:
http://www.continuousphysics.com/bullet_jenga.zip
No linux or other platform builds yet.
New Bullet jenga demo (windows only)
-
- Site Admin
- Posts: 4221
- Joined: Sun Jun 26, 2005 6:43 pm
- Location: California, USA
Re: New Bullet jenga demo (windows only)
Ok let`s see, Im new in the forum and only want to say that bullet seems to be great! I've been using it in my old demos (with ODE), any way the run very very fast that the official and tuhopuu versionsErwin Coumans wrote:New Bullet jenga demo:
http://www.continuousphysics.com/bullet_jenga.zip
No linux or other platform builds yet.

Great work Erwin

-
- Posts: 2
- Joined: Fri Jul 15, 2005 2:59 am
-
- Posts: 67
- Joined: Mon Jul 25, 2005 8:56 am
-
- Posts: 26
- Joined: Sat Jul 23, 2005 12:56 am
- Location: LA
-
- Site Admin
- Posts: 4221
- Joined: Sun Jun 26, 2005 6:43 pm
- Location: California, USA
brute force place holders: penetration depth estimation
The jenga stack was supposed to show fairly stable stacking of convex objects, with no deactivation. This means GJK and contact manifold generation 
The current Bullet implementation has a few place holders that are brute force implemented. For example no real broadphase yet, no kdop tree yet, and no proper penetration depth calculation.
Also the jenga demo has no time of impact enabled, so penetration depth estimation is probably killing the fps after the collapse.
So as you can see, it is early stages.
At the moment I'm more focussed to get all above features implemented, that should fix most performance problems.
Still, the jenga stack should perform pretty well. Moving the platform down should bring the jenga stack into contact with the implicit cylinder. So it should tumble over.
- Does it collapse immediatly or after a longer while ?
- Do the fps issues happen immediatly or after a longer while ?

The current Bullet implementation has a few place holders that are brute force implemented. For example no real broadphase yet, no kdop tree yet, and no proper penetration depth calculation.
Also the jenga demo has no time of impact enabled, so penetration depth estimation is probably killing the fps after the collapse.
So as you can see, it is early stages.
At the moment I'm more focussed to get all above features implemented, that should fix most performance problems.
Still, the jenga stack should perform pretty well. Moving the platform down should bring the jenga stack into contact with the implicit cylinder. So it should tumble over.
- Does it collapse immediatly or after a longer while ?
- Do the fps issues happen immediatly or after a longer while ?
-
- Posts: 67
- Joined: Mon Jul 25, 2005 8:56 am
Sorry, it doesn't really "collapse", it probably "tumbles over" instead, as you said.- Does it collapse immediatly or after a longer while ?
- Do the fps issues happen immediatly or after a longer while ?
Here's a picture below. In this situation I have ~1 fps. (It's a 2.4 Ghz machine with a Radeon 9800).

- Pierre