Hi.
I'm just learning about different BroadPhase collision strategies. And now btDbvtBroadphase keeps my attention.
Is there any advantages for using btDbvtBroadphase over other strategies like SAP (btAxisSweep3)?
Well this question is more concerned about performance, because it is clearly that the main advantage about btDbvtBroadphase is that it could manage large worlds.
And it seems that btDbvt is pretty efficient for updating body locations.
About btDbvtBroadphase and other approaches
-
- Posts: 109
- Joined: Thu Dec 14, 2006 4:27 pm
- Location: Colombia
-
- Posts: 78
- Joined: Mon Nov 13, 2006 1:44 am
Re: About btDbvtBroadphase and other approaches
Compared to SAP, here's a short pro/con list:
Pros:
- Fast insertion, removal, update.
- Unbounded world space.
- Additional queries support: Ray-casting, Culling, Occlusion, LOD, etc...
- One data structure for physics and everything else in a game engine.
Cons:
- Lot of cache misses compared to SAP (can be improved).
- An hand optimized DbvtBroadphase should be slower than an hand optimized SAP on medium load (5k bodies) when used to broadphase only.
The current btDbvtBroadphase is a bit faster SAP in benchmarking cases on PC, but its mainly due to implementation, and i don't think those results extrapolate on 'slow branching' targets like xbox360.
Nathanael.
Pros:
- Fast insertion, removal, update.
- Unbounded world space.
- Additional queries support: Ray-casting, Culling, Occlusion, LOD, etc...
- One data structure for physics and everything else in a game engine.
Cons:
- Lot of cache misses compared to SAP (can be improved).
- An hand optimized DbvtBroadphase should be slower than an hand optimized SAP on medium load (5k bodies) when used to broadphase only.
The current btDbvtBroadphase is a bit faster SAP in benchmarking cases on PC, but its mainly due to implementation, and i don't think those results extrapolate on 'slow branching' targets like xbox360.
Nathanael.
You do not have the required permissions to view the files attached to this post.
-
- Posts: 109
- Joined: Thu Dec 14, 2006 4:27 pm
- Location: Colombia
Re: About btDbvtBroadphase and other approaches
Thanks a lot Nathanael.
I didn't make benchmarking about it before. But now I'm more interested in your Dynamic tree approach.
So thanks.
I didn't make benchmarking about it before. But now I'm more interested in your Dynamic tree approach.
So thanks.