Search found 7 matches

by niad
Sun Feb 10, 2019 12:19 pm
Forum: General Bullet Physics Support and Feedback
Topic: performance of collisions of 'voxel' shapes
Replies: 6
Views: 868

Re: performance of collisions of 'voxel' shapes

If you can get Erwins method to work well I'd stick with that, the approach I use is more complicated etc Looking at the profile result you posted, it spends pretty much all of the time in the nearphase callback. Some of that overhead is because it assumes each node has its own transform(not true fo...
by niad
Thu Feb 07, 2019 8:21 am
Forum: General Bullet Physics Support and Feedback
Topic: performance of collisions of 'voxel' shapes
Replies: 6
Views: 868

Re: performance of collisions of 'voxel' shapes

My recommendation is to create a new shape type(VoxShape) that stores the voxel grid internally.


This is what I do, and I have no such performance issues, even with many voxels colliding.
by niad
Tue Jan 15, 2019 9:28 am
Forum: General Bullet Physics Support and Feedback
Topic: btbvhtrianglemeshshape on separated objects
Replies: 2
Views: 295

Re: btbvhtrianglemeshshape on separated objects

That would probably just create a giant entry in the broadphase, that everything would overlap, and slow your program down- With only 30-80 verts, that sounds like a very low detail mesh, you could probably use convex hulls. Triangle meshes tend to be easy to penetrate since they are just surfaces a...
by niad
Tue Jan 15, 2019 9:17 am
Forum: General Bullet Physics Support and Feedback
Topic: Multiple transformations
Replies: 1
Views: 195

Re: Multiple transformations

If you are just using it for collision detection (everything is inactive) perhaps the broadphase doesn't see a need to update the AABB's after they are initially inserted. Try calling btCollisionWorld::updateAabbs() after you have moved your bodys. You probably need m_forceUpdateAllAabbs on if you a...
by niad
Tue Jan 15, 2019 9:04 am
Forum: General Bullet Physics Support and Feedback
Topic: btConvexPointCloudShape not scaling
Replies: 3
Views: 270

Re: btConvexPointCloudShape not scaling

I don't use that class, but glancing at the code(localGetSupportingVertexWithoutMargin) I'd say the scaling is applied too late-- shouldn't matter if you are doing uniform scaling but with (3,6,1) I'd guess it would produce incorrect results. Try changing this code in localGetSupportingVertexWithout...
by niad
Sun Jan 13, 2019 7:08 pm
Forum: General Bullet Physics Support and Feedback
Topic: In btDbvtBroadphase, why is deferred collide not on by default?
Replies: 2
Views: 322

Re: In btDbvtBroadphase, why is deferred collide not on by default?

Erwin responded on github-- I guess Bullet has some policy of never changing defaults I guess that is why this is off -- Also why m_forceUpdateAllAabbs is enabled by default Badytheprogram : while that will help other unrelated computations, it won't have much effect on bullets broadphase(inserting ...
by niad
Thu Jan 10, 2019 10:05 pm
Forum: General Bullet Physics Support and Feedback
Topic: In btDbvtBroadphase, why is deferred collide not on by default?
Replies: 2
Views: 322

In btDbvtBroadphase, why is deferred collide not on by default?

Recently I was pushing bullets btDbvtBroadphase more than I normally do, in particular I was using it to drive a particle system. It would spawn a few thousands small dynamics near each other, they would only live a few seconds, then die, and more would be constantly spawning in. btDbvtBroadphase st...