Search found 8 matches
- Wed May 25, 2011 1:28 pm
- Forum: Questions and Help
- Topic: Sphere-terrain optimization.
- Replies: 8
- Views: 5262
Re: Sphere-terrain optimization.
I've implemented the first version of the sphere-terrain special case. The other mesh types will come soon, in addition to some more testing. You can grab it on the development fork: http://bepuphysics.codeplex.com/SourceControl/network/Forks/RossNordby/Development The simulation that used to take ...
- Mon May 23, 2011 8:09 am
- Forum: Questions and Help
- Topic: Sphere-terrain optimization.
- Replies: 8
- Views: 5262
Re: Sphere-terrain optimization.
Part of the problem I think was that I was running those tests in Debug mode. In release mode, using the latest development version this time (I was using 0.15.2 before), with nothing changed except for TerrainDemo.cs: using System; using BEPUphysics.Collidables; using BEPUphysics.Entities.Prefabs; ...
- Mon May 23, 2011 4:37 am
- Forum: Questions and Help
- Topic: Sphere-terrain optimization.
- Replies: 8
- Views: 5262
Re: Sphere-terrain optimization.
By the way, I don't know how your whole architecture is set up, but you can get some pretty significant gains from using internal multithreading. A fork-and-join type approach where the physics engine has full access to the Xbox's 3 cores will take the per-update time down a lot, though other compu...
- Mon May 23, 2011 4:17 am
- Forum: Questions and Help
- Topic: Sphere-terrain optimization.
- Replies: 8
- Views: 5262
Re: Sphere-terrain optimization.
Could I ask what the original terrain test looked like? I may be able to use it to do additional optimizing in the future. Here's a copy of my TerrainDemo.cs, in the worst case this reaches ~14ms on my E5300 3.0Ghz single threaded. using System; using BEPUphysics.Collidables; using BEPUphysics.Enti...
- Sun May 22, 2011 3:35 pm
- Forum: Questions and Help
- Topic: 100-200ms delay when exploding beside inactive spheres?
- Replies: 1
- Views: 1983
100-200ms delay when exploding beside inactive spheres?
I place 400 spheres on a terrain in the terrain demo, and allow them to sit for a minute. Then I move close to a group of about 10 spheres that have collected in a pocket of the terrain and shoot the testing-ball-thingie, and explode the ball beside the spheres. This causes a significantly long fram...
- Sun May 22, 2011 1:55 pm
- Forum: Questions and Help
- Topic: Sphere-terrain optimization.
- Replies: 8
- Views: 5262
Sphere-terrain optimization.
Hi Norbo, you dropped by on this thread the other day: http://www.olhovsky.com/2011/05/physics-engine-integrated-and-sphere-terrain-collision-done/ You seemed to suggest that 400 spheres takes 25ms to update on the XBOX. So I did some testing to see how performance works out on the XBOX 360 in my ow...
- Tue May 10, 2011 6:33 pm
- Forum: General
- Topic: Regarding BEPU algorithms used.
- Replies: 1
- Views: 4837
Regarding BEPU algorithms used.
I've been reading about how physics engines work a bit, before jumping in and trying to blindly use BEPU. I'm just curious about a few things, which I'm sure I can discover from the source, but it would be nice if you could just tell me :) Does BEPU use Projected Gauss-Siedel? If not, then what? Is ...
- Fri Apr 15, 2011 8:56 am
- Forum: Questions and Help
- Topic: Character movement improvement.
- Replies: 1
- Views: 2112
Character movement improvement.
First of all, this project is really cool, and I'll certainly donate if I decide to use it, and maybe I'll donate anyway even if I don't. I've been working on a DBP competition submission for 6 weeks or so, and I've started to realize that my own basic physics simulation is not enough to handle incr...