BEPUphysics

Modifying angularMomentum directly should do the trick, and as for the player characters, that's something that will be slightly tricky. Two spheres can roughly approximate an ellipsoid, though the process of making it act like a player character will likely require some fancy footwork. Physically s...

Yeah that's a good idea, I'll put that in v0.2.0. Thanks.

Currently, the engine uses discrete collision detection, so it can miss collisions between small, fast objects as you have observed. The iteration count corresponds to the collision response solver iterations, so increasing or decreasing it won't be much help. You could artificially increase the upd...

Alright, compound bodies are done a bit late, but I'm going to start doing some final testing and make some changes to the demos project. Hopefully, today will be the day.

After quite a bit more searching, I've finally got it to freeze while I have a profiler running, and it is showing a very specific location for the problem. It's seperate from where my previous tests indicated, though very loosely related. This is much more definitive than my exception tests, so I h...

Fe_Yoshi wrote:And you do realize there isn't even a General sub-forum for ANYTHING, right?

Theeeeeere is a reason for that!

This extreme weirdness deserves a post of its own over a simple edit. I am encountering what appear to be logical impossibilities- in my experience, computers don't take too kindly to logical impossibilities. While everything works happy dandy in debug mode, the freezes that occur under "start ...

o y ty! Hopefully I can fix this catastrophic bug soon, then.

Yes and no, it seems. I went back and re-tried the full code sample you posted; it does indeed freeze under non-debug mode with both configurations. However, using only the original snippet placed into the original demos source code, it works fine in Debug configuration under both debug and start wi...

Under what conditions does the space.update() method throw a NaN exception? Is there a stability issue (excessive jitter, falling through objects, etc) besides the fact that cubes freeze under a release build in certain circumstances? If not, using a debug build configuration should help- I cannot r...

Thanks, I'll look through it.

Alright, I've gotten it to freeze with that set up and in non-debug mode. I'll try to figure out what's going on. Though with that one giant cube, it's apparent its not the wedging problem, and it isn't an infinite loop within the controller handling. This might take a bit of time. I can't think off...

I ran the sample code provided, but I was unable to replicate a freeze. Are there any other adjustments to the simulation settings (gravity, timescale, external forces and such) outside of the code sample? Right now, my best guess as to why you might see a freeze is that the bottom cube gets wedged ...

Every once in a while I find myself wanting a little more than NProf, and I have checked out Ants a few times; I'll probably end up grabbing it at some point. Fortunately, I know where the bottlenecks are right now so my optimizations can be targeted without Ants, but I foresee it being very useful ...

I think my wording might have been a little misleading on the stress test- my FPS shows 0 during the major computation parts, and my 1830-block test never went above 0, even at the start. I think the film was run in either .5 or 1 timescale, then I sped it up to 'real time' with the video render. In...