Domain: bulletphysics.com
Stories and comments across the archive that link to bulletphysics.com.
Comments · 12
-
Re:Good luck with that
It's also worth noting that Bullet physics implemented GPU acceleration via CUDA (same as PhysX) themselves...10 months ago: http://www.bulletphysics.com/wordpress/?p=50
-
Re:Anti-trust?
Bullet Physics http://www.bulletphysics.com/wordpress/ claims to, and ohmigod, has support for Nvidia CUDA acceleration(just like PhysX). That ATI and other card makers have not yet provided an API that an open source and competing (#3 behind Havok) product finds desirable, is not Nvidia's fault either.
-
Re:Havok
Bullet was made CUDA-acceleration ready last year for Nvidia users, by Bullet physics developers themselves. http://www.bulletphysics.com/wordpress/?p=50
Nvidia has beta/developer OpenCL drivers out. I can't find any download location for AMD's reported OpenCL drivers, despite mention of them everywhere, not even on developer.amd.com. It seems like AMD's backing the wrong pony, and being a bit disingenuous to boot. Par for the course?
It seems like they're officially 'taking a swipe' for the fact that Nvidia wasn't paying Bullet Physics devs to add OpenCL first, even though Bullet supported Nvidia all along.
Also hilariously worth noting, the Bullet physics website (quoted by parent) has absolutely no mention of this "push" by AMD at the time of writing.
-
Re:Good luck with that
An OpenCL implementation of Bullet physics is coming. It's Open Source and is already being used in commercial games -- once it gets GPU acceleration there will probably be little demand for PhysX.
I completely agree and look forward to applying my M.E. and C.S. backgrounds for my work.
-
Re:Havok
Havok is a better engine anyway.
But that's the problem with corporate buyings anyway. Even if its kinda wrong to stop supporting the other platforms, they have every right to do so.
Have Havok. I'll take Bullet http://www.bulletphysics.com/wordpress/ I rather like AMD making Bullet OpenCL ready.
-
Re:Good luck with that
An OpenCL implementation of Bullet physics is coming. It's Open Source and is already being used in commercial games -- once it gets GPU acceleration there will probably be little demand for PhysX.
-
Bullet Physics for the Win!
I don't have any affiliation with the project other than I've used it in my homegrown game engine that has never left my hard drive. It is however rather easy to use. When I was looking for a physics engine, Bullet turned out to be the best license, code base, and documentation set out there for no cost.
-
Bite the...
-
Re:Probable Patent Infringement
That's pretty much what a physics engine does, and there are already a number of open source physics libraries out there (ODE and Bullet are the most well supported as far as I know, the former has been used in a few big budget commercial titles). Someone just needs to port the back-end to CUDA and off we go... Easier said than done, I reckon.
I recall hearing chatter about CUDA bindings for Bullet but I'm not sure if anything came of that. -
Re:why not GPL it?
Because there is an absolutely unfathomable amount of brilliant mathematics running behind the scenes. The kind of stuff that competitor physics engine authors would love to pore over.
Even if they can't use that work, they can certainly learn some of the tricks that Havok uses.
Now, if what you want is open source physics, check out ODE ( http:://ode.org ) and bullet ( http://www.bulletphysics.com/ ) both are fully open source, both are well documented, and both are quite good, but aim at different usages.
ODE is great for robotics simulation, and is decent for games. ODE's strength is in joint types ( motors, sliders, etc ). ODE has a clean C api and is very easy to use.
Bullet is faster and more stable than ODE; however it is targeted more towards gaming. It's weak in joint types, but has much better performance when dealing with large groups of bodies. Bullet also has soft-body simulation as well, which is very very impressive. Bullet has a C++ api which, well, is well designed but perhaps over-designed. Finally, bullet supports Collada export, so you can wire up rigs in Blender and use them in your game.
I've been using ODE for five years. I love it. But recently I worked with the bullet folks to port their demo app to cocoa ( they have a windowsy demo, which uses GLUI, but was fugly on OS X ). I learned a lot about bullet doing that and I can see myself moving that direction in the future, when the joint types mature.
Anyway; I just wanted to make it clear that there are valid open source physics engines out there. -
How does it compare
What I want to know is: how does it compare to the existing Open Source physics libraries, such as Bullet (which was made by an ex-Havok developer)?
-
Re:ODENow moving into the realm of I-have-no-idea-what-I'm-talking-about. I hope you are talking about yourself there. If you read the date that mail is from it is Feb 2006, a lot has happened in that 18 months. If you care to take a look at Bullet's feature list : you will see that is now supports: Projected Gauss Siedel (quickstep) and Generic 6 Degree of Freedom Constraint , Motors, Limits
In a recent project of mine I created an ODE implementation, but it was painfully slow. I changed my implementation to use Bullet and got about a 5x improvement in performance. For me Bullet was superior, espically in the realms of convex collisions & dynamics.