PhysX support?

I saw a thread from last year about PhysX support, but it looks like it's not in the tree (at least, at http://code.google.com/p/jmonkeyengine/source/browse/, which was the latest code I could find).



Is anyone working on this, or planning to? Or does anyone know of any other Java bindings for PhysX? I couldn't find any.



Thanks!


  • Chris

i know there's an empty java Physx port up on google code that someone initiated but never actually put any code to…  It's kind of hard to make a port of a closed source library.  There's the jME physics engine, however, which is open-source like the rest of jME :smiley:



If it's hardware accelerated physics you're looking for, you will want to keep an eye on how OpenCL progresses.  You can apply for Nvidia's developer program… if you get accepted, they have the sdk's for OpenCL up for download.  It's still not in release drivers for hardware vendors though, so you wouldn't be able to deploy it immediately

jmephysics is a separate project, here would be the place to look for it:

http://code.google.com/p/jmephysics/source/browse/#svn/trunk/impl



But its not there.

Irrisor the creator of jmephysics is not actively developing jmephysics anymore.



ODE is the only complete implementation.

@sbook: Thanks! What I'd really like is an API that I can write to that will allow hardware acceleration, but doesn't require HW accel. And I'm really hoping to use an open source API like jME Physics, rather than being tied to one vendor like I'd be with PhysX. I'll look into OpenCL.



@Core-Dump: Thanks for the pointer. I had found that project, but since it's inactive, I thought I'd see if anybody knew about any current efforts.

http://jphysx.com

@Uree:  regarding jphysx . . . I dunno . . . looks like a semi-dead project to me.  Only a Linux binding unless I’m reading the activity logs wrong.



@chrisl and everybody else:



I had talked a while back about getting a PhysX binding for JMEPhysics up and working, but . . . I think it’s a lower priority.



As far as JMEPhysics in general goes.  I’m currently working on that.  It’s not QUITE dead yet, but things have been going slowly.  I’m getting close to a complete-ish JBullet implementation, and I’m currently ramping up to either start work on integrating with Bullet Physics which is making a BIG push at the moment to use OpenCL (with the cooperation of Intel, Apple, NVidia AND ATI!) sometime in the near future.  Either that or Erlend is putting me in touch with some physics gurus that I’m hoping to tap to help out with building our own engine straight into JMEPhysics.  Not sure quite how far that’ll go, but I would imagine that at SOME POINT if we build our own engine, we might think about OpenCLifying it, but that’s a WAAAAAAAAAAAAAAAYY long way down the road.



So . . . Keep your eye on JMEPhysics, even if it’s not there quite yet.

Falken, that sounds like some awesome plans! :slight_smile:

Is this mujib222 some kind of news bot ? he posted a similar article in a project Darkstar thread lol.

Looks like a spam bot that tries to get a foothold with semi-related posts?

renanse said:

Looks like a spam bot that tries to get a foothold with semi-related posts?


Yeah, that's usually the case with posts like that..  The dead giveaway is always links in their signature.