We closed this forum 18 June 2010. It has served us well since 2005 as the ALPHA forum did before it from 2002 to 2005. New discussions are ongoing at the new URL http://forum.processing.org. You'll need to sign up and get a new user account. We're sorry about that inconvenience, but we think it's better in the long run. The content on this forum will remain online.
IndexProgramming Questions & HelpOther Libraries › Fast Point-Spring Library
Page Index Toggle Pages: 1
Fast Point-Spring Library? (Read 875 times)
Fast Point-Spring Library?
Mar 5th, 2010, 6:31pm
 
Hello all,

I'm working on a project with point-spring physics in 3D space. I was wondering if anyone was aware of a fast - really fast, point-spring solver/library. I've been experimenting with Verlet Integration and the respective Relaxation methods, and the result is a realtime performance of a few thousand springs at 30fps with about 15-25 iterations, which is ok, but i'd be interested if anyone knew of anything even faster. I dont need collisions etc., just plain point-spring. Is Traer's really fast? In general, what is the fastest implementation of a point-spring solver?

Thanks.
Re: Fast Point-Spring Library?
Reply #1 - Mar 5th, 2010, 11:18pm
 
yconst wrote on Mar 5th, 2010, 6:31pm:
i'd be interested if anyone knew of anything even faster.


Do you mean 'anything' that is still Java (and free)

If you want fast Physics, get a hardware physics accelerator, and/or consider GPGPU processing.

Short of that, if you're currently using a library with more features than you need, consider modifying the library to remove everything you don't need (or build new code from just the parts that you want) and try to optimise that.

Else consider using a C++ physics library which could possibly offer better native CPU performance.

-spxl
Re: Fast Point-Spring Library?
Reply #2 - Mar 6th, 2010, 5:39am
 
Thanks for the reply

Quote:
Do you mean 'anything' that is still Java (and free)?


Yea, more or less...

To be more precise, what I'm tryin to do is an approximation of the deformation of a spaceframe under load, so I thought that this could be iteratively approximated using a normal relaxation solver instead of implementing a complex finite element analysis.
So:
Hardware is out of the question coz it needs to be distributable
I'm currently using my own implementation of Verlet Integration + Relaxation so only the mimimum is present in terms of features.
Using C is an option, although more like a last ditch effort Tongue

Still, the question remains I believe:

What is (in your opinion) the Fastest (and Free Tongue) point-spring implementation in Java or Processing...

Thanks again!
Re: Fast Point-Spring Library?
Reply #3 - Mar 6th, 2010, 7:17am
 
Unfortunately I can't help you there. I only had a general kind of answer to test your 'anything' option.

I don't know any spring simulators aside from a very basic one I made in Flash (Actionscript 2).

You can see it at subpixels.com/flash and you'll probably have a good idea how it works and how it isn't likely to help you much.  Wink

-spxl
Page Index Toggle Pages: 1