SV: [ODE] Particle/cloth/skeleton (was: Re: Choleski factorization)

[david@csworkbench.com]

Yeah, I realize there will be a lot of special case functions that will
have to be written to handle the cases where some macros take an
operator.... but I think it would probably be better to get away from
macros and move towards inline functions anyway, just so improvements like
these can be made.

David

>> SIMD: I wasn't considering rewriting all of the code with SIMD
>> alternatives, but just rewriting Vector and Matrix
>> multiplication and dot
>> product functions with a SIMD alternative.  Then possibly
>
> Sounds good to me. However as I pointed out before. Converting
> ODE to use SSE is not an easy task because in a lot of places the
> computations are done componen wise. Also there is quite a few
> macros that takes the operation as input. I assume that all those
> would have to be replaced with a new macro. Then there are some
> other issues also. In the end it would be a lot of work but it
> would also surely be worth it.
>
>> Verlet:
>
> If what you want is to simulate softbodies then I think you should think
> about what Jeff Lander said. He was playing around with the
> "Lots of particles and distance constraints" idea also when he
> figured out that he would need so many constraints that it would
> be better to just go with a finite elements method.
> He mentions that he's writing a paper on the subject. I asked
> him last week about the paper but I still havent got any reply from him.
> The original position this is mentioned in can be found here
> http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&selm=MlY18.4
> 52%24Vn6.280076%40paloalto-snr1.gtei.net
>
> /Joakim E. - http://www.planestate.net
>
> _______________________________________________
> ODE mailing list
> ODE@q12.org
> http://q12.org/mailman/listinfo/ode