Re: [Algorithms] N-body processing

[Adam M. <amo...@dp...>]

Don't call it stupid, I do that!  But of course I have at most tens of
bodies, not hundreds.  What you may want is a storage format for a sparse
matrix.  Sparse matrices are an important construct used a lot by sci
computation people, so there is no need for you to invent anything new.

--
--Adam Moravanszky
http://www.n.ethz.ch/student/adammo

>The most obvious and stupid solution would be something like an array of
N*N
>structures, with N = the number of objects in the simulation. Of course,
>this is absolutely out of question. My standard test sets N = 1024, to give
>you some ideas.
>
>I'm curious about the solutions you people could imagine. Something with an
>hash-table maybe, but then what would be a good hash-key? This also can be
>seen as the reference/owner problem one can have with shared
>materials/textures/etc, since an (owner, ref) couple is nothing more than a
>pair of colliding/touching/linked/related entities. I don't think I have
any
>brilliant ideas for the moment, but I'm curious about yours....
>
>Cheers,
>
>Pierre