[Haskell-cafe] Squashing space leaks

Daniel Fischer daniel.is.fischer at web.de
Fri May 6 17:39:00 EDT 2005


Am Freitag, 6. Mai 2005 16:58 schrieb Greg Buchholz:
> Daniel Fischer wrote:
> > The problem is that a prime crept in in Josef's code,
> > so to calculate the positions and velocities, the updated versions of the
> > planets are used, it should be
>
>     Ah, yes, thanks for the correction.
>
> > Besides, offset_momentum does not use the parameter n at all and I think
> > that all the indexing in the computation of the potential energy is
> > rather inefficient.
>
>     I wasn't too worried about the energy function, since it only gets
> called twice, whereas the "advance" get called potentially millions of

Of course, and for only five bodies...

> times (and must be the source of the space leak).  You can see one
> profile I ran (for n=1000 instead of 1000000) at...

I did profiling for n = 20000 and n = 100000, if I read the profiles 
correctly, Josef's code reduces memory usage by a factor of 2 resp. 3, which 
ain't bad. Maybe one could reduce space usage further by not creating the 
list of iterations but instead make it a tail recursive function? I'll try.
>
>     http://sleepingsquirrel.org/nbody/nbody.prof
>
> Thanks,
>
> Greg Buchholz

One question: the energy of the system should be constant - that little 
physics I know. What I don't know is whether the change in energy in this 
simulation is within usual, reasonable bounds or not. If not, is it due to a 
too long time interval or should one use a more sophisticated algorithm?

Cheers,
Daniel


More information about the Haskell-Cafe mailing list