Re: Barnes-Hut treecode (was Re: Galaxy Interaction Simulations)

Steinn Sigurdsson (steinly@topaz.ucsc.edu)
10 Mar 94 11:07:56

In article <CMFGFz.AzK@usenet.ucs.indiana.edu> sdslavin@pegasus2.astro.indiana.edu (Shawn Slavin) writes:

In article <STEINLY.94Mar9120625@topaz.ucsc.edu>,
Steinn Sigurdsson <steinly@topaz.ucsc.edu> wrote:
>In article <1994Mar9.005526.28636@jarvis.csri.toronto.edu> wayne@csri.toronto.edu (Wayne Hayes) writes:

[Stuff deleted. ;)]

> >To model globular clusters with unsoftened particles (ie including
> >2bod relaxation) is not possible,

> You mean not possible with current machines/techniques, or fundamentally
> impossible, even in principle? I was under the impression that it's
> only a matter of not enough CPU cycles to accurately model large-N
> systems without softening.

>Well, in a very real sense modeling large N collisional systems
>is intractable, there are two problems, the true trajectories
>diverge from the calculated ones, and you can show that if there
>are bound subsystems then at any resolution (in the point mass
>approximation) there will be a perturbation of a bound system by an
>unbound particle that is significant at your resolution scale but
>can't be accurately modeled at that scale - there is however good reason to believe
>neither issue matters all that much.

Well, I sat in on a colloquium today, which happened to be given by Piet Hut.
The basic answer was: Simulations of globular clusters of 10^5 stars will have
to wait for GRAPE, as you need ~10^19 floating point operations to go from
initial conditions to just past core collapse => you need a teraflop-class
computer.

HARP, not GRAPE :-)

Also, the focus of his talk was binaries in gc's, which I found quite
interesting. These two, three, ... body interactions are really important in
gc dynamical evolution. Adding the code to handle close interactions of
scattering + binary formation/hardening in 3-body interactions really chews up
your CPU time. (i.e. I was assuming you are using a softened potential
[pretty much the backbone of the treecode], but now you may want to include [I
would anyway] close interactions and binary formation.)

I am not using a treecode to model globulars, the code I mentioned
earlier is conceptually quite different - see Hernquist & Ostriker
1992 for the algorithm.
The problem I was alluding to is how to handle the 3(&4) body
interactions. In a full N-body code, you can't track the binaries,
some you have to carry as a "single" particle and treat close
approaches separately. The problem then becomes how to do that
self-consistently. The current bet is that it doesn't matter too
much if you don't include all levels of binary perturbation.
There are other hybridization approaches where you combine
different N-body approximations to do different part of the physics
but this is not the place to discuss it.

* Steinn Sigurdsson Lick Observatory *
* steinly@lick.ucsc.edu "standard disclaimer" *
* The laws of gravity are very,very strict *
* And you're just bending them for your own benefit - B.B. 1988*