Re: SPH Glass?

From: Chris Power <cpower_at_astro.swin.edu.au>
Date: Thu, 22 Feb 2007 10:14:19 +1100 (EST)

Hi Gabe,

I've made glass files using GADGET that have been used in both collisional
and collisionless runs. GADGET's approach to glass making is very simple
-- it just sets up a Poisson distribution of particles in a periodic box
and evolves the equations of motion for the system of particles assuming
that the sign of the gravitational constant is reversed. Once the
distibution has reached equilbrium -- which it should do very quickly, but
you can check this by looking at the rms velocity and force distibutions
-- your glass is made.

GADGET implicitly assumes that the glass-making will be collisionless. If
you look in init.c, you can see what is happening:

void seed_glass(void)
{
    ....

  All.TotNumPart = MAKEGLASS;
  partmass = All.Omega0 * (3 * All.Hubble * All.Hubble / (8 * M_PI *
All.G))
    * (All.BoxSize * All.BoxSize * All.BoxSize) / All.TotNumPart;


    ....

  header.npartTotal[1] = All.TotNumPart;
  header.mass[1] = partmass;

    ....
}

You can see here the key lines, header.npartTotal[1] and header.mass[1] --
the code assumes that the glass particles are type==1 or halo particles.
I suspect that the errors you got arose because you set the softenings for
the non-SPH particles to zero.

I hope this helps,

Cheers,

Chris


----------------------------------------------------------------------
Dr. Chris Power
----------------------------------------------------------------------
Centre for Astrophysics & Supercomputing Tel: +61 (3) 9214 8302
Swinburne University of Technology Fax: +61 (3) 9214 8797
Mail #39, P.O. Box 218,
Hawthorn, Victoria, 3122
Australia
----------------------------------------------------------------------


On Wed, 21 Feb 2007, Gabriel Altay wrote:

>
> Hello everyone,
>
> Im trying to use the public version of Gadget2 to make an SPH glass.
> Ive read through the other posts that involve glasses and Ive followed
> the recommendations for the Makefile and the parameter file, but Im
> getting an error during the first time step when I try to run the code.
> see below
>
> ***************************************
> type=1 dmean=3125 asmth=976.562 minmass=847.015 a=0.0078125
> sqrt(<p2>)=0 dlogmax=inf
> displacement time constraint: 0.01 (0.01)
>
> glass-making: dmean= 3125 global disp-maximum= 0 rms= nan
>
> Begin Step 1, Time: 0.00788689, Redshift: 125.793, Systemstep:
> 7.4388e-05, Dloga: 0.00947662
> domain decomposition...
> NTopleaves= 127
>
> No domain decomposition that stays within memory bounds is possible.
> ****************************************
> The fact that dlogmax=inf and rms=nan cant be a good thing.
>
> When I used the same Makefile and parameter file but changed to halo
> particles instead of gas particles (by setting the softening lengths to
> zero for every particle except halo pars ... is this the correct way to
> specify what kind of particles to use? I understand that the initial
> conditions file is ignored), the code ran just fine. Has anyone else
> encountered problems when making an SPH glass? I suppose a solution
> would be to just calculate the smoothing lengths separately after the
> halo particles have reached equilibrium, but the earlier posts made it
> seem that you could do it all in one step.
>
>
> thanks
> Gabe
>
>
>
>
> -----------------------------------------------------------
>
> If you wish to unsubscribe from this mailing, send mail to
> minimalist_at_MPA-Garching.MPG.de with a subject of: unsubscribe gadget-list
> A web-archive of this mailing list is available here:
> http://www.mpa-garching.mpg.de/gadget/gadget-list
>
Received on 2007-02-22 00:35:59

This archive was generated by hypermail 2.3.0 : 2023-01-10 10:01:30 CET