Since this isn't yt related question anymore, let's move this to a private conversation. <div><br></div><div>Sam<br><br><div class="gmail_quote">On Sun, Feb 28, 2010 at 1:45 PM, Agarwal, Shankar <span dir="ltr"><<a href="mailto:sagarwal@ku.edu">sagarwal@ku.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi Sam,<br>
<br>
If I choose option 1, I guess I will need a grid of 512*4=2048. That will resolve 200Mpc/2048=97 Kpc<br>
<br>
In that case,<br>
<br>
MaximumRefinementLevel = 2<br>
CellFlaggingMethod = 2 4<br>
MinimumOverDensityForRefinement = 2.0 4.0<br>
<br>
<br>
How important is MinimumOverDensityForRefinement ? I can't decide why I should not use 4.0 8.0 to reduce computing time. And how do these two numbers relate to the VirialOverdensity=200?<br>
<br>
Is this something that I will have to figure out only by running a simulation? I am hoping somebody already has experience with large box simulations.<br>
<div class="im"><br>
Shankar<br>
<br>
<br>
-----Original Message-----<br>
From: <a href="mailto:yt-users-bounces@lists.spacepope.org">yt-users-bounces@lists.spacepope.org</a> on behalf of Sam Skillman<br>
</div><div><div></div><div class="h5">Sent: Sun 2/28/2010 1:39 PM<br>
To: Discussion of the yt analysis package<br>
Subject: Re: [yt-users] Dealing with HOP groups<br>
<br>
Actually yes. You could check this yourself by examining the last column in<br>
HopAnalysis out. If max_r < 2*(1/512)=3.9e-3 then yes, it will skip it. In<br>
one of my simulations, a 5e11 Msun halo has a max_r of 2.64e-03<br>
corresponding to 168 kpc/h. The halo profiler was not built for profiling<br>
halos with < 2 cells in radius. In my mind you have at least four options:<br>
<br>
1. Run with AMR to resolve the dark matter potential so that you can use<br>
the halo profiler as built. Moderate computational cost. I'm guessing this<br>
was the case for Stephen's simulation where a decent fraction of the halos<br>
in the HopAnalysis were able to be profiled.<br>
2. Run with a gigantic unigrid such that you can resolve these halos with ><br>
2 (and preferably many) cells. This is a very computationally expensive<br>
route as you'd need more than a 2048^3 unigrid in a 200 Mpc box. I do not<br>
recommend this option.<br>
3. Write a dark matter particle profiler that uses the particle positions<br>
(and not grid cells) to bin up the profile. Little computational cost but<br>
requires coding on your part.<br>
4. Run the same size unigrid on a smaller physical volume to capture a<br>
different range of the mass function.<br>
<br>
Cheers,<br>
Sam<br>
<br>
<br>
<br>
On Sun, Feb 28, 2010 at 12:07 PM, Agarwal, Shankar <<a href="mailto:sagarwal@ku.edu">sagarwal@ku.edu</a>> wrote:<br>
<br>
> Hi Sam,<br>
><br>
> Here is how I am using it right now...<br>
><br>
> hp.add_halo_filter(HP.VirialFilter,must_be_virialized=True,<br>
> overdensity_field='ActualOverdensity',<br>
> virial_overdensity=200,<br>
> virial_filters=[['TotalMassMsun','>=','5e11']],<br>
> virial_quantities=['TotalMassMsun','RadiusMpc'])<br>
> hp.make_profiles(filename="VirialHaloes.out")<br>
><br>
> And i get 500.<br>
><br>
><br>
> Do you think it has something to do with the grid spacing (for 512 ^3 grid<br>
> in 200 Box, it is 390 Kpc)<br>
><br>
> HaloProfiler may be skipping small halos completely since they are smaller<br>
> than 390 Kpc.<br>
><br>
><br>
> Shankar<br>
><br>
><br>
> -----Original Message-----<br>
> From: <a href="mailto:yt-users-bounces@lists.spacepope.org">yt-users-bounces@lists.spacepope.org</a> on behalf of Sam Skillman<br>
> Sent: Sun 2/28/2010 12:34 PM<br>
> To: Discussion of the yt analysis package<br>
> Subject: Re: [yt-users] Dealing with HOP groups<br>
><br>
> Hi Shankar,<br>
><br>
> Do you have a line that looks like:<br>
> hp.add_halo_filter(HP.VirialFilter,must_be_virialized=True,<br>
> overdensity_field='ActualOverdensity',<br>
> virial_overdensity=200,<br>
> virial_filters=[['TotalMassMsun','>=','1e14']],<br>
> virial_quantities=['TotalMassMsun','RadiusMpc'])<br>
><br>
> If so, please post it. The number of halos > 10^14 Msun is fairly close to<br>
> 500 in a 200 Mpc box. If you want the halo profiler to use all halos, set<br>
> virial_filters=[['TotalMassMsun','>=','1e14']] accordingly.<br>
><br>
> Sam<br>
><br>
> On Sun, Feb 28, 2010 at 11:08 AM, Agarwal, Shankar <<a href="mailto:sagarwal@ku.edu">sagarwal@ku.edu</a>><br>
> wrote:<br>
><br>
> > Hi Stephen,<br>
> ><br>
> > I read<br>
> ><br>
> <a href="http://lists.spacepope.org/pipermail/yt-users-spacepope.org/2009-March/000110.html" target="_blank">http://lists.spacepope.org/pipermail/yt-users-spacepope.org/2009-March/000110.html</a><br>
> ><br>
> > Do you remember the cosmology parameters for your run: particles, grid,<br>
> > box, refinement, AMR ON or OFF ?<br>
> ><br>
> ><br>
> > I have run a 200 Mpc/h box, 512^3 particle, 512^3 grid, AMR-OFF<br>
> simulation<br>
> > and I get about 30,000 halos using HOP. When I try to finding virial<br>
> halos,<br>
> > I get only 500. I may have to increase the grid to 1024^3.<br>
> ><br>
> > Let me know how you got 40,000 virial halos.<br>
> ><br>
> > Thanks<br>
> > Shankar<br>
> > _______________________________________________<br>
> > yt-users mailing list<br>
> > <a href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a><br>
> > <a href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org" target="_blank">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a><br>
> ><br>
><br>
><br>
><br>
> --<br>
> Samuel W. Skillman<br>
> DOE Computational Science Graduate Fellow<br>
> Center for Astrophysics and Space Astronomy<br>
> University of Colorado at Boulder<br>
> samuel.skillman[at]<a href="http://colorado.edu" target="_blank">colorado.edu</a><br>
><br>
><br>
> _______________________________________________<br>
> yt-users mailing list<br>
> <a href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a><br>
> <a href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org" target="_blank">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a><br>
><br>
><br>
<br>
<br>
--<br>
Samuel W. Skillman<br>
DOE Computational Science Graduate Fellow<br>
Center for Astrophysics and Space Astronomy<br>
University of Colorado at Boulder<br>
samuel.skillman[at]<a href="http://colorado.edu" target="_blank">colorado.edu</a><br>
<br>
</div></div><br>_______________________________________________<br>
yt-users mailing list<br>
<a href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a><br>
<a href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org" target="_blank">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a><br>
<br></blockquote></div><br><br clear="all"><br>-- <br>Samuel W. Skillman<br>DOE Computational Science Graduate Fellow<br>Center for Astrophysics and Space Astronomy<br>University of Colorado at Boulder<br>samuel.skillman[at]<a href="http://colorado.edu">colorado.edu</a><br>
</div>