Hi,
thanks for pointing that out. With the approach I'm suggesting however
the softening lengths do not change during the simulation, they are set
in the IC file (like the topicstarter is already doing, only using
different particle types). It would be true when using the Sph part of
the varying softening lengths, where the softening is related to the
-varying- smoothing length. But as I gather the simulation is entirely
collisionless.
Sander
Daniel Pfenniger wrote:
>
> Hi,
> Note that varying softening during a run breaks conservation of
> energy, a sever change in the physics. A softened particle can be
> seen as a distributed mass producing the softened potential. Changing
> the softening radius amounts to displace spherically the mass, which
> changes the binding energy of the whole system.
> Dan
>
> Sander Valcke wrote:
>>
>> Hi all,
>>
>> I don't know if it would be possible to just add a softening variable
>> to the P struct in this case: allvars.h (523):
>>
>> #ifdef ADAPTIVE_GRAVSOFT
>> FLOAT Softening;
>> #endif
>>
>> I'm not very familiar with the gravity code but it seems that setting
>> UNEQUALSOFTENINGS and ADAPTIVE_GRAVSOFT_FORGAS allows varying
>> softening lengths when some code is modified to extend this beyond
>> the sph particles.
>>
>> the following code is found in forcetree.c (line 585):
>>
>> if(All.ForceSoftening[pa->Type] > maxsoft)
>> maxsoft = All.ForceSoftening[pa->Type];
>>
>> if this would be encapsulated in an extra ADAPTIVE_GRAVSOFT flag:
>>
>> #ifndef ADAPTIVE_GRAVSOFT
>> <insert code above>
>> #else
>> if (pa->Softening > maxsoft)
>> maxsoft = pa->Softening;
>> #endif
>>
>> (or something similar and more general)
>>
>> a similar change in potential.c lines 108/109, gravtree.c lines
>> 132/133, forcetree.c lines 585/586, 1158/1159, 1174/1175, 1230/1244,
>> 1455/1456, 1514/1528, 2070/2071, 2140/2154, 2334/2335, 2416/2430
>> (note: the two-line pieces of code are identical, as are the 15 line
>> pieces), timestep.c around 477. Basically at every place where
>> ADAPTIVE_GRAVSOFT_FORGAS is found.
>>
>> Then there would just need to be 2 more small additions, for reading
>> and writing the softening length from and to snapshots in the case
>> ADAPTIVE_GRAVSOFT is set.
>>
>> Anyway that's what I make of it : )
>>
>> kind regards,
>>
>> Sander
>>
>> Yves Revaz wrote:
>>> Hi Florian,
>>>
>>> One possibility could be simply to extend the number of type to a value
>>> higher than 6.
>>> Unfortunately in the code the number of type is not coded by a constant
>>> but by
>>> the number 6, so you will have to change this value in a lot of places
>>> in the code :-( .
>>>
>>> Then of course, you will have to add new parameters in the gadget param
>>> file,
>>> corresponding to the new types (and read them correctly in begrun.c and
>>> set_softenings in gravtree.c).
>>>
>>> SofteningGas
>>> SofteningHalo
>>> SofteningDisk
>>> SofteningBulge
>>> SofteningStars
>>> SofteningBndry
>>> SofteningXxxx
>>> etc...
>>>
>>> SofteningGasMaxPhys
>>> SofteningHaloMaxPhys
>>> SofteningDiskMaxPhys
>>> SofteningBulgeMaxPhys
>>> SofteningStarsMaxPhys
>>> SofteningBndryMaxPhys
>>> SofteningXxxxMaxPhys
>>> etc...
>>>
>>>
>>> Be also carefull to the way the snapshot header will be written, as
>>> npart (and npartTotal)
>>> will have a larger size.
>>>
>>> I hope it will help,
>>>
>>> yves
>>>
>>>
>>>
>>>
>>> Florian Beutler wrote:
>>>
>>>> hello
>>>> I am working with a DM halo containing particles with different
>>>> mass (shell model). I would like to have different softening values
>>>> for particles with different mass. The softening value is connected
>>>> to the particle type therefore I am using particle types 1-5 for
>>>> halo particles (I think these particle types are all collisionless
>>>> and I can use them all as halo type particles... or am I wrong?).
>>>> Is there any possibility to extend the particle types to more then 5?
>>>> thanks and regards
>>>> Florian Beutler
>>>>
>>>
>>>
>>>
>>
>
>
>
>
> -----------------------------------------------------------
> 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
--
Eagles may soar, but weasels don't get sucked into jet engines.
Received on 2009-05-28 10:22:39