Re: Re: Run-time error: A timestep of size zero was assigned

From: Volker Springel <volker_at_MPA-Garching.MPG.DE>
Date: Mon, 19 Nov 2007 10:00:27 +0100

Hi,

I agree with Juan. This error is almost certainly caused by using a
softening length equal to 0 for a particle type that is used in the
simulation. This causes a division by zero in the gravity calculation
(from the self-force term), which means that the forces come out as
infinity. But even if they were made finite by filtering the self-force
term explicitly, the timestep criterion would give a zero timestep for
zero softening...

Volker


Juan Carlos Muñoz Cuartas wrote:
>
> Hi,
>
> I had this error time ago, and the cause was an error
> in my parameter file. Look at your parameter file and
> verify that the softening length of every particle is
> not zero, it is needed to compute the integration time
> step.
>
> best!
>
> --- Christopher Erick Moody <cmoody_at_caltech.edu>
> escribió:
>
>> Hi,
>>
>> I am running a fairly simple simulation of 1500
>> particles in an
>> NFW profile. I have a similar error: 'A timestep of
>> size zero was
>> assigned on the integer timeline!' that always
>> occurs on the first
>> timestep. It was suggested that the particle in
>> question (1495) is an
>> outlier and may cause problem. But if I remove it,
>> the error simply
>> moves to another particle (1348). Removing that
>> particle means that
>> particle 1493, and then particle 1492 have errors.
>> None of these
>> particles are outliers, and are relatively close to
>> other particles,
>> so I'm not sure why GADGET-2 would have problems
>> with them to begin
>> with.
>>
>> Any help is appreciated,
>> chris
>>
>> (the error is copied below:)
>>
>> Begin Step 0, Time: 0, Systemstep: 0
>> domain decomposition...
>> NTopleaves= 162
>> work-load balance=1 memory-balance=1
>> domain decomposition done.
>> begin Peano-Hilbert order...
>> Peano-Hilbert done.
>> Start force computation...
>> Starting non-periodic PM calculation (grid=0).
>> done PM.
>> Tree construction.
>> Tree construction done.
>> Begin tree force.
>> tree is done.
>> Begin tree force.
>> tree is done.
>> force computation done.
>>
>> Error: A timestep of size zero was assigned on the
>> integer timeline!
>> We better stop.
>> Task=0 Part-ID=1495 dt=nan tibase=1.86265e-09
>> ti_step=-2147483648
>> ac=nan xyz=(-2507.5|-1560.8|-253.42)
>> tree=(nan|nannan)
>>
>> pm_force=(0.000301103|0.000178601|3.49493e-05)
>> task 0: endrun called with an error level of 818
>> pm_force=(0.000443893|0.000335812|0.000159569)
>> task 0: endrun called with an error level of 818
>>
>>
>>
>>
>> From: Hugo Vicente Capelato <hugo_at_das.inpe.br>
>> Date: Tue, 13 Jun 2006 09:52:02 -0300
>>
>> Hi,
>>
>> I´m sorry, I did´nt had time yet to regard what is
>> happening in detail
>> with my simulations and give you a more elaborated
>> answer. I´m terribly
>> busy these times. Anyway, from past simulations for
>> which I encountered
>> the same error messages, I think I´m having some
>> particles escaping to very
>> large distances, as suggested by Ives, and that is
>> the cause of my troubles.
>>
>> To fix this, I think you must remove these particles
>> by hand and then reestart
>> the simulation ...
>>
>> Cheers, H.
>>
>> Quoting Yago Ascasibar <yago_at_aip.de>:
>>
>>>
>>> Hi,
>>>
>>> I've also had such a problem, and I never really
>> understood why. Actually,
>>> it
>>> seems likely that I do have some isolated
>> particles near the boundaries.
>>> Could you please elaborate a little bit more on
>> what may be happening?
>>> Thanks a lot,
>>> Yago
>>>
>>> On Saturday 10 June 2006 12:18, Yves Revaz wrote:
>>>> 2) Check that the snapshot from which you
>> re-start the simulation do not
>>>> contains Nan or Inf numbers, or isolated
>> particles at very large
>>>> distance from the
>>>> center of mass or very large velocities.
>>>
>>>
>>>
>>>
> -----------------------------------------------------------
>>> 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
>>>
>> --
>> HUGO VICENTE CAPELATO
>> Divisao de Astrofisica
>> Instituto Nacional de Pesquisas Espaciais
>> Av. dos Astronautas, 1758
>> 12.227-010 Sao Jose dos Campos
>> Phone numbers:
>> at INPE : 55 12 3945-7207 /-7200 /-6804
>> fax : 55 12 3945-6811
>> at IAG : 55 11 3091-2706
>> at home : 55 11 3726-2793
>> portable: 55 11 8206-5690
>> e-mail: hugo-at-das.inpe.br,hugo-at-astro.iag.usp.br
>> -------------------------------------------------
>> This mail sent through IMP: http://horde.org/imp/
>>
>>
>>
>>
>>
> -----------------------------------------------------------
>>
>> 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
>>
>
>
> ===============================================
> Juan Carlos Muñoz Cuartas
> Grupo de Fisica y Astrofisica Computacional
> Estudiante de Maestria
> Instituto de Fisica, Universidad de Antioquia
> Medellin-Colombia
> http://urania.udea.edu.co/jcuartas
> ===============================================
>
>
> Comparte video en la ventana de tus mensajes (y también tus fotos de Flickr). Usa el nuevo Yahoo! Messenger versión Beta.
> http://e1.beta.messenger.yahoo.com/
>
>
>
>
> -----------------------------------------------------------
>
> 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 2007-11-19 10:00:17

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