[mesa-users] Timestep control

[Warrick Ball]

Great, thanks for these responses.  It was more a curiosity on my part, 
but good to have the posts as reference.

Cheers,
Warrick

On Wed, 27 Aug 2014, Bill Paxton wrote:

> I routinely use the control 'max_years_for_timestep' in my inlists.  You can of course also set it from your run_star_extras.
>
> Alternatively, if you'd like to go deeper, take a look at run1_star in star/job/run_star_support for the following lines:
>
>               result = star_evolve_step(id, first_try, just_did_backup)
>               if (result == keep_going) result = star_check_model(id)
>               if (result == keep_going) result = extras_check_model(id, id_extra)
>               if (result == keep_going) result = star_pick_next_timestep(id)
>
> This happens for each step -- it does "star_evolve_step", then calls check_model in star, followed by your extras check_model.
> After those, it calls star_pick_next_timestep which sets s% dt_next, the timestep for the next step.
>
> Then after a few more housekeeping tasks, run1_star calls extras_finish_step.
> So if you change dt_next in your extras_finish_step it will change the next timestep.
> You should probably only decrease dt_next -- trying to force larger timesteps is likely to give convergence problems.
>
> Hope that helps.
>
> Cheers,
> Bill
>
>
>
> On Aug 27, 2014, at 9:30 AM, Dean Townsley wrote:
>
>> Hi Warrick,
>>
>> On 08/27/2014 01:37 AM, Warrick Ball wrote:
>>> Are there already hooks for the maximum timestep?  Or is it enough
>>> simply to add something like
>>>
>>>  s% dt = min(dt, my_dt_limit)
>>>
>>> to extras_check_model in run_star_extras?
>>>
>>
>> I think you should probably do it in extras_finish_step() instead, and
>> it is s% dt_next.  This is what we have done to control the timestep for
>> accreting WD setups.
>>
>> I think s%dt is the either the timestep currently being attempted (when
>> extras_check_model() is called) or the timestep just finished (when
>> extras_finish_step() is called).
>>
>> There is also a string somewhere that indicates what is constraining the
>> timestep and is printed in the summary info to the terminal, but I don't
>> know what that is called.  That should probably be changed too, to
>> something like "mylimit" or something more meaningful.
>>
>> Dean
>>
>>
>> ------------------------------------------------------------------------------
>> Slashdot TV.
>> Video for Nerds.  Stuff that matters.
>> http://tv.slashdot.org/
>> _______________________________________________
>> mesa-users mailing list
>> mesa-users at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>
>


------------
Warrick Ball
Postdoc, Institut für Astrophysik Göttingen
wball at astro.physik.uni-goettingen.de
+49 (0) 551 39 5069