[Mesa-users] Questions about the debugging controls

Rob Farmer robert.j.farmer37 at gmail.com
Thu Aug 19 09:25:13 UTC 2021


Hi,
>I still get this message, which confused me a bit because surely this
means that eps_nuc should no longer limit the timestep?

Its not setting the timestep as you set the limit to -1, but it will still
print debugging information.

max resid dlnE_dt     2  0.366E-03

Is telling me mesa is struggling in the envelope of your model (zone=2) as
the residual in the energy equation (dlne_dt) is 0.366E-03 (smaller is
better).
I would look at what the outer layers of your star is doing (and if this is
with your new eos then check that it is reporting the correct derivatives).

Rob

On Wed, 18 Aug 2021 at 18:24, Jeremy Sakstein <sakstein at hawaii.edu> wrote:

> Hi Rob,
>
> Here is an example output for a step that is causing problems (likely the
> model will not progress far beyond this point):
>
>                                       log eps nuc ratio         673
>  2306      100799   -2.9351265085002717D-07    2.9351265085002717D-07
>  1.0312068975460102D+00    1.0312062006175886D+00
>
> hydro_call_number, s% dt, dt, dt/secyer, log dt/yr, tim       80022
>  6.5794075228711705D+01    6.5794075228711705D+01    2.0848520988102703D-06
>   -5.6809247488191792D+00    1.1498761104281389D+13
>
> 100800    1  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.604E-05    max resid dlnE_dt     2  0.336E+00  avg corr  0.332E-05
>   max corr lnT     1 -0.151E-01  lg dt/yr -5.68  max corr, avg+max resid
> 100800    2  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.179E-04      max resid du_dt     2  0.109E+01  avg corr  0.212E-04
>   max corr lnd     1  0.186E+00  lg dt/yr -5.68  max corr, avg+max resid
> 100800    3  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.638E-05      max resid du_dt     2  0.397E+00  avg corr  0.238E-04
>   max corr lnd     1 -0.208E+00  lg dt/yr -5.68  max corr, avg+max resid
> 100800    4  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.157E-04    max resid dlnE_dt     2  0.111E+01  avg corr  0.145E-04
>   max corr lnd     2 -0.743E-01  lg dt/yr -5.68  max corr, avg+max resid
> 100800    5  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.138E-05    max resid dlnE_dt     2  0.924E-01  avg corr  0.121E-05
>   max corr lnd     1  0.753E-02  lg dt/yr -5.68  max corr, avg+max resid
> 100800    6  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.326E-07    max resid dlnE_dt     2  0.228E-02  avg corr  0.312E-06
>   max corr lnd     2  0.218E-02  lg dt/yr -5.68  okay!
>      100800   8.866758  2.342E+05 -22.657234   4.482143  24.966366
> 0.000000   0.000000   0.000000   0.983259   0.000000  -3.012370   2306
> 4887
>   -5.680925   4.285447  -0.127391   2.490563   7.119873 -99.000000
>  24.966366   0.000000   0.793755   0.016741   0.047990   0.208284      6
>    0
>  3.6437E+05  21.191831   6.177020   4.477692   6.697846  -4.026484
>  21.357420   0.103959   0.086107   0.000951  9.520E-01 -0.393E-01  max
> increase
>                        log_rel_run_E_err       -5.4381278585262685
>                                rel_E_err    2.5507532745735840D-12
>                              center fe56    6.1523181848720156D-07
>                           fe_core_infall        0.0000000000000000
>
>                                       log eps nuc ratio         673
>  2306      100800   -3.0946312842115191D-07    3.0946312842115191D-07
>  1.0312076323484072D+00    1.0312068975460102D+00
>
> hydro_call_number, s% dt, dt, dt/secyer, log dt/yr, tim       80023
>  6.9083778990147295D+01    6.9083778990147295D+01    2.1890947037507840D-06
>   -5.6597354497492409D+00    1.1498761104350473D+13
>
> 100801    1  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.115E-05    max resid dlnE_dt     2  0.870E-01  avg corr  0.592E-05
>   max corr lnd     1  0.295E-01  lg dt/yr -5.66  max corr, avg+max resid
> 100801    2  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.102E-07    max resid dlnE_dt     2  0.656E-03  avg corr  0.138E-05
>   max corr lnd     1  0.814E-02  lg dt/yr -5.66  max corr, avg+max resid
> 100801    3  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.189E-09    max resid dlnE_dt     2  0.133E-04  avg corr  0.512E-07
>   max corr lnd     2  0.300E-03  lg dt/yr -5.66  avg+max resid
> 100801    4  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.815E-11    max resid dlnE_dt  2189  0.371E-08  avg corr  0.139E-09
>   max corr lnT     1 -0.628E-06  lg dt/yr -5.66  max resid
> 100801    5  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.814E-11    max resid dlnE_dt  2245  0.437E-08  avg corr  0.112E-13
>   max corr lnd     1 -0.623E-11  lg dt/yr -5.66  max resid
> 100801    6  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.810E-11    max resid dlnE_dt  2246  0.436E-08  avg corr  0.158E-13
>   max corr lnd   113 -0.969E-12  lg dt/yr -5.66  okay!
>                                       log eps nuc ratio         673
>  2306      100801   -3.2565049140316067D-07    3.2565049140316067D-07
>  1.0312084055873316D+00    1.0312076323484072D+00
>
> hydro_call_number, s% dt, dt, dt/secyer, log dt/yr, tim       80024
>  7.2537967939654663D+01    7.2537967939654663D+01    2.2985494389383233D-06
>   -5.6385461506793026D+00    1.1498761104423010D+13
>
> 100802    1  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.685E-05    max resid dlnE_dt     2  0.377E+00  avg corr  0.372E-05
>   max corr lnT     1 -0.151E-01  lg dt/yr -5.64  max corr, avg+max resid
> 100802    2  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.205E-04      max resid du_dt     2  0.124E+01  avg corr  0.236E-04
>   max corr lnd     1  0.210E+00  lg dt/yr -5.64  max corr, avg+max resid
> 100802    3  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.198E-04      max resid du_dt     2  0.880E+00  avg corr  0.343E-04
>   max corr lnd     1 -0.272E+00  lg dt/yr -5.64  avg+max corr+resid
> 100802    4  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.413E-04    max resid dlnE_dt     2  0.269E+01  avg corr  0.224E-04
>   max corr lnd     2 -0.114E+00  lg dt/yr -5.64  max corr, avg+max resid
> 100802    5  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.646E-05    max resid dlnE_dt     2  0.356E+00  avg corr  0.337E-05
>   max corr lnd     1  0.128E-01  lg dt/yr -5.64  max corr, avg+max resid
> 100802    6  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.494E-06    max resid dlnE_dt     2  0.326E-01  avg corr  0.210E-05
>   max corr lnd     2  0.112E-01  lg dt/yr -5.64  max corr
> 100802    7  coeff  1.0000  slope  0.000E+00  f  0.000E+00  avg resid
>  0.513E-08    max resid dlnE_dt     2  0.366E-03  avg corr  0.183E-06
>   max corr lnd     2  0.124E-02  lg dt/yr -5.64  okay!
>                                       log eps nuc ratio         673
>  2306      100802   -3.4120550040021100D-07    3.4120550040021100D-07
>  1.0312092157613792D+00    1.0312084055873316D+00
>
>
> Cheers, and thanks for looking at this for me,
>
> Jeremy
>
>
>
>
> On Tue, Aug 17, 2021 at 8:34 AM Rob Farmer <robert.j.farmer37 at gmail.com>
> wrote:
>
>> > I want to use report_hydro_solver_progress correct?
>>
>> Yes
>>
>> > I'm using 12778
>>
>> Always useful information to provide as options can change names over time
>>
>> Rob
>>
>>
>> On Tue, 17 Aug 2021 at 19:06, Jeremy Sakstein <sakstein at hawaii.edu>
>> wrote:
>>
>>> Thanks Rob,
>>>
>>> I will do this and send you the output.
>>>
>>> Just to check: I want to use report_hydro_solver_progress correct? I
>>> couldn't find a control called report_solver_progress in the controls (I'm
>>> using 12778).
>>>
>>> Cheers,
>>>
>>> Jeremy
>>>
>>> On Tue, Aug 17, 2021 at 4:53 AM Rob Farmer <robert.j.farmer37 at gmail.com>
>>> wrote:
>>>
>>>> grep -rinI 'log eps nuc ratio' star/p*/*
>>>> star/private/timestep.f90:1703:            if (s% report_all_dt_limits)
>>>> write(*,4) 'log eps nuc ratio', &
>>>>
>>>> That leads to the function check_dlog_eps_nuc_change, note there are
>>>> some divide by eps_nuc's in there. So having 0's floating around (when you
>>>> set eps_nuc_factor to 0) is not going to help. I would suggest just using
>>>> s% dxdt_nuc_factor = 0d0  which turns off composition change from burning
>>>> but keeps eps_nuc set. That way your star won't just collapse as it runs
>>>> out of nuclear energy.
>>>>
>>>> Try setting
>>>> report_solver_progress = .true.
>>>>
>>>> and send the output for a step that seems to cause problems.
>>>>
>>>> Rob
>>>>
>>>>
>>>>
>>>> On Mon, 16 Aug 2021 at 20:31, Jeremy Sakstein <sakstein at hawaii.edu>
>>>> wrote:
>>>>
>>>>> Hi everyone,
>>>>>
>>>>> I've been having some convergence issues so I used the debugging
>>>>> controls:
>>>>>
>>>>> report_why_dt_limits
>>>>> report_all_dt_limits
>>>>> report_hydro_dt_info
>>>>>
>>>>> To try to figure out what is causing them. These helped a lot and I
>>>>> have been able to push the model much further but I am still unable to get
>>>>> full convergence so I wanted to ask a couple of questions.
>>>>>
>>>>> Currently, when I turn on report_why_dt_limits it just tells me
>>>>>
>>>>> "timestep_hold > model_number, so no timestep increase"
>>>>>
>>>>> If I also turn on report_all_dt_limits it tells me "log eps nuc ratio"
>>>>> followed by some ratios so I assume that eps_nuc is what is limiting the
>>>>> timestep? Is this the correct interpretation of this information?
>>>>>
>>>>> When I set
>>>>>
>>>>>    delta_log_eps_nuc_limit = -1
>>>>>    delta_log_eps_nuc_hard_limit = -1
>>>>>
>>>>> I still get this message, which confused me a bit because surely this
>>>>> means that eps_nuc should no longer limit the timestep?
>>>>>
>>>>> As an extreme test, I updated extras_start_step to set
>>>>>
>>>>> s% dxdt_nuc_factor = 0d0
>>>>>
>>>>> s% eps_nuc_factor = 0d0
>>>>>
>>>>> which I think should effectively turn off nuclear burning. Indeed, the
>>>>> abundances don't seem to change but I still get the same eps_nuc warning
>>>>> from report_all_dt_limits.
>>>>>
>>>>> Is someone able to tell me whether these changes are doing what I
>>>>> expected given my output?
>>>>>
>>>>> It would be great if someone could suggest either further ways I can
>>>>> determine what is limiting the timestep, or how to stop eps_nuc limiting
>>>>> the timestep if this is really what is happening.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Jeremy
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Jeremy Sakstein (He/Him/His)
>>>>> Assistant Professor of Theoretical Physics
>>>>> Department of Physics & Astronomy
>>>>> University of Hawaiʻi at Mānoa
>>>>> Watanabe Hall
>>>>> 2505 Correa Road
>>>>> Honolulu, HI, 96822, USA
>>>>> http://www.jeremysakstein.com
>>>>> _______________________________________________
>>>>> mesa-users at lists.mesastar.org
>>>>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>>>>
>>>>>
>>>
>>> --
>>> Jeremy Sakstein (He/Him/His)
>>> Assistant Professor of Theoretical Physics
>>> Department of Physics & Astronomy
>>> University of Hawaiʻi at Mānoa
>>> Watanabe Hall
>>> 2505 Correa Road
>>> Honolulu, HI, 96822, USA
>>> http://www.jeremysakstein.com
>>>
>>
>
> --
> Jeremy Sakstein (He/Him/His)
> Assistant Professor of Theoretical Physics
> Department of Physics & Astronomy
> University of Hawaiʻi at Mānoa
> Watanabe Hall
> 2505 Correa Road
> Honolulu, HI, 96822, USA
> http://www.jeremysakstein.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20210819/6f881a47/attachment.htm>


More information about the Mesa-users mailing list