[Mesa-users] do_element_diffusion = .false. != diffusion_class_factor(:) = 0

[RICHARD H D TOWNSEND]

Now I'm confused -- I though the defaults were read as a namelist, which is why we need extra code in set_default_controls().

> On Dec 14, 2017, at 2:35 PM, Rob Farmer <r.j.farmer at uva.nl> wrote:
> 
> Setting arrays like that works in the default files because they get compiled into the code not read as a name list.
> 
> Rob
> 
> 
> On 14 Dec 2017 21:28, "Warrick Ball via Mesa-users" <mesa-users at lists.mesastar.org> wrote:
> Hi Earl, Rich, et al.,
> 
> That's interesting and unexpected (at least for me).  An unfortunate
> by-product is then that the defaults files use this "feature". e.g.
> 
>        xa_function_weight(:) = 0
> 
> in &controls (r10108), among many other places.
> 
> Is there any way to make it obvious to users that this doesn't actually
> work?  Writing a comment at each one of these statements in the inlists
> doesn't seem the right way to do it.  Maybe there's a way to crash if the
> user tries?
> 
> Cheers,
> Warrick
> 
> ------------
> Warrick Ball
> Postdoc, School of Physics and Astronomy
> University of Birmingham, Edgbaston, Birmingham B15 2TT
> wball at bison.ph.bham.ac.uk
> +44 (0)121 414 4552
> 
> On Thu, 14 Dec 2017, mesa-users at lists.mesastar.org wrote:
> 
> >
> >
> >> On Dec 14, 2017, at 12:47 PM, Evan Bauer <ebauer at physics.ucsb.edu> wrote:
> >>
> >> Hi Earl,
> >>
> >> It looks like your inlist command is only setting the first class factor to 0. That is
> >>      diffusion_class_factor(:) = 0
> >> somehow only gets diffusion_class_factor(1) and leaves the rest of them untouched. Since the default number of diffusion classes is five, I think your inlist needs
> >>      diffusion_class_factor(1) = 0
> >>      diffusion_class_factor(2) = 0
> >>      diffusion_class_factor(3) = 0
> >>      diffusion_class_factor(4) = 0
> >>      diffusion_class_factor(5) = 0
> >>
> >> Somebody who understands fortran namelists better than me can probably explain exactly why this is happening, but hopefully the above is an adequate solution.
> >>
> >
> > Array assignment in namelists doesn't work; never quite figured out whether this is standard-mandated, or an issue with the compiler(s) we use. That's why, for array-valued controls, we have to default initialize them explicitly rather than doing it from the namelists in star/defaults. Have a look e.g. at set_default_controls() in ctrls_io.f90.
> >
> > cheers,
> >
> > Rich
> >
> >
> >> Cheers,
> >> Evan
> >>
> >>
> >>> On Dec 14, 2017, at 3:58 AM, Earl Bellinger <bellinger at mps.mpg.de> wrote:
> >>>
> >>> Dear mesa-users,
> >>>
> >>> I am wondering if someone can help me to understand why turning off
> >>> element diffusion yields a different result than setting the diffusion
> >>> factor to 0? I would expect, perhaps naively, for them to yield the
> >>> same result.
> >>>
> >>> Some simple inlists are attached for calculating the first Gyr of
> >>> evolution for a solar mass track. There's a small difference in
> >>> several quantities, for example the effective temperature changes from
> >>> 5643.763 to 5643.913. This is somewhat different from running it with
> >>> diffusion class factor of 1, which gives Teff of 5643.915.
> >>>
> >>> Best regards,
> >>> Earl Bellinger
> >>>
> >>> ---
> >>> earlbellinger.github.io
> >>> Earl Bellinger, Ph.D. Candidate
> >>> Stellar Ages & Galactic Evolution Group
> >>> Max Planck Institute for Solar System Research
> >>> <inlist_diff_off><inlist_diff_0><diff_off.data><diff_0.data>
> >>
> >
> >
>