[mesa-users] 3723 is current as of Nov 7, 2011

[Bill P. <pa...@ki...>]

Hi,

Yes, I know there was a release just last week.  You can skip this one if you like.

The main reason for a release now is that I've added a small tool to compare
the results from mesa eos/kap/net to the results from other codes for verification.
It works by reading a model in FGONG format, then at each point it uses the
(T,Rho.composition) from the FGONG as arguments to eval the mesa eos/kap/net
and compares the mesa results to the ones given in the FGONG.  You will find
this in star/test_suite/compare_microphysics.  Edit 'inlist_compare_microphysics'
to set the options for mesa using the usual star inlist scheme.  As usual, send
me questions if you'd like to give this a try.

This release also has a new set of mesa opacity tables derived from the same old sources,
but this time at exactly double the (T,Rho) resolution of the OPAL/OP tables.  I've used the
OPAL interpolation code to produce the extra points to go in the tables.   Of course
it would be much better if we just had much higher resolution tables from OPAL/OP
rather than doing this silly fake-high-resolution, but until we have those, this will at
lease make us less vulnerable to interpolation differences.

Speaking of interpolation issues, I've changed the default for interpolation of opacities
in Z from linear to cubic -- again, a better solution would be more tables from OPAL/OP
at more Z's, but until then...

Another changed default is to use the actual non-integer atomic weights when calculating number
density from mass density for nuclear reactions instead of using the baryon number. This
means that chem_isotopes_filename defaults to 'isotopes.data_real' rather than
'isotopes.data_approx' in run_star_defaults.   The use of "real" instead of "approx" will
boost PP rates by about 1.5% and CNO rates by about 0.7% -- less than 0.1% change for others.

Finally, the mesasdk tar.gz's have been updated by Rich Townsend based on 1st round experience.
If your's isn't broken, you don't to reinstall the sdk.  The changes are for ease of installation in the future.

Cheers,
Bill