[mesa-users] mismatch between structure files and rows of history file

[Dave Spiegel]

I'm running MESA on a 64-bit (RedHat) Linux machine, with 8 threads.

I've encountered a puzzling issue.
I ran the 1.0 solar mass pre-MS to WD test, and set profile_interval=10 and
history_interval=10.
Right off the bat, there's a mismatch between row k of the history file and
profile #k, because the 1st row of the history file is model #1, whereas
profile1.data is model #10, but this is easy to correct for.

A more puzzling mismatch occurs later because two rows of the history file
end up having the same model number.  Specifically, in this run, rows 1995
and 1996 both have model_number 19910, but different values for physical
quantities.  Row 1996 has the physical quantities that correspond to the
profile file that has this model number (profile1994.data).

One final issue that confused me is that occasionally the number of models
between writes to the history file and between profile????.data writes is
not exactly 10 (i.e., not exactly what is set in the inlist).  For
instance, in the history file, row 81 is model_number 800; row 82 is
model_number 807; and row 83 is model_number 810.  In general, these
nonuniform intervals between writes to the history.data file and between
written profile????.data files occur the same way for each and do not lead
to mismatches between the history file rows and the structure files.

I'm trying to understand why these issues occur.  Do they represent
intended behavior of the code, or do they indicate some sort of bug(s)?
 Please let me know if I can provide any more useful information about my
system or about the run.

Thanks very much,
Dave

-- 
Dave Spiegel, Ph.D.
School of Natural Sciences, Astrophysics
Institute for Advanced Study
Bloomberg Hall
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20120918/75113bbd/attachment.html>