[mesa-users] The file "scan_grid_results.data"

[João Faria]

Hi Kuldeep,

I reproduced the problem with the inlists you sent (in MESA 6022). And I 
also think those might be the lines of code causing the issue.

I don't see a problem with having non-consecutive sample numbers since 
the chi^2 values will ultimately be sorted in the file 
"scan_grid_results.data". When I tried increasing sample_number before 
the return (in mesa/star/astero/src/run_star_extras_astero.f, around 
line 260) it worked fine and the results file came out with the two 
samples.


Cheers,
João

On Wed 28 May 2014 02:30:54 PM WEST, Kuldeep Verma wrote:
>
> Hi Everyone,
>
> Since nobody replied to my previous email, I was impelled to look
> myself in the code. It seems from the following piece of code in
> "mesa/star/astero/src/run_star_extras_astero.f"
>
> if (best_chi2 < 0) then
> write(*,*) 'failed to find chi^2 for this run'
> call zero_best_info
> best_chi2 = 999999d0
> return
> end if
>
> sample_number = sample_number + 1
>
> that the variable "sample_number" is not being updated when the
> program fails to calculate the total chi-square. This is perhaps to
> name the successful samples (the samples that passes all the observed
> constraints and the total chi-square is successfully calculated during
> the evolution) in a sequence like 1, 2, 3, 4, ..... But this creates a
> problem. There will be no entry in the file "scan_grid_results.data"
> for all the successful samples that preceded a failed sample (the
> total chi-square is not calculated during the evolution). To
> understand this, let us suppose that the total chi-square is
> calculated successfully for the second last grid point and not for the
> last grid point. Then the quantities sample_chi2(i), sample_age(i),
> sample_Teff(i) etc. have values corresponding to the best model at the
> end of the second last grid point (i = sample_number at this point),
> and get replaced by initializing values of best_chi2, best_age,
> best_Teff etc. after the run (since the sample_number remains the
> same). In this way the information about the best parameters of second
> last grid point is lost from the file "scan_grid_results.data". To see
> this, please download the inlists and starting model from my previous
> email in the test suite "example_astero" and run it. There are two
> models in the grid. The total chi-square is calculated for the first
> model and the best parameters can be seen in the file
> "scan_grid_results.data" at the end of its evolution. The second model
> does not satisfy all the observed constraints and therefore the total
> chi-square is not found during the evolution. At the end of the run,
> we have 999999 for chi-square and initializing values for other
> parameters in file "scan_grid_results.data".
>
> It is quite possible that I have misunderstood the code. If so, please
> forgive for such a long email and in light me about what I get after
> running the attached inlists. I'm using latest version of MESA, 6208.
>
> With Best Wishes,
> Kuldeep
>
>
> On Wed, May 21, 2014 at 9:45 PM, Kuldeep Verma <kuldeepv89 at gmail.com
> <mailto:kuldeepv89 at gmail.com>> wrote:
>
> Hi Everyone,
>
> Can someone please point me out, where (in the code) or how the
> file "scan_grid_results.data" is written while running a grid of
> models using the test suite "example_astero"? I ran a grid of a
> couple of models, the first model successfully calculates the
> chi-square (~14) and the second fails. I was expecting the first
> model to be present in the file "scan_grid_results.data", but
> turns out that the second model replaces the first one with of
> course a chi-square equal to 999999. I have attached the required
> inlists and starting model, in case someone wants to have a look
> or try running it.
>
> Thanks in advance for any help.
>
> With Best Wishes,
> Kuldeep
>
>
>
>
> ------------------------------------------------------------------------------
> Time is money. Stop wasting it! Get your web API in 5 minutes.
> www.restlet.com/download
> http://p.sf.net/sfu/restlet
>
>
> _______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users