[mesa-users] critical problem with history interval

[Grant Newsham]

Jakub,

I also ran varying the number of threads and found no difference - or at
least no difference worth bothering about.

Are you just seeing differences in your plots - because it could be a
problem in whatever you are using to plot the data. Look at the raw data
per model step.

Anyway - let me know - I am about to go to sleep for the night - I am
interested in this since it may affect what I am working upon.

Regards,

Grant

On Sat, Jan 12, 2013 at 3:28 AM, Jakub Ostrowski <
ostrowski at astro.uni.wroc.pl> wrote:

> Hi!
>
> Bill, Grant, I have done more calculations today and today I can't
> recreate my initial problem related to differences in history_interval.
> This is actuality pretty weird because yesterday I was able to recreate it
> a few times. On my Mac I have tried OMP_NUM_THREADS = 8 and 4 and for both
> settings I get the same tracks.
>
> But the problem still exits. Take a look at attached figure (I plotted
> only runs with history_interval = 1 there). Red line is from MacBook - it's
> still different than Bill's tracks and I don't believe that it's real. My
> first concern is the weird loop after depletion of central helium and the
> second one - I don't see a reason why luminosity drops so low after helium
> depletion. It looks rather unreal, but I can't get other track on my laptop
> today!
>
>
>
> So I calculated the same inlists on supercomputer, this time also with 8
> threads. The result was identical so I plotted it only one time (green
> line). As you can see this is different than my Mac's result and more
> similar to Bill's track. And it looks better (but this might not be true of
> course…).
>
> OK, so what should I do now? The discrepancy between these two machines is
> huge, I can't recreate original problem on my Mac and I'll have access to
> my Ubuntu PC on Monday - I don't know what to expect on that machine. I
> can't work until it's figured out.
>
> Aaron, I'll try to do something about the contraction after TAMS when my
> main issue is resolved. I'll definitely let you know if I was able to
> achieve some compatibility.
>
> Cheers,
> Jakub
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130112/397be304/attachment.html>