[Mesa-users] System crashes when evolving beyond zams

[YASH MEHUL MEHTA]

So I did a bit of experimentation:
Running with ./rn and with ./re 1000 both results in the system crashing exactly at step 1011. This is true with and without garbage collection.
Interestingly, I was able to achieve step 1023 exactly once, after restarting my laptop, force closing all tasks, and maximizing RAM (this run was without garbage collection, and was done using ./re 1000). But I was not able to reproduce this despite using the same setup multiple times.

I have made absolutely no changes to the default install except setting z=0.0001 and setting terminating-at-zams to false (and later, introducing garbage collection).
There is no error message. Either the virtual machine simply crashes, or the entire laptop shuts down. I have never faced this problem in the past despite some heavy computation.
________________________________
From: Warrick Ball <W.H.Ball at bham.ac.uk>
Sent: 16 April 2021 20:31
To: YASH MEHUL MEHTA <yashmehta at iisc.ac.in>
Cc: mesa-users at lists.mesastar.org <mesa-users at lists.mesastar.org>
Subject: Re: [Mesa-users] System crashes when evolving beyond zams

External Email


Hi,

If it is a lack of RAM, you can start by restarting your run from a photo.  If you haven't changed the options, the `photos/` directory will contain files that are either `x` followed by a 3-digit number or just a 4-digit number. e.g. `x050` or `1000`.  You can try restarting from one of these towards the end of your run, before the crash.  You say you reached model 1011, so try

     ./re x950

or perhaps

     ./re 1000

If that gets further than the original run with `./rn`, it's probably because you don't have enough RAM.  To keep your runs going, you can try using the "garbage collector" by adding e.g.

     num_steps_for_garbage_collection = 100 ! every 100 steps; you might need to adjust this

to `&star_job` and perhaps

     report_garbage_collection = .true. ! instead of the default .false.

to see that the garbage collection is working.

The garbage collector works by de-allocating the EoS handles.  Those that are still needed for the evolution are automatically reloaded, so in effect this drops those handles that aren't still required, thereby freeing some memory.

This isn't guaranteed to work.  There might be a situation where the model simply needs more EoS tables than you have RAM, but this might let you squeeze by.

All that said, if the model still crashes at the same point after a restart, it might not be the RAM and you need to send more information: inlists to reproduce the error and the full error message.

Cheers,
Warrick

___________

Warrick Ball
Postdoc, School of Physics and Astronomy
University of Birmingham, Edgbaston, Birmingham B15 2TT
W.H.Ball at bham.ac.uk
+44 (0)121 414 4552

On Fri, 16 Apr 2021, mesa-users at lists.mesastar.org wrote:

> Hi
> My name is Yash Mehta, I am running mesa-r15140 on my virtual machine with 7000MB RAM. (I am running ubuntu16.04 on my windows10 home, 8gb Dell XPS 15)
> Evolving a star upto zams is causing no issues, but when trying to evolve a 15M_sun star with initial z=0.0001 beyond zams, the system crashes somewhere midway (typically after 1011th iteration; zams is reached at 938th iteration)
> Right before the system crashes, the terminal shows (after the data about the 1011th iteration):
> "write /home/yash/Downloads/mesa-r15140/data/eosDT_data/cache/mesa-FreeEOS_00z10x.bin"
> "write /home/yash/Downloads/mesa-r15140/data/eosDT_data/cache/mesa-FreeEOS_02z10x.bin"
> Is the issue simply lack of RAM or could it be something else? Also, is there a way out?
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20210416/a5fdb18d/attachment.htm>