[Mesa-users] Termination issues with hydrogen burn in 1M star

[Warrick Ball]

Hi Kaitlyn,

I think `1M_pre_ms_to_wd` should only take ~6GB of RAM at any point.  Certainly 16GB should be enough.  How much actual RAM does your laptop have?  Do you possibly have to restart the Ubuntu emulator for changes to the RAM allocation to take effect?

Can you try running any (or all) of the test `timing`, `irradiated_planet`, `make_metals` or any of the `rsp` test cases?  They should all use less than about half as much RAM as `1M_pre_ms_to_wd`.

I'm afraid I don't have any experience with running MESA under Windows myself (be it VM / WSL / anything else), so I'm just clutching at possible issues.

Cheers,
Warrick

___________

Warrick Ball
Senior Research Software Engineer
University of Birmingham
W.H.Ball at bham.ac.uk

On Mon, 24 Jul 2023, Kaitlyn L wrote:

> CAUTION: This email originated from outside the organisation. Do not click links or open attachments unless you recognise the sender and know the content is safe.
>
>
> Hi Warrick,
>
> Thank you for the reply. I figured out how to allocate more ram to
> Ubuntu and allocated 8GB. When I tried to run it, I still received the
> same message. I even tried allocating 16GB, but nothing changed. I
> believe it could have something to do with the amount of physical RAM
> I have, although I'm not sure.
>
> Thanks,
> Kaitlyn Liang
>
>>
>> On Jul 19, 2023, at 3:54 PM, Warrick Ball <W.H.Ball at bham.ac.uk> wrote:
>> Hi Kaitlyn,
>
>
>>
>>> ... to the `&controls` part of the inlist.
>>
>> My mistake: this should be added to the `&star_job` part of the inlist, not `&controls`.
>>
>> Cheers,
>> Warrick
>>
>> ___________
>>
>> Warrick Ball
>> Senior Research Software Engineer
>> University of Birmingham
>> W.H.Ball at bham.ac.uk
>>
>>> On Wed, 19 Jul 2023, Warrick Ball wrote:
>>>
>>> Hi Kaitlyn,
>>>
>>> The "Killed" message is usually a sign that MESA has run out of RAM.  If possible, try to increase the amount of RAM available to the emulator.  The system should have at least 8GB available to it, though you might get by okay with 6GB.
>>>
>>> Alternatively, you can try adding the control [1]
>>>
>>>   num_steps_for_garbage_collection = 1000 ! default 0, might need even less
>>>
>>> to the `&controls` part of the inlist.  It looks like you're crashing on the main sequence, though, which isn't usually heavy on RAM, so you might find you need to decrease this to 100, or perhaps even 50.  MESA will run more slowly (it'll pause to drop then reload the equation of state tables) but should hopefully crash less.
>>>
>>> Cheers,
>>> Warrick
>>>
>>> [1] https://docs.mesastar.org/en/release-r23.05.1/reference/star_job.html#num-steps-for-garbage-collection
>>>
>>> ___________
>>>
>>> Warrick Ball
>>> Senior Research Software Engineer
>>> University of Birmingham
>>> W.H.Ball at bham.ac.uk
>>>
>>>> On Wed, 19 Jul 2023, Kaitlyn L via Mesa-users wrote:
>>>> CAUTION: This email originated from outside the organisation. Do not click
>>>> links or open attachments unless you recognise the sender and know the
>>>> content is safe.
>>>> Hi,
>>>> I've been running the 1M_pre_to_wd code from the mesa-r23.05.1 test suite
>>>> and am repeatedly getting the following error:
>>>> write
>>>> /home/kaitlynliang/mesa-r23.05.1/data/eosDT_data/cache/mesa-FreeEOS_02z30x.bin
>>>> ./rn1: line 6:  1142 Killed                  ./star
>>>> DATE: 2023-07-19
>>>> TIME: 15:13:28
>>>> ******************************************************************
>>>> failed to create end_core_h_burn.mod when running
>>>> inlist_to_end_core_h_burn_header
>>>> ******************************************************************
>>>> Surprisingly, start.mod gets written to disk, but crashes before
>>>> end_core_h_burn.mod is written to disk.
>>>> MESA is being run inside a Ubuntu emulator running on top of Windows 10 HP
>>>> Laptop. I'm a highschool student working with an experienced MESA user,
>>>> but neither of us can figure it out.
>>>> Thank you,
>>>> -Kaitlyn Liang
>