[CIG-SEISMO] Why SPECFEM2D can't deal with adjoint simulation with all boundaries set free?

[#HU LEI#]

Hi, Prof.Komatitsch and whom it may concern,

After typing the commands which you suggested and trying another run with all boundaries set free, I found that the error message in the screen-shot below is all that I got. I exported the simulation log into a file which is attached in this email and named as 'running_log'. From this file, I found that the forward modeling is good while the adjoint simulation stopped at time step = 1000, which indicates that the adjoint simulation did run for a period, but met a problem somewhere during its process.

I strongly suggest that you may first run this 'EXAMPLES/LuoYang_fluid_solid_kernel' case on your own laptop by setting 'STACEY_ABSORBING_CONDITIONS     = .false.' in your SPECFEM2D to see if you also meet this problem, just in case that I might have made mistakes with some setting.

Warm regards,

Ray
[https://outlook.office.com/owa/service.svc/s/GetFileAttachment?id=AAMkADY1YTk5NTlmLTk4ZWEtNGVmYS04YzkwLTZlZTYyYmFjMzEwZABGAAAAAACLRP8oPOz8RZpU%2B40JfAOoBwBoaJqhvzRZSrqLvSbQOhMFAAAAAAEJAABoaJqhvzRZSrqLvSbQOhMFAAGT81WhAAABEgAQAKMGNZSb1GdLkhaVgejOryw%3D&X-OWA-CANARY=iaIYowiveUSiP8GwwwCYEkDcdT4-gtUYGyS1_AahLX2kSpotOoSVpeV9Ww8aknfXNw5QyU5-BNA.&isImagePreview=True]

________________________________
From: Dimitri Komatitsch <komatitsch at lma.cnrs-mrs.fr>
Sent: Friday, March 2, 2018 6:38 AM
To: #HU LEI#
Subject: Re: Why SPECFEM2D can't deal with adjoint simulation with all boundaries set free?


Dear Ray,

Thanks for your message. Free boundaries on all the sides of a model are
rarely used in geophysics, thus it is probably a minor bug just coming
from the fact that such a configuration has never been tested by anybody
else. Probably easy to fix.

Could you type this:

./configure --enable-debug
make clean
make all

and try the run again, and let us know if you get any useful error
message at run time?

Also, could you please send your message below to the mailing list?
(cig-seismo at geodynamics.org). Someone may know the answer.

Thanks,
Best regards,
Dimitri.

On 03/01/2018 08:34 AM, #HU LEI# wrote:
> Hi, Prof.Komatitsch,
>
>
> I am a PHD student from NANYANG TECHONOLOGICAL UNIVERSITY. I found that,
> in the case of  'EXAMPLES/Luoyang_fluid_solid_kernal', SPECFEM2D can't
> run properly if I change only all the four boundary conditions from
> Stacey ABC to free.
>
> And I am curious about why. the images indicating the setting and error
> is indicated below.
>
> Hope to hear from you soon.
>
>
> Thanks and regards,
>
>
> Ray
>
>
>
>

--
Dimitri Komatitsch, CNRS Research Director (DR CNRS)
Laboratory of Mechanics and Acoustics, Marseille, France
http://komatitsch.free.fr
Dimitri Komatitsch<http://komatitsch.free.fr/>
komatitsch.free.fr
CNRS Research Director (Directeur de Recherche CNRS) Laboratory of Mechanics and Acoustics, CNRS/University of Aix-Marseille, France (je suis français, vous pouvez m ...


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180305/3947584d/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: running_log
Type: application/octet-stream
Size: 78546 bytes
Desc: running_log
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20180305/3947584d/attachment-0001.obj>