Tutorial about the coupling simulation with SYRTHES

[c.jauselabert]

Hello everyone,

I am trying to get used to Code_Saturne and Syrthes and I wanted to do the tutorial called "3disks2d".

The first part concerns a computation with Syrthes alone and this is where I have a problem.

Actually, the setting of the computation goes well but when I launch the simulation, it is stuck at 14% at "Conduction Initialization" (see joined file).

Does someone know what is wrong? Is that a common problem?

Thank you all for your help.

Regards,

Clément

[Jacques Fontaine]

Hello,

You should look into all the listings to get the error.
Regards,

[stefann]

I have the exact same problem.

In listing file:

Code: Select all

*** SYRTHES MESH
                           |--------------------|------------------|
                           |   Volumic mesh     |  Boundary mesh   |
      ---------------------|--------------------|------------------|
      | Dimension          |               2    |               2  |
      | Number of nodes    |            1632    |      unused      |
      | Number of elements |            2904    |             360  |
      | Nb nodes per elt   |               3    |               2  |
      ---------------------|--------------------|------------------|

 *** verif_maill : number of elements reoriented : 0

In console output:

Code: Select all

  ---------------------------
  Start SYRTHES preprocessing
  ---------------------------

Updating the mesh file name.. 
   -> OK


  -------------------------
  Start SYRTHES computation
  -------------------------

Execution of SYRTHES.. 
    -> number of processors for conduction = 1 
Segmentation fault (core dumped)

Any idea how to get more verbose output in order to diagnose this seg fault?

Thanks,

Stefan

[Brian Angel]

Hello,

Can you please upload your model onto this forum so that I can have a look at it.

Regards,

Brian Angel.

[stefann]

Hi:

I was following the tutorial here: http://code-saturne.org/cms/sites/defau ... -disks.pdf. I made it as far as page 20 when I ran into the segmentation fault. I attached my files as requested.

I believe this is a syrthes problem, as I didn't make it to the section of the tutorial where you couple with code_saturne.

Thanks


Stefan

[Brian Angel]

Hello,

I'll have a look over the next day or so and come back to you.

Best regards,

Brian Angel.

[Brian Angel]

Hello,

I've used the files that you have uploaded to set up your case on my machine. However, it doesn't run for which reasons it is not clear. So, I set up a new case using the mesh that you supplied. This runs okay and gives results of the temperature distribution in the three discs. The attached file contains the run on my machine which is using Syrthes V4.1 and Ubuntu 13.04.

Can you please try this case on your machine and let me know what happens.

Regards,

Brian Angel.

[Brian Angel]

With the files this time.

[stefann]

Hello:

I ran the syrthes.gui command opened the file you provided and started the run. The progress bar stops at 14% as shown in Clement's screenshot.

I killed the run and closed the gui. I ran the command line which was echo'd from the gui run:

Code: Select all

./syrthes.py -n 1 -d Test_renuda.syd -v ensight -l listing_syrthes

Here was the output:

Code: Select all

SYRTHES4 home directory: /opt/syrthes4.1.1-ubuntu/arch/Linux_x86_64
 MPI home directory: /opt/syrthes4.1.1-ubuntu/extern-libraries/opt/openmpi-1.4.3/arch/Linux_x86_64

  -----------------------------------
  Prepare SYRTHES execution directory
  -----------------------------------

 Building the executable file syrthes.. 
ar xv /opt/syrthes4.1.1-ubuntu/arch/Linux_x86_64/lib/libsyrthes_seq.a mainsyrthes.o
x - mainsyrthes.o
gcc -o syrthes   -O3 -D _FILE_OFFSET_BITS=64 -D_FILE_OFFSET_BITS=64 \
	-I/opt/syrthes4.1.1-ubuntu/arch/Linux_x86_64/include -I/opt/syrthes4.1.1-ubuntu/arch/Linux_x86_64/bib_material_syrthes  -D _FILE_OFFSET_BITS=64 *.o \
	/opt/syrthes4.1.1-ubuntu/arch/Linux_x86_64/lib/libsyrthes_seq.a -lm    

  *****  SYRTHES compilation and link completed *****

  SyrthesCase summary:

    Name =                         SYR
    Data file =                    Test_renuda.syd
    Update Data file =             True
    Do preprocessing =             True
    Debug =                        False
    Case dir. =                    /home/stefan/test/3disks2D/solid
    Execution dir. =               /home/stefan/test/3disks2D/solid
    Data dir. =                    /home/stefan/test/3disks2D/solid
    Source dir. =                  /home/stefan/test/3disks2D/solid
    Post dir. =                    /home/stefan/test/3disks2D/solid/POST

    Conduction mesh dir. =         /home/stefan/test/3disks2D/solid/
    Conduction mesh name =         3rond2d.syr

    Total num. of processes =      1
    Logfile name            =      /home/stefan/test/3disks2D/solid/listing_syrthes
    Echo =                         True
    Parallel run =                 False
    Do preprocessing =             True

   SyrthesParam summary
    Param file name =            Test_renuda.syd
    Conduction mesh name =       3rond2d.syr
    Radiation mesh name =        None
    Result prefix. =             3tond2d
    Restart =                    False
    Coupling =                   False
    Interpreted functions =      False


  ---------------------------
  Start SYRTHES preprocessing
  ---------------------------

Updating the mesh file name.. 
   -> OK


  -------------------------
  Start SYRTHES computation
  -------------------------

Execution of SYRTHES.. 
    -> number of processors for conduction = 1 
Segmentation fault (core dumped)

  Error while running syrthes
Stop Syrthes execution.

Is there a way to compile syrthes with debug on to see where the crash occurs?


Thanks

Stefan

[Yvan Fournier]

Hello,

Yes, in the setup.ini of Syrthes, it is possible to add debug options (I have not done it recently, but have done it in the past).

Otherwise, even without a debug version, running under Valgrind (if the code is small enough) or under a debugger should at least provide a stack trace, without the line numbers, but a least with the source file names, which is a start, and may provide some insight into the crash.

To find the exact calling command (which you need to adapt for a debugger), use the "run_solver" script from the execution directory.

Regards,

Yvan