CS v2.1.0 and subroutines

[Alexandre Guilloux]

Hello,

I have some problems with CS v2.1.0 and the subroutine usclim.f90 (same problem with usproj.f90).

The compilation of the subroutine didn't seem to be a problem. I attached the compile.log file.

But during the calculation Code_Saturne doesn't consider the subroutine.
 
I don't know which informations do you need to resolve my problem. So tell me if you need more file.
 
Best regards

Alexandre

[Yvan Fournier]

Hello,

I did no see anything suspicious.

Do you at least have the messages you print in the "listing" file (but not the correct boundary condition), or do you not even have the message ?

If the calculation stops on error, could you post the "run_solver.sh" file in the execution directory ?

Best regards,

  Yvan

[Alexandre Guilloux]

Hello,

I did not have the messages in my "listing" file, my "condition_limite.dat" file is not create, and the boundary condition is not correct.

The subroutine is compiled but it's like Code_Saturne doesn't use it during the calculation.

Futhermore, I haven't error message in the execution directory.

I will keep searching and I will keep you informed of my problem.

Best regards,

Alexandre

[Alexandre Guilloux]

Hello,

I made the Tutorial Case03, case which use usclim.f90 subroutine.

I have the same phenomenon, Code_Saturne compile the subroutine but doesn't use it during the calculation. For information, I have a constant inlet temperature (300°C).

I add some files of the calculation, but I don't have error message.
 
Best regards,

Alexandre

[Yvan Fournier]

Hello,

I assume that for the Tutorial case 3, you adjusted to selection criteria in usclim.f90 to use the correct mesh group names (numbers) instead of 'ENTREE'.

In any case, you should still have a message in you "listing" file.

Also, you did not post your "run_solver.sh" file.

Best regards,

  Yvan

[Alexandre Guilloux]

Hello,

When I run the Tutorial case 3, I use the subroutine usclim.f90 which is in the repertory "/code_saturne-2.1.0/examples/2-full_domain/case3" (not my subroutine).

I post the "run_solver.sh" file. There is no error during my calculation, so I had to recompile the subroutine.
 
During the installation of Code_Saturne on my new machine I had some problems with the open mpi library. I don't know if the problem is link. So I post the setup file on my installation.

Thanks for your help

Alexandre

[Yvan Fournier]

Hello,

You posted the cs_solver executable, but I would like to check the run_solver.sh...

Regards,

  Yvan

[Alexandre Guilloux]

Hello,

Sorry I didn't understand. I post the run_solver.sh file.

Regards,

Alexandre

[Alexandre Guilloux]

Hello,
I still have my problem with the library mpi and the compilation of subroutines.
All my test are make with Tutorial "Full domaine" case 3
.
 
I have two installations of code_saturne 2.1.4, with variant configurations:
 
Version 01 :
Installation options:
-compc = gcc
-compF = gfortran
-mpiCompC = mpicc
-mpi = none no no
 
Behaviour of this Version
- The Fortran subroutines are compile but they are not use during the calculation. (the compile.log file shows mpicc is used to compile
)
- The calculation with several processors works.
(The run_solver.sh file shows MPICH2 is used)
 
Version 02:
 Installation options:

-compc = usr/bin/gcc

-compF = usr/bin/gfortran

-mpiCompC = none

-mpi = /usr/lib/openmpi yes no
 
Behaviour of this Version
- The Fortran subroutine are compile and use during the calculation.
(the compile.log file shows gcc is used to compile)

- The calculation with several processors bug in a random way.  The run_solver.sh file shows openMPI is used)
 
I am not comfortable with this library problem. I don't know if it's possible but I would like:
 
- gcc as compilator fortran
- MPICH2 as mpi library
 
But I can't do this in the setup file.
 
Thanks for your help,
 
Best regards,

Alexandre

[Alexandre Guilloux]

A another file