Scotch link error

[jasonhnicholson]

Using the code_saturne v7.3 branch. I am using pop!_os 22.04 which is Ubuntu 22.04 based. You can assume Ubuntu 22.04 and that should be good enough.

Code: Select all

uname -a

linux jhnlt002 6.5.6-76060506-generic #202310061235~1697396945~22.04~9283e32 SMP PREEMPT_DYNAMIC Sun O x86_64 x86_64 x86_64 GNU/Linux

I have already been able to build hdf5, med, and cgns. I cannot get scotch to build. My setup_ini is attached pointing to what I started from after several tries. My install_saturne.log is attached is very short.

I can build scotch fine from its own directory:

Code: Select all

cmake .
make

[Yvan Fournier]

Hello,

Yes, this was fixed for the v8.0 branch and release.

v7.3 is a feature/development branch, and was retired in June, when 8.0 was released, so I suggest upgrading to v8.0.

Best regards,

Yvan

[jasonhnicholson]

I have compile issues with v8.0.2 with both Ubuntu 20.04 and Ubuntu 22.04. I upgraded my OS to try to get Code_Saturne to compile. The available Code_Saturne package is 6.0.2-2build1 in both Ubuntu 20.04 and 22.04 which was released sometime in 2019. The tutorials are written for v7 and v8. I have gone backward to get a version of Code_Saturne to compile. I will try 8.0.2 again. Compiling 8.0.2 failed. I attached the logs. Please, help if you can. For now, I will go back to 6.0.2.

What's occurred to me is to publish a container and a package if I can get Code_Saturne to compile. If I can get into the tutorials and find that Code_Saturne is a good option for solving compressible liquid problems, then I would work on the package and container pieces related to Code_Saturne.

I want to simulate hydraulic pumps/motors and valves across 0-500 bar and 30-90C. These liquids are rho = f(T, p) and viscosity = (f, p). I saw that Code_Saturne can support bulk viscosity functions as well; I am wondering what the impact of adding bulk viscosity to a pump simulation would be. Does it add enough energy damping when a fluid is hydrostatically compressed (rhetorical question)? All these questions are interesting but I need to understand how to use Code_Saturne to simulate basic fluid flow problems first.

[Yvan Fournier]

Hello,

Could you post the "config.log" from the build directory of code_saturne (code_saturne-8.0.build), not MED ? It seems the MED "file" library has compiled OK but the code is also looking for the optional MEDCoupling layer, which should not be the case by default unless perhaps some environment variables relative to Salome are set. An in any case, "configure" can be re-run from that directory, adding "--without-medcoupling" to the invocation command line given line 7 of config.log.
(then run make && make install).

Otherwise, I do not know why code_saturne has not been updated in the Debian repositories. V7.0 is available in "experimental", but even that is 2 years old.

For useful Ubuntu packages for code_saturne, check this : https://github.com/code-saturne/code_sa ... n-packages

For info related to a container, check here also : https://github.com/code-saturne/code_sa ... -container. The packages referenced seem old, so an update would be useful.... We have not worked on this recently on our side. Docker is avoided on our infrastructures for security reasons, and Singularity (which can use similar rules) is used instead, but building a container has not bee our priority.

I am not sure when the tutorials will be updated but the changes on the GUI side between code_saturne 7.0 and 8.0 have been minor (mostly some additions, few changes), so they should still be quite useful even if not quite up to date.

Best regards,

Yvan