Windows with WSL software installation troubleshooting

[LaTrisse-blip]

Hi,

After completing the software installation with the tutorial for Windows with WSL, I'm not able to get a fully passed test (73% instead of 100%, see the log attached.
ctest_output.txt

I've tried with apt, I've tried with candi. Each time, I got the same result. Moreover, the simulations which failed have always been the same. At some point during the previous "cmake" step, the prompt tells me Doxygen cannot be found, could this be related? It's the first warning message I got throughout the whole process.

Did another user ever report a similar issue? Would you describe getting the 100% tests passed as mandatory before using the software?

Thanks for your time.

[blaisb]

Dear @LaTrisse-blip :)

The tests are portable to some extent. Sometimes on radically different compiling environment or etc., they might differ for some slight variation of numbers. This rarely happens, but this could be an hypothesis. The tests are something we use a lot internally for development, but running them is absolutely not mandatory for the software to run.

Generally (99.999% of the times), you are able to compile Lethe and you are able to run an example to completion, it means the installation is perfectly fine and you can proceed on with your simulations.

What I would suggest is that you try running one of the many examples. For example, the easiest one to run is the 2D lid-driven cavity : https://chaos-polymtl.github.io/lethe/documentation/examples/incompressible-flow/2d-lid-driven-cavity-flow/lid-driven-cavity-flow.html

Try running it first without any mpi using :
lethe-fluid cavity.prm

Then try running it with two cores using :
mpirun -np 2 lethe-fluid cavity.prm

If both work, then everything is good to go. You can even compare the results using the python script. This example should just take a few seconds. If the first one works and not the second one, then it means you have an mpi-related issue. Please post the error message in that case, i'll help you out.

We suggest running the tests kinda as a "double check" measure, but sometimes tiny issue with MPI or something similar can prevent them from being repeatable. In all honesty, I have never tried to run the test from a window machine using WSL. I know that on CLANG on my Mac M4, the test do not all work perfectly.

Thank you for trying out Lethe. If you have any more question, please do write here, i'll be glad to help you.