Closed bcm0 closed 3 years ago
h5py uses pytest-mpi (that's why I wrote it).
I suspect in your case though the problem is hypothesis starts with different random numbers or uses a different seed - you probably need to configure it (or the environment it runs in) to make sure both processes use the same ordering of numbers.
I found hypothesis seeds and now my problem is fixed.
Thank you very much!
mpirun -n 2 python -m pytest --with-mpi --hypothesis-seed 1234
Like suggested in the docs I run
mpirun -n 2 python -m pytest --with-mpi
.The input arguments get mixed and the assert fails. Most likely this is due to the tests being run twice with
mpirun -n 2
. I would love to see a working example.