Using previous depletion results in future runs

Hello,

I want to run a more complicated depletion run and the supercomputer I work with has yet to work with multithreading. At the most I could get was one cpu, even though I would try to ask for more. I tried to get OpenMPI to work or increasing the particle count, but neither worked for multithreading. I ran into issues such as ERROR: Failed to open HDF5 file with mode ‘w’: summary.h5, when I tried to use OpenMPI to multithread.

I have tried to use ThreadPoolExecutor from concurrent.futures library in Python. However, the code never worked. Is there a simpler way of multithreading? Or is it possible to use MPI as I have been trying? Has anyone else figured a way of using MPI for multithreading or has addressed the problem I am having?

Because I never figured this issue out, I moved onto trying to break the depletion run into smaller runs that can be ran on one cpu and can be under the time limit for the supercomputer. I want all the small runs to use the previous depletion run results, so it would be like running the bigger depletion run without the struggles of trying to get the supercomputer to multithread. Is there a way in OpenMC that I can start the depletion run where the previous run finished? If so, how do I do so in Python?

Thanks

If I don’t misunderstand what you say, you can try to set the prev_res in CoupledOperator, see details in openmc.deplete.CoupledOperator — OpenMC Documentation

Thank you. That helped a lot and it worked for a while. However, I changed the month intervals from 6 months to 2 months and I ran into some issues. The code would run for the time I gave it, but when I tried to run the code again starting at the previous depletion results, it stopped after 3 minutes. I looked into the slurm file and I get this error: openmc.exceptions.DataError: Nuclide “Xe123” has not been loaded yet. I have checked to see if I made any other changes that could have cause it, but I did not find any. Have you run into this issue before? Is there a simple fix for this that I could put in my code?

This look like Xe123 is missing in your nuclide library. You can download the Official Data Libraries in here.