tally_fission_rates mpi problem

[John Mayers]

Hello, dear developers.

While working, I encountered the following problem: when using tally_fission_rates from process.py in MPI mode, the program crashes.

I looked into the code and found the reason. The problem is in this line: num_fsrs = int (geometry.getNumTotalFSRs())

Since there are fewer fsrs in the domain than in the entire geometry, a crash occurs. I tried to solve this problem by replacing getNumTotalFSRs with getNumFSRs and after calculating, perform the MPI.Allreduce operation with the summation operation. The program worked, but at the exit I got complete nonsense. I conduct tests on the C5G7 benchmark and when trying to get the fission reaction rates in this way, I got an array of arbitrarily scattered values that absolutely does not agree with the geometry.

Tell me, please, is there any easy way to fix this problem? I tried to solve it myself by adding an algorithm similar to the one implemented in the Solver class for calculating fission rate reactions, however, there is a getLocalFSRId method that does not work through Python interfaces, since it requires a variable of type long to enter. Is there another way to do this?

Thank you in advance for your reply.

Best Wishes,

John Mayers