I would like that the simulation will be faster, so that more particles can be simulated.
Problem is that when the simulation is running there appears warning: Particle (number of particle) underwent maximum number of events. I have a limited model with vacuum boundary conditions (the same problem was with reflective boundary conditions). Despite the warning the simulation is performed and I get the results, but it seems to me that therefore the simulation is slower.
In order to make the simulation faster, I also used parallelization for multi-core computing. I used the command:
openmc.run(threads=24)
Despite 24 available kernels, the simulation runs at a maximum of 8 kernels.
Sorry I missed your message when it first came through. Were you ever able to figure out the issue with your model? The error message likely indicates that there is some modeling error, but without seeing your model it’s impossible to know.
As for the question about running faster using parallelism, the maximum speedup you’re able to get depends on the hardware you’re running on. What does your hardware setup look like?
Hi Ingrid,
I meet the same problem with you. [warning: Particle (number of particle) underwent maximum number of events.]
Have you figured it out? Hope to get your reply!
This error happens if a particle has undergone a million events, an event being a collision or a surface crossing. This is usually indicative of some error in the definition in your model. Things to consider would be:
Have you applied boundary conditions to the outermost surfaces in your problem?
Are the densities of your materials reasonable?
Have you checked for overlapping cell definitions? You can do this by running openmc -g or openmc --geometry-debug. If it finds two cells that occupy the same location in the geometry, it will report an error. Without this check turned on, if you have two cells that overlap one another, the behavior is undefined.
If you are comfortable sharing your model, someone might be able to help you debug it as well.
We welcome any and all questions, especially if you are having trouble using the code, and will do our best to help, but please try to be considerate of others on this mailing list. If you have a question, asking it multiple times in different threads does not make it more likely to get an answer, and will only lead to more confusion (and fill up people’s inboxes).