Open wbinventor opened 8 years ago
What behavior do you want a distribcell on an outer
to have? It would be really difficult to add one element in the results array per instance of outer
universe, but it would be pretty easy to lump all of the outer
results into a single value in the array.
@smharper - the first of your two proposals is what I think would be best as it would be consistent with how the current distribcell implementation works for cells in non-outer
universes.
That will require dynamically resizing the tally results
arrays so it will likely interfere with #559. If it's okay with you, I'm not going to address this until #559 is closed
@smharper sure that sounds like a good plan.
It appears that the tallies with a "distribcell" filter on cells in lattice
outer
universes do not work. Although the code runs fine with such tallies, the results array in the tally's entry in the statepoint file is empty.I recognize that this is a corner case and that it is tough to imagine a case where one would want to use "distribcell" tallies in lattice
outer
universes. However, I could see one use case if one wished to tally in each of the gaps between fuel assemblies, which may be most conveniently defined with the latticeouter
attribute.The reason this came up is that I added an
outer
universe to some of my lattices per the discussion with @smharper on #546. When one generates anopenmc.MGXS
library with "distribcell" domains, tallies will be created for those cells in latticeouter
universes. Everything is fine unless one tries to loop over the domains in the library to perform operations on theMGXS
or baseTallies
in downstream data processing (e.g., distribcell tally summations per #550).