FW-CADIS StatePoint output

[nelsonag]

Description

When running TRRM for FW-CADIS purposes in a model that has user-specified tallies , the forward solution statepoint is overwritten by the adjoint solution statepoint.

I have found it useful to be able to observe the FW-CADIS forward flux distribution to compare with the final solution obtained in MC, and as such I have had to run TRRM three times, a forward solution with tallies on , and then the two solutions needed for TRRM. This is ultimately wasteful as its just to recover information on disk that is overwritten.

Alternatives

Either the Forward TRRM statepoint is given a different filename from the adjoint, or, the two statepoint files are somehow merged (perhaps by adding "adjoint" to the name of tallies from the adjoint run). The former is simple, the latter is complicated. Both will suffice.

Compatibility

No impact

[jtramm]

This is a great point! Yes, as it is now, it's quite wasteful that the forward flux statepoint is lost when running in adjoint mode. As you said, it would be nice to compare the forward/reverse solutions without having to rerun things multiple times.

Probably I would prefer to (conditionally) rename the output files to be forward.statpoint.1000.h5, forward_tallies.out, forward_plot_1.vtk etc when the code detects it has been run in random ray adjoint mode, as this is simplest to implement. I'll give this a try in the next few weeks!