Deprecated and deprecating functionalities

[GiudGiud]

This issue is meant to discuss which functionalities you think have been deprecated in the code, and which functionalities will or should be deprecated.

• In the 3D MOC branch, I think we should deprecate the backup CMFD solver. It's not necessary to have a secondary solver which is a priori less stable than the main one (to show/prove), since it has a coarser group structure. It would clean up the CMFD file a bit to delete it
• In the 3D MOC branch, On-The-Fly transport (transport at the same time as ray-tracing, on the fly), seems to be only partly implemented, possibly because it was working at some point then got deprecated. As MOCkingbird and John's code are going this way, we might want to fix that option.
  "Reducing 3D MOC Storage Requirements with Axial On-the-fly Ray Tracing" by G. Gunow et al. shows it's 20% slower than the current way but the solver is 5 times faster now than then so that might have changed.
• The CPU solvers are now vectorized. VectorizedSolver which follows a different approach should be re-tested and most probably deleted
• The spectrumCalculator, initialization sweeps and chi guess, used to get a "better" starting point for the simulation, dont seem to be required anymore, even for the 3D full core case which is notoriously hard to converge. Besides, initialization sweeps are expensive and chi guess creates 0 fluxes which aren't great for CMFD.

[GiudGiud]

As an update, OTF transport has been fixed and is still 20% slower than ray tracing for the whole s,z plane followed by transport of track angular fluxes.