- mpi results and non mpi results can differ around eps
- multiprocessing has initial high cpu values with model.solver
- simpli.pixel_size does not call property
- nan values detected in light signal (for 90degree fibers?)
- simpli simulation has high initial copying and allocation time
- doc generation warnings about Classes: stub file not found
- overlap warning only with multiple fiber bundles
- fiber_bundles have 3 copies, 1 in python, and 2 in cpp (fbs_, fbs_org_)
- interpolation for mu, dn
- stokes vs jones
- polarization value px, py
- test cells
- test mpi on multiple nodes
- warning if fiber radius, segment length and r_min are questionable
- warning if collision close to 0 but "never ending"
- memory warning for fiber_bundles (gets quite big for small segment lengths)
- non linear splitting and merging
- segment length per fiber, automatic determination for each fiber fun(f_radius)