Feature ddi fix lattice constant

[MSallermann]

DDI:

• added second mu_s scaling to DDI gradient -> is now correctly calculated as the force experienced by the spins (before it was the effective field)
  • the distance between spins is now correctly scaled with the lattice constant
  • made the DDI unit test more general (mu_s != 1, lattice_constant != 1 and non-normalized bravais-vectors)

Geometry (maybe discuss if this makes sense):

• geometry->cell_atoms is now not longer scaled with the lattice_constant in the constructor of the geometry class. This makes the behaviour of the bravais vectors and basis vectors more uniform. This only changes how a distance is calculated within the code and nothing for the user! Before it was: dist = lattice_constant * ( da * bravais_vector[0] + ... ) + cell_atoms[i] now it is: dist = lattice_constant * ( da * bravais_vector[0] + ... + cell_atoms[i] )

Other:

• fixed a small error in the calculation of the topological charge where the position of the first basis atom was calculated wrongly (would not have mattered 99.9% of the time because it should always lie at the lattice site)

[codecov[bot]]

Codecov Report

Merging #495 into develop will decrease coverage by 0.07%. The diff coverage is 72.09%.

@@             Coverage Diff             @@
##           develop     #495      +/-   ##
===========================================
- Coverage    51.51%   51.43%   -0.08%     
===========================================
  Files           82       82              
  Lines         9473     9454      -19     
===========================================
- Hits          4880     4863      -17     
+ Misses        4593     4591       -2

[coveralls]

Coverage decreased (-0.1%) to 84.454% when pulling e8d408b639db908d72dc219ee2b8a4e12ab8e4e2 on MSallermann:feature_ddi_fix_lattice_constant into d06f4af87cac5ac611f3dbf19a4bfe61b22c9118 on spirit-code:develop.

[GPMueller]

Please test the CUDA version once more after my changes and let me know if everything works as expected.