Phonons: look if we can use Hessian from DFT

[farhi]

VASP:

• use IBRION=5 - 6
• see https://cms.mpi.univie.ac.at/vasp/vasp/IBRION_5_IBRION_6.html

however it looks like the displacements are done still, but internally which saves the generation of the wavefunctions.

[farhi]

The Density-Functional Perturbation Theory (DFPT) does not bring better performences, except for saving the time to generate wavefunctions. So, as it highly depends on the code used, we prefer to rely on ASE/PhonoPy to handle this exactly the same way in a generic implementation.

[farhi]

Concerning QE documentation:

• http://www.quantum-espresso.org/wp-content/uploads/2013/06/phonons_tutorial_shanghai1.pdf
• http://www.fisica.uniud.it/~giannozz/QE-Tutorial/tutorial_phon.html