Phonon scattering is so different

[caolz]

Firstly I trained a neural network potential, after that I used tutorial 8: Thermal conductivity calculations with the SSCHA, got the final_dyn and saved it in phononpy format.

This diagram is the one I drew using tutorial 1 to read final_dyn, which is basically the same as VASP using phonopy.

I drew this picture using FORCE_CONSTANTS saved as a phonopy format, and tried VASP format vs QE format, but the difference is too great as well.

I wonder if the developers haven't seen anything similar?

[mesonepigreco]

Hi, phonopy export is broken unfortunately. There is a workaround implemented by @DjordjeDangic here

#87

Check if this works for you