Is isotropic minimization really what we want?

[samwaseda]

atomistics/src/atomistics/calculators/lammps/commands.py

Line 17 in c548214

LAMMPS_MINIMIZE_VOLUME = "fix ensemble all box/relax iso 0.0"

As this page states, iso 0 means the sum of the diagonal components of the pressure tensor becomes zero, while aniso 0 means each of the diagonal components must become zero. Is iso 0 really what we want?

[jan-janssen]

I guess we primarily took it from pyiron_atomistics:
pyiron/pyiron_atomistics@75b6b17

Make LammpsControl pressure handling make sense.
Behavior for the "pressure" option is now:

• scalar -> "iso"-style simulation
• list: Voigt notation of pressure tensor
• isotropic, hydrostatic pressures are allowed in any case
• otherwise lists with None in them will be forbidden if the cell has to
  be rotated
• everything goes for a non-rotated cell

[samwaseda]

yes and now it's hard-coded, which from my point of view is problematic

[jan-janssen]

yes and now it's hard-coded, which from my point of view is problematic

Sounds good, do you want make a pull request to make it more compatible to what we had before?

[samwaseda]

Fine