VASP doesn't save stresses to HDF when run on ISIF=2 (default)

[pmrv]

I'm running some VASP calculations and would like to have the full stress tensor after the (static) calculation. I've checked that my OUTCAR does contain the information. (Let me know if you need an example, github doesn't let me upload tar archives).

I've traced the code and it seems we do parse the stresses, but then only save the averaged pressure. I've modified that code in this PR to save also the stress to the log_dict, which I thought should be enough to get it into HDF5, but from my quick testing that doesn't seem to be the case.

Any ideas where to put this so it works?

[coveralls]

Pull Request Test Coverage Report for Build 1341110897

• 2 of 3 (66.67%) changed or added relevant lines in 2 files are covered.
• 451 unchanged lines in 8 files lost coverage.
• Overall coverage increased (+1.5%) to 69.769%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pyiron_atomistics/vasp/base.py	1	2	50.0%

Files with Coverage Reduction	New Missed Lines	%
pyiron_atomistics/atomistics/structure/strain.py	1	97.09%
pyiron_atomistics/init.py	1	98.0%
pyiron_atomistics/atomistics/job/structurecontainer.py	17	74.39%
pyiron_atomistics/atomistics/structure/factory.py	20	76.8%
pyiron_atomistics/atomistics/master/murnaghan.py	32	72.15%
pyiron_atomistics/sphinx/base.py	99	82.07%
pyiron_atomistics/lammps/interactive.py	137	35.44%
pyiron_atomistics/atomistics/structure/atoms.py	144	73.95%

Totals
Change from base Build 1212254905:	1.5%
Covered Lines:	11410
Relevant Lines:	16354

---

💛 - Coveralls

[sudarsan-surendralal]

Since this is inside an elif statement, this segment of the code would work only when vasprun files are absent. You would need to put this in if statement block:

pyiron_atomistics/pyiron_atomistics/vasp/base.py

Line 1947 in d1d3ed1

if outcar_working:

[pmrv]

This works, but the pressures come out in units of kB. I'll have to check first how to convert.

[samwaseda]

Just for me to understand: Is stress a per-atom quantity? Or is it simply meant to be pressure with the opposite sign?

[pmrv]

Just for me to understand: Is stress a per-atom quantity? Or is it simply meant to be pressure with the opposite sign?

Here stress is the full tensor in voigt notation, what is saved as pressures is the average stress.

[samwaseda]

Here stress is the full tensor in voigt notation, what is saved as pressures is the average stress.

OH! That's a suuuuuuuuuper questionable definition. At least it's not compatible with the LAMMPS definition at all. What's actually behind the idea of having the average pressure in the first place, instead of having nothing or the full tensor?

[pmrv]

@sudarsan-surendralal Are you aware of anyone in our group that extensively uses VASP "pressures" via pyiron? I'd like the output of lammps and vasp to be consistent, which in this case would be saving the full 3x3 matrix under "pressures". For know I can leave the field in VASP untouched, but it'd be good to identify a time window after which we can make the breaking change "stress" -> "pressures".

[sudarsan-surendralal]

@sudarsan-surendralal Are you aware of anyone in our group that extensively uses VASP "pressures" via pyiron? I'd like the output of lammps and vasp to be consistent, which in this case would be saving the full 3x3 matrix under "pressures". For know I can leave the field in VASP untouched, but it'd be good to identify a time window after which we can make the breaking change "stress" -> "pressures".

Sorry, I'm not aware if anyone uses pressures from VASP (maybe @jan-janssen). I'm not too familiar with the definitions of stress and pressure in the context of DFT simulations but I remember reading (from this book) that the definitions are not that straightforward at least wrt. LAMMPS. I have the feeling that this needs to be discussed in our developers meeting or even in a purely scientific meeting since there could be non-pyiron users who have a better idea.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[pmrv]

I'll merge this today as-is and work around the inconsistent naming on my end for a while, but we should be this on the radar.