Open ndaelman-hu opened 11 months ago
Note: when the ionic positions are updated and the main file is vasprun.xml
, the <kpoints>
should be derived from <calculations>
, else from the "standalone" tag (under <modeling>
).
This is apparently a new VASP feature (as of version 6?), so especially with vasprun.xml
one has to distinguish between the placement of <kpoints>
: under modeling
or calculation
(only in newer versions, not sure about the single-point case).
Correction, the cause is not the VASP version, but the routines chosen to perform frequency / phonon calculations. I came to this conclusion after finding another example that uses IBRION = 6: entry_id = y0MhHRV_TIb_zamDSdM2QB6o84bF
During a geometry optimization or frequency calculation, the symmetry of the cell can change. When symmetry-reduction is active, the number of reduced k-point can change by ionic update. This causes errors when reshaping eigenvalues extracted from the OUTCAR, which are contingent on the (original) no. irreducible k-points:
The frequency calculation that brought this error to light: trash:///freqsafd_pbe.zip