Internal errors: mkpoints_full.F & CHECK_GEN_LAYOUT

[EmilHThomsen]

I continually reproduce the same error when running a HSE06 electronic band structure calculation for my system.

Here is the direct !BUG! output:
-----------------------------------------------------------------------------
| _ ____ _ _ _____ _ |
| | | | _ \ | | | | / ____| | | |
| | | | |_) | | | | | | | __ | | |
| |_| | _ < | | | | | | |_ | |_| |
| _ | |_) | | |__| | | |__| | _ |
| (_) |____/ \____/ \_____| (_) |
| |
| internal error in: mkpoints_full.F at line: 984 |
| |
| internal error in CHECK_GEN_LAYOUT: G-vectors changed at ik:65 old: |
| 5228 new: 5221 |
| |
| If you are not a developer, you should not encounter this problem. |
| Please submit a bug report. |
| |
-----------------------------------------------------------------------------

Calulation details:

VASP version: 6.5.0 (16Dec24)

System:
Materials Project ID: mp-30943 (MgPSe3)

Workflow:

• 1. Structural relaxation (EDIFFG=-0.02, EDIFF=1E-6, SYMPREC=1E-6)

• 2. PBE SCF

• 3. HSE06 SCF

• 4. HSE06 electronic band structure using KPOINTS_OPT

The HSE06 SCF converges successfully and produces valid WAVECAR/CHGCAR files.

The subsequent band-structure calculation fails reproducibly with:

internal error in CHECK_GEN_LAYOUT:
G-vectors changed at ik:65
old: 5228
new: 5221

The calculation reaches approximately k-point batch 41/180 before terminating.

Relevant INCAR:

• SYSTEM = mp-30943 HSE06 band structure

  ENCUT = 450
  PREC = Accurate
  EDIFF = 1E-06

  LHFCALC = True
  AEXX = 0.25
  HFSCREEN = 0.2
  HFRCUT = -1
  PRECFOCK = Fast
  ALGO = Normal
  TIME = 0.5

  ISMEAR = 0
  SIGMA = 0.05

  IBRION = -1
  NSW = 0

  ISPIN = 1

  ISYM = -1
  SYMPREC = 1E-6

  LREAL = False
  LASPH = True
  ADDGRID = True

  LORBIT = 11
  LMAXMIX = 4

  ISTART = 1
  ICHARG = 0

  LWAVE = False
  LCHARG = False

  NELM = 120
  NELMIN = 4

  KPAR = 1
  NCORE = 1

  KPOINTS_OPT_NKBATCH = 1

The KPOINTS_OPT file was generated using pymatgen HighSymmKpath on the relaxed structure with symprec=1E-6.

Tests already performed:

• ISYM = -1, 0, 3
  ISPIN = 1 and ISPIN = 2
  KPOINTS_OPT_NKBATCH present and absent
  Failure was reproduced using the explicit zero-weight path method rather than only KPOINTS_OPT.
  Fresh PBE SCF and fresh HSE06 SCF
  Standardized primitive structure generated using pymatgen SpacegroupAnalyzer with symprec=1E-6

The error remains reproducible.

The calculation uses:

• 4x4x4 (gamma-sampled) SCF mesh
  KPOINTS_OPT line-mode path generated by Pymatgen (on the Standardized, relaxed, primitive structure).
  WAVECAR restart from converged HSE06 SCF

I can provide POSCAR, INCAR, KPOINTS, KPOINTS_OPT, OUTCAR, and full stdout if needed.

[EmilHThomsen]

I did not post many input files so they are attached below.

I want to highlight, that it runs the individual Kpoint calculations until Kpoint no. 41 (completes 40/180), before this error occurs.

Here is my KPOINT_OPT file:
High-symmetry path (KPOINTS_OPT) for hybrid band structure
20
Line_mode
Reciprocal
0.0 0.0 0.0 ! \Gamma
0.5 0.0 0.0 ! L

0.5 0.0 0.0 ! L
0.5 0.2163817878187575 -0.2163817878187575 ! B_1

0.7836182121812425 0.5 0.2163817878187575 ! B
0.5 0.5 0.5 ! Z

0.5 0.5 0.5 ! Z
0.0 0.0 0.0 ! \Gamma

0.0 0.0 0.0 ! \Gamma
0.35819089390937875 0.0 -0.35819089390937875 ! X

0.6418091060906213 0.35819089390937875 0.0 ! Q
0.5 0.5 0.0 ! F

0.5 0.5 0.0 ! F
0.6418091060906213 0.6418091060906213 0.2163817878187575 ! P_1

0.6418091060906213 0.6418091060906213 0.2163817878187575 ! P_1
0.5 0.5 0.5 ! Z

0.5 0.0 0.0 ! L
0.7836182121812425 0.35819089390937875 0.35819089390937875 ! P

My Kpoint_file:
Automatic kpoint scheme
0
Gamma
4 4 4