internal error in: rot.F at line: 793

[zhaoru_sun1]

Dear Support,

When I set "ALGO = All" to optimize the structure of the Li slab in vasp6.2.0, it showed the following error message:

Code: Select all

 
 -----------------------------------------------------------------------------
|                     _     ____    _    _    _____     _                     |
|                    | |   |  _ \  | |  | |  / ____|   | |                    |
|                    | |   | |_) | | |  | | | |  __    | |                    |
|                    |_|   |  _ <  | |  | | | | |_ |   |_|                    |
|                     _    | |_) | | |__| | | |__| |    _                     |
|                    (_)   |____/   \____/   \_____|   (_)                    |
|                                                                             |
|     internal error in: rot.F  at line: 793                                  |
|                                                                             |
|     EDWAV: internal error, the gradient is not orthogonal 1 1 -1.817e-3     |
|                                                                             |
|     If you are not a developer, you should not encounter this problem.      |
|     Please submit a bug report.                                             |
|                                                                             |
 -----------------------------------------------------------------------------

INCAR

Code: Select all

#-------------------------[common]
 ISTART = 0
 NCORE = 16
# PREC = Normal # default
 PREC = Accurate
# LREAL = .FALSE. # default
# LREAL = Auto
 ENCUT = 500
# ENCUT = 800 # volume opt

#-------------------------[smearing]
# ISMEAR = 1 # default, for metal, Methfessel-Paxton smearing order N
# SIGMA = 0.2 # default
# ISMEAR = 0 # Gaussian smearing
# SIGMA = 0.05

#-------------------------[electron relaxation]
# ALGO = Normal # default, DAV
# ALGO = Fast # DAV + RMM
# ALGO = VeryFast # RMM
# ALGO = Damped
 ALGO = All # CG
# NELM = 60 # default
 NELM = 500
# EDIFF = 1.0E-4 # default
 EDIFF = 1.0E-5

#-------------------------[ionic relaxation]
# NSW = 0 # default
 NSW = 500
# IBRION = 0 # default, MD
 IBRION = 1 # quasi-Newton
# IBRION = 2 # CG
# EDIFFG =  EDIFF×10 # default
# EDIFFG = -0.02
 EDIFFG = -0.01
# ISIF = 2 # default, atom position
# ISIF = 3 # volume
# IOPTCELL =  1 0 0 1 1 0 1 1 1   # 0=fix, 1=relax

#-------------------------[spin]
# ISPIN = 1 # default
 ISPIN = 2
# MAGMOM = NIONS * 1.0 # default
# MAGMOM = 300*0 8*-4 7*4  1000*0
# LORBIT = None # default
 LORBIT = 11 # mag per atom output

#-------------------------[surface]
# LDIPOLE = .TRUE.  # slibe dipole correction
# IDIPOLE = 3   # z direction

#-------------------------[file]
# LWAVE = .TRUE. # default
 LWAVE = .FALSE. # WAVECAR not write
# LCHARG = .TRUE. # default
 LCHARG = .FALSE. # CHAGCAR not write

KPOINTS

Code: Select all

A
0
M
7 7 1

POSCAR

Code: Select all

Converted by xsd2pos.py
1.0
     3.44486000     0.00000000     0.00000000
     0.00000000     3.44486000     0.00000000
     0.00000000     0.00000000    65.59047010
  Li
  30
Direct
     0.00000000     0.00000000     0.05400830
     0.00000000     0.00000000     0.10652904
     0.00000000     0.00000000     0.15904978
     0.00000000     0.00000000     0.21157052
     0.00000000     0.00000000     0.26409126
     0.00000000     0.00000000     0.31661200
     0.00000000     0.00000000     0.36913274
     0.00000000     0.00000000     0.42165348
     0.00000000     0.00000000     0.47417422
     0.00000000     0.00000000     0.52669496
     0.00000000     0.00000000     0.57921570
     0.00000000     0.00000000     0.63173644
     0.00000000     0.00000000     0.68425718
     0.00000000     0.00000000     0.73677792
     0.00000000     0.00000000     0.78929866
     0.50000000     0.50000000     0.02774793
     0.50000000     0.50000000     0.08026867
     0.50000000     0.50000000     0.13278941
     0.50000000     0.50000000     0.18531015
     0.50000000     0.50000000     0.23783089
     0.50000000     0.50000000     0.29035163
     0.50000000     0.50000000     0.34287237
     0.50000000     0.50000000     0.39539311
     0.50000000     0.50000000     0.44791385
     0.50000000     0.50000000     0.50043459
     0.50000000     0.50000000     0.55295533
     0.50000000     0.50000000     0.60547607
     0.50000000     0.50000000     0.65799681
     0.50000000     0.50000000     0.71051755
     0.50000000     0.50000000     0.76303829

All the related files have been attached. Not sure if this is a bug or not. Thank you!

[merzuk.kaltak]

Dear Zhaoru Sun,

I have successfully converged your input files with vasp-6.2.0 using two different toolchains (software-stacks).
gnu_mkl.zip is obtained with vasp-6.2.0 compiled with gfortran-9.3.0 + openmpi-4.0.5 + Intel-MKL-2019.5.281 and run on two nodes with 8 cpus each connected with infiniband cables.
pxe.zip is obtained with vasp-6.2.0 compiled with intel-parallel-studio-xe-2020.2.254 and run on 8 nodes with 8 cpus each (same number of MPI ranks as you have chosen for your run).
Both tests give the same results (up to a specific precision), compare the CONTCAR or OSZICAR files for instance.

I strongly suspect that something is not properly set up with your environment.
It seems you are using some Intel software stack. Which version of Intel-compilers, Intel-MPI and Intel-MKL are you using?
Is it listed on our recommended toolchain list?