TD-DFT aborts

[camille]

Hi all,

I wanted to compute a TD-DFT calculation so I followed this tutorial (http://cms.mpi.univie.ac.at/vasp/vasp/B ... tions.html)

For the first calculation, I had no problem (I mean, I think) because the computations finished normally
Here was my INCAR (in bold are the required tag):
____________________________
SYSTEM = CaGaSiO

################### ELECTRONIC SCF CALCULATION ######################

PREC = ACCURATE # Usually Normal for geom. optimisation and Accurate for Etot and DOS calc.
ENCUT = 600 # Cutoff energy of the plane wave basis set (eV)
IALGO = 38 # Algorithm electronic density minimization.

LREAL= FALSE # Auto for geom. opt. and .FALSE. for Etot and DOS calc.
# ISYM = 0 # Do the calculation without sym.
# NELM = 15
# NELECT = 1816
# ICHARG = 11 # = 11 non AUTO
################### GEOMETRY OPTIMIZATION ######################
NELM = 80
IBRION = 2 # Algorithm for the geom. optim.
NSW = 0 # Max. number of iterations / 0 if no geom. optim.
ISIF = 0
# EDIFF = 1.0E-6 # 2 = only atomic positions / 6 = only cell param. / 3 = Full geom. optim.
EDIFFG =-1E-02 # Convergence criterion for Geom. Optim. +:energy (eV) / -:force (eV/A)
# EDIFFG =0.1E-02 # Convergence criterion for Geom. Optim. +:energy (eV) / -:force (eV/A)

################### DOS SMEARING ######################

# ISMEAR = 0 ; sigma = 0.05 # partial occupancy of the wave function (Geom. Optim.)

ISMEAR = -5 # tetrahedron method with Blochl corrections for DOS or Etot
NEDOS = 2000
RWIGS = 1.76 1.22 1.11 0.66
EMIN = -5
EMAX = 10
NPAR=1
LORBIT = 11

LOPTICS = .TRUE.
NBANDS = 220

################### MAGNETISM ########################

#ISPIN = 2
#MAGMOM = 12*0 1*2 7*4 12*-4 48*0
#LDAU = .TRUE.
#LDAUTYPE = 2
#LDAUL = -1 2 2 -1
#LDAUU = 0.0 4.00 3.94 0.
#LDAUJ = 0.0 0.0 0.0 0.
#NUPDOWN = 6

################### NMR CALC. ######################

# LEFG = .TRUE.
# QUAD_EFG = -204.0 150.0 0.0 0.0


################### Hybrid ######################

HSE
LHFCALC = .TRUE. ; HFSCREEN = 0.2 ; AEXX = 0.375
ALGO = ALL ; TIME = 0.2 ; LDIAG = .TRUE.
PRECFOCK = Normal # (coarse grid for HF, soft augmentation charge)
__________


After this computation, I started the second one but it aborts systematically:

OUTCAR :
Found 12 irreducible k-points:

Following reciprocal coordinates:
Coordinates Weight
0.000000 0.000000 0.000000 1.000000
0.500000 -0.000000 -0.000000 1.000000
0.000000 0.500000 -0.000000 1.000000
0.500000 0.500000 -0.000000 1.000000
0.000000 0.000000 0.333333 1.000000
0.500000 -0.000000 0.333333 1.000000
0.000000 0.500000 0.333333 1.000000
0.500000 0.500000 0.333333 1.000000
-0.000000 0.000000 -0.333333 1.000000
0.500000 0.000000 -0.333333 1.000000
0.000000 0.500000 -0.333333 1.000000
0.500000 0.500000 -0.333333 1.000000

Following cartesian coordinates:
Coordinates Weight
0.000000 0.000000 0.000000 1.000000
0.063316 -0.000453 -0.000000 1.000000
-0.000453 0.063316 0.000000 1.000000
0.062863 0.062863 -0.000000 1.000000
-0.000000 0.000000 0.063934 1.000000
0.063316 -0.000453 0.063934 1.000000
-0.000454 0.063317 0.063935 1.000000
0.062863 0.062863 0.063934 1.000000
0.000000 -0.000000 -0.063934 1.000000
0.063317 -0.000454 -0.063935 1.000000
-0.000453 0.063316 -0.063934 1.000000
0.062863 0.062863 -0.063934 1.000000

TETIRR: Found 72 inequivalent tetrahedra from 72
files read and symmetry switched off, memory is now:
total amount of memory used by VASP on root node 177566. kBytes
========================================================================

base : 30000. kBytes
nonl-proj : 79672. kBytes
fftplans : 4759. kBytes
grid : 15620. kBytes
one-center: 373. kBytes
HF : 518. kBytes
nonlr-proj: 5781. kBytes
wavefun : 40843. kBytes

Basis sets for responsefunctions:
=================================
k-point 1 : 0.0000 0.0000 0.0000 plane waves: 5937
k-point 2 : 0.5000-0.0000-0.0000 plane waves: 5932
k-point 3 : 0.0000 0.5000-0.0000 plane waves: 5932
k-point 4 : 0.5000 0.5000-0.0000 plane waves: 5868
k-point 5 : 0.0000 0.0000 0.3333 plane waves: 5903
k-point 6 : 0.5000-0.0000 0.3333 plane waves: 5902
k-point 7 : 0.0000 0.5000 0.3333 plane waves: 5902
k-point 8 : 0.5000 0.5000 0.3333 plane waves: 5932
k-point 9 : -0.0000 0.0000-0.3333 plane waves: 5903
k-point 10 : -0.5000 0.0000-0.3333 plane waves: 5902
k-point 11 : -0.0000-0.5000-0.3333 plane waves: 5902
k-point 12 : -0.5000-0.5000-0.3333 plane waves: 5932

maximum and minimum number of plane-waves per node : 5937 5868

maximum number of plane-waves: 5937
maximum index in each direction:
IXMAX= 13 IYMAX= 13 IZMAX= 8
IXMIN= -13 IYMIN= -13 IZMIN= -8

WARNING: aliasing errors must be expected set NGX to 54 to avoid them
WARNING: aliasing errors must be expected set NGY to 54 to avoid them
WARNING: aliasing errors must be expected set NGZ to 34 to avoid them
exchange correlation table for LEXCH = 8
RHO(1)= 0.500 N(1) = 2000
RHO(2)= 100.500 N(2) = 4000
allocating 0 responsefunctions rank= 5940
NQ= 1 0.0000 0.0000 0.0000,
parallelization over bands

Bands included in the BSE
VB(min)= 1 VB(max)= 92
CB(min)= 93 CB(max)= 184

parameters for screened Coulomb W: AEXX= 0.3750000 HFSCREEN= 0.2000000
BSE (scaLAPACK) single prec attempting allocation of 6.877 Gbyte rank= 101568

total amount of memory used by VASP on root node 7054938. kBytes
========================================================================

base : 30000. kBytes
nonl-proj : 79672. kBytes
fftplans : 4759. kBytes
grid : 15620. kBytes
one-center: 373. kBytes
HF : 518. kBytes
nonlr-proj: 5781. kBytes
wavefun : 40843. kBytes
response : 0. kBytes
bse : 6877372. kBytes

________________

With this incar file :
SYSTEM = CaGaSiO

################### ELECTRONIC SCF CALCULATION ######################

PREC = ACCURATE # Usually Normal for geom. optimisation and Accurate for Etot and DOS calc.
ENCUT = 600 # Cutoff energy of the plane wave basis set (eV)
IALGO = 38 # Algorithm electronic density minimization.

LREAL= FALSE # Auto for geom. opt. and .FALSE. for Etot and DOS calc.
# ISYM = 0 # Do the calculation without sym.
# NELM = 15
# NELECT = 1816
# ICHARG = 11 # = 11 non AUTO
################### GEOMETRY OPTIMIZATION ######################
NELM = 80
IBRION = 2 # Algorithm for the geom. optim.
NSW = 0 # Max. number of iterations / 0 if no geom. optim.
ISIF = 0
# EDIFF = 1.0E-6 # 2 = only atomic positions / 6 = only cell param. / 3 = Full geom. optim.
EDIFFG =-1E-02 # Convergence criterion for Geom. Optim. +:energy (eV) / -:force (eV/A)
# EDIFFG =0.1E-02 # Convergence criterion for Geom. Optim. +:energy (eV) / -:force (eV/A)

################### DOS SMEARING ######################

# ISMEAR = 0 ; sigma = 0.05 # partial occupancy of the wave function (Geom. Optim.)

ISMEAR = -5 # tetrahedron method with Blochl corrections for DOS or Etot
NEDOS = 2000
RWIGS = 1.76 1.22 1.11 0.66
EMIN = -5
EMAX = 10
# NPAR=1
LORBIT = 11

#LOPTICS = .TRUE.
NBANDS = 220

################### MAGNETISM ########################

#ISPIN = 2
#MAGMOM = 12*0 1*2 7*4 12*-4 48*0
#LDAU = .TRUE.
#LDAUTYPE = 2
#LDAUL = -1 2 2 -1
#LDAUU = 0.0 4.00 3.94 0.
#LDAUJ = 0.0 0.0 0.0 0.
#NUPDOWN = 6

################### NMR CALC. ######################

# LEFG = .TRUE.
# QUAD_EFG = -204.0 150.0 0.0 0.0


################### Hybrid ######################

HSE
LHFCALC = .TRUE. ; HFSCREEN = 0.2 ; AEXX = 0.375
ALGO = TDHF ; TIME = 0.2 ; LDIAG = .TRUE.
PRECFOCK = Normal # (coarse grid for HF, soft augmentation charge)

_______________

I am starting to be crazy because I really don't see the mistake. Do you see any problem in my INCAR's file ?

Thx in advance

Best,

[admin]

HF is only implemented for NPAR=number of nodes.
Otherwise communication between nodes would be required during the
calculation of the exchange potential, which would be utterly slow.

[camille]

Hey,

I'll try it.

Last small question, what's the tag for TD-HSE ? Just : TD-HSE ?

Best,

[camille]

Update,

Well, the first step seems OK.
However, the second one is still aborting (see below).

Do you see any reason ?

Best,

--- FROM OUTCAR ---

parameters for screened Coulomb W: AEXX= 0.2500000 HFSCREEN= 0.2000000
BSE (scaLAPACK) single prec attempting allocation of 0.067 Gbyte rank= 8192

total amount of memory used by VASP on root node 139733. kBytes
========================================================================

base : 30000. kBytes
nonl-proj : 4147. kBytes
fftplans : 16740. kBytes
grid : 9842. kBytes
one-center: 49. kBytes
HF : 9. kBytes
nonlr-proj: 517. kBytes
wavefun : 11321. kBytes
response : 0. kBytes
bse : 67108. kBytes

BLAS level 3 operations / number of FFT's:
number of FFTs for wave wavefunctions 800. fft
number of operations in four-orbital integrals -94.46 Gflops, 221200. fft
BSESET: cpu time 84.81: real time 84.82
BSESUM: cpu time 0.00: real time 0.00

total amount of memory used by VASP on root node 111635. kBytes
========================================================================

base : 30000. kBytes
nonl-proj : 4147. kBytes
fftplans : 16740. kBytes
grid : 9842. kBytes
one-center: 49. kBytes
HF : 9. kBytes
nonlr-proj: 517. kBytes
wavefun : 0. kBytes
response : 0. kBytes
bse : 50331. kBytes

BSEDIAG: cpu time 257.59: real time 258.12
BSEOSZI: cpu time 12.30: real time 12.39

--- INCAR ---
################### ELECTRONIC SCF CALCULATION ######################
PREC = ACCURATE # Usually Normal for geom. optimisation and Accurate for Etot and DOS calc.
ENCUT = 400 # Cutoff energy of the plane wave basis set (eV)
IALGO = 38 # Algorithm electronic density minimization.
ADDGRID = T # improve accuracy by adding features into the grid
LREAL= FALSE # Auto for geom. opt. and .FALSE. for Etot and DOS caAUTO
#ISYM = 0 # Do the calculation without sym.
# NELECT = 1816 # number of elec

ISMEAR = 0 ; sigma = 0.01 # partial occupancy of the wave function (Geom. Optim.)
NBANDS = 128
# ISMEAR = -4 # tetrahedron method with Blochl corrections for DOS or Etot
NEDOS = 4000
# RWIGS = 1.27 0.82 0.37
EMIN = -20
EMAX = 20
# NPAR = 1 # full parallel
LORBIT= 11 # LORBIT = 0 | 1 | 2 | 5 | 10 | 11 | 12 RWIGS required for 0-5
# NBANDS = 24

HSE
LHFCALC = .TRUE. ; HFSCREEN = 0.2 ; AEXX = 0.25
ALGO = TDHF ; TIME = 0.2 ; LDIAG = .TRUE.
PRECFOCK= FAST # (coarse grid for HF, soft augmentation charge)


--- OSZICAR ---
energies w=

--- job-vasp-out ---
LDA part: xc-table for Pade appr. of Perdew
found WAVECAR, reading the header
POSCAR, INCAR and KPOINTS ok, starting setup
FFT: planning ...
reading WAVECAR
the WAVECAR file was read successfully
initial charge from wavefunction
the WAVEDER file was read successfully
energies w=

responsefunction array rank=2952
LDA part: xc-table for Pade appr. of Perdew
allocating 0 responsefunctions rank= 2952
allocating 0 responsefunctions rank= 2952
Doing 0 frequencies on each core in blocks of 0
allocating two-electron 4 orbital integral table 32 32 32 32
BSE (scaLAPACK) single prec attempting allocation of 0.067 Gbyte rank= 8192
BSE setting up matrix
|.........|....... BSE redistributing all elements

BSE diagonalizing matrix (scaLAPACKaware)
BSE calculating oscillator strength
VASP MPI job : end

___________________________________________________________