Page 1 of 1

EPC g(q) MoS2-ML : SIGSEGV: Segmentation fault

Posted: Thu Mar 14, 2019 10:04 am
by SaraFiore
Dear Experts,
I am trying to compute the electron-phonon coupling matrix elements of a monolayer of MoS2
But when I try to run EPW.X I get the following error

Code: Select all

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x2b921484b26f in ???
#1  0x2b9214893c15 in ???
#2  0x2b921489610b in ???
#3  0x5abe49 in setup_dgc_
   at /scratch/local/q-e-qe-6.3/LR_Modules/setup_dgc.f90:53
#4  0x436c1e in epw_setup_
   at /scratch/local/q-e-qe-6.3/EPW/src/epw_setup.f90:210
#5  0x406133 in epw
   at /scratch/local/q-e-qe-6.3/EPW/src/epw.f90:102
#6  0x405fac in main
   at /scratch/local/q-e-qe-6.3/EPW/src/epw.f90:23


Any help will be appreciated! thank you!!!

I used

mpirun -np 4 pw.x < scf.in > scf.out
mpirun -np 4 pw.x < nscf.in > nscf.out
mpirun -np 4 epw.x -npool 4 < epw.in > epw.out
(I also tried with -np 1)
while for phonons
[...]
#SBATCH --nodes=4
#SBATCH --ntasks-per-core=2
#SBATCH --ntasks-per-node=12
#SBATCH --cpus-per-task=2
[...]
srun ph.x -in ph.in


and as inputs

(PP from http://www.quantum-simulation.org/potentials/sg15_oncv/)

scf.in

Code: Select all

&control
     calculation='scf'                   
     title ='MoS2'
     restart_mode='from_scratch'
     pseudo_dir='../PHONON/'
     outdir='./'
     prefix='MoS2',
  /
  &system
     ibrav=0
     nat=3,
     ntyp=2,
     ecutwfc = 150,
     ecutrho = 1500,
  /
  &electrons
     conv_thr = 1.0d-13                                 
  /
 ATOMIC_SPECIES
  Mo  95.94    Mo_ONCV_PBE-1.0.upf
  S   32.065   S_ONCV_PBE-1.1.upf

ATOMIC_POSITIONS angstrom
Mo          -0.0000000000       1.8386990572      11.5773510440
S            1.5923600935       0.9193495287      10.0139409592
S            1.5923600935       0.9193495287      13.1407611289

K_POINTS automatic
16 16 1 0 0 0

CELL_PARAMETERS angstrom
      3.1847201870      -0.0000000000       0.0000000000
     -1.5923600935       2.7580485859       0.0000000000
      0.0000000000       0.0000000000      23.1547020881


ph.in

Code: Select all

phonons of MoS2
 &INPUTPH
  alpha_mix(1) =   4.0000000000d-01
  epsil = .true.
  fildrho = 'drho'
  fildvscf = 'dvscf'
  fildyn = './MoS2.dyn'
  ldisp = .true.
 !recover = .true.
  nq1 = 8
  nq2 = 8
  nq3 = 1
  outdir = './'
  prefix = 'MoS2'
  tr2_ph =   1.0000000000d-18
/


nscf.in

Code: Select all

&control
     calculation='nscf'                   
     pseudo_dir='../PHONON.'
     outdir='./'
     prefix='MoS2',
     wf_collect      = .false.

  /
  &system
     ibrav=0
     nat=3,
     ntyp=2,
     ecutwfc = 150,
     ecutrho = 1500,
     nbnd = 11

  /
  &electrons           
    diagonalization = 'david'
    mixing_beta     = 0.7
    conv_thr        = 1.0d-10             
  /
 ATOMIC_SPECIES
  Mo  95.94    Mo_ONCV_PBE-1.0.upf
  S   32.065   S_ONCV_PBE-1.1.upf

ATOMIC_POSITIONS angstrom
Mo          -0.0000000000       1.8386990572      11.5773510440
S            1.5923600935       0.9193495287      10.0139409592
S            1.5923600935       0.9193495287      13.1407611289

CELL_PARAMETERS angstrom
      3.1847201870      -0.0000000000       0.0000000000
     -1.5923600935       2.7580485859       0.0000000000
      0.0000000000       0.0000000000      23.1547020881
K_POINTS crystal
256
  0.00000000  0.00000000  0.00000000  3.906250e-03
  0.00000000  0.06250000  0.00000000  3.906250e-03
  0.00000000  0.12500000  0.00000000  3.906250e-03
  0.00000000  0.18750000  0.00000000  3.906250e-03
  0.00000000  0.25000000  0.00000000  3.906250e-03
  0.00000000  0.31250000  0.00000000  3.906250e-03


epw.in

Code: Select all

--
&inputepw
  prefix      = 'MoS2'
  amass(1)    =  95.94
  amass(2)    = 32.065
  outdir      = './'
  dvscf_dir   = '../PHONON/save'

  elph        = .true.
  kmaps       = .false.
  epbwrite    = .true.
  epbread     = .false.
  epwwrite    = .true.
  epwread     = .false.

  wannierize  = .true.
  nbndsub     =  11
  nbndskip    =  0
  num_iter    = 300
  proj(1)     = 'Mo:d'
  proj(2)     = 'S:p'   

  elecselfen  = .false.
  phonselfen  = .false.
  a2f         = .false.
  prtgkk      = .true.


  fsthick     = 7.0
  eptemp      = 20
  degaussw    = 0.05
 
  filqf       = 'path.dat'
  nkf1        = 1
  nkf2        = 1
  nkf3        = 1

  nk1         = 16
  nk2         = 16
  nk3         = 1
  nq1         = 8
  nq2         = 8
  nq3         = 1
 /   
  10
   0.000000000000000E+00   0.000000000000000E+00   0.000000000000000E+00
   0.000000000000000E+00   0.144337567296724E+00   0.000000000000000E+00
   0.000000000000000E+00   0.288675134593449E+00   0.000000000000000E+00
   0.000000000000000E+00   0.433012701890173E+00   0.000000000000000E+00
   0.000000000000000E+00  -0.577350269186897E+00   0.000000000000000E+00
   0.124999999999993E+00   0.216506350945087E+00   0.000000000000000E+00
   0.124999999999993E+00   0.360843918241811E+00   0.000000000000000E+00
   0.124999999999993E+00   0.505181485538535E+00   0.000000000000000E+00
   0.249999999999985E+00   0.433012701890173E+00   0.000000000000000E+00
   0.249999999999985E+00   0.577350269186897E+00   0.000000000000000E+00


Respectfully,

Sara Fiore

Re: EPC g(q) MoS2-ML : SIGSEGV: Segmentation fault

Posted: Thu Mar 14, 2019 3:17 pm
by Vahid
I see two issues:

1. mpirun -np 4 epw.x -npool 4 < epw.in > epw.out is fine but for the preliminary scf and nscf calculations, the -npool 4 is missing. Maybe this is not required in the latest version of EPW.

2. The ONCV PSPs are generally Hamman's norm conserving PSPs. Did you generate US PSPs? If not, the ecutrho is unnecessary for NC PSPs and if specified, should be 4*ecutwfc not 10*ecutwfc.

Cheers,
Vahid

Vahid Askarpour
Department of Physics and Atmospheric Science
Dalhousie University,
Halifax, NS, Canada

Re: EPC g(q) MoS2-ML : SIGSEGV: Segmentation fault

Posted: Fri Mar 15, 2019 3:14 pm
by SaraFiore
Dear Vahid Askarpour,
Thank you for you answer!
using
mpirun -np 4 pw.x -npool 4 < scf.in > scf.out
mpirun -np 4 pw.x -npool 4 < nscf.in > nscf.out
mpirun -np 4 epw.x -npool 4 < epw.in > epw.out
works!

I am still having problems with the path.dat file, but it is probably just a matter of format.
I'll let you know if I am able to perform the whole calculation

Thank you again

Bets regards

Sara Fiore

ETH Zrich | PhD Student | Doctorate at D-ITET
Rechn. Modellierung Nanostrukturen