Segmentation fault (core dumped)

[muhammadhasan]

Hi,

I am getting an error message (sdc.txt, attached) while running 'yambo', (p2y was successful). Basically the same input file running well if I reduce k points. I have run using 3375, 4913, 5832 k-points successfully.. However, when I start running using k-points 6859 or beyond, the segmentation fault problem happens. Here is my nscf file, which run successfully even at higher (6859 and beyond) k-points.

Code: Select all

 &control
  calculation = 'nscf'
  restart_mode = 'from_scratch'
  prefix = 'ZG-au_300'
  pseudo_dir = './'
  outdir = './out/'
  verbosity = 'high'
/
&system
  ibrav = 0
  nat =   27
  ntyp =    1
  ecutwfc =  50.00
  occupations = 'smearing'
  smearing = 'mp'
  degauss =   0.19D-02
  force_symmorphic = .true.
  nbnd = 280
  nosym = .TRUE.
  noinv = .TRUE.
/
&electrons
  diago_thr_init=5.0e-6,
  diago_full_acc = .true.
  diagonalization = 'david'
  mixing_mode= 'plain'
  mixing_beta = 0.30
  conv_thr =   1D-5
/
ATOMIC_SPECIES
   Au  196.967 Au_ONCV_PBEsol-1.0.upf
CELL_PARAMETERS (angstrom)
     -6.11369997      0.00000000      6.11369997
      0.00000000      6.11369997      6.11369997
     -6.11369997      6.11369997      0.00000000
ATOMIC_POSITIONS (angstrom)
K_POINTS crystal
6859
.....

Would you have any suggestion how to resolve the issue? I have run it parallel and serial (LOG.txt) in both ways, however same error is noticing.
Thank you

Best
Md J Hasan
PhD student
Mechanical Engineering
University of Maine

[Daniele Varsano]

Dear Hasan,

it can also be a memory issue anyway, from the log file, it seems there is a problem with your k grid that is seen as not-connected.
Can you please post your yambo input and report files (r-*)?

Best,

Daniele

[muhammadhasan]

Hi Professor,

I have attached all necessary input files (PW, yambo). However, when I run 'yambo" after "p2y", It has not generated any report files (r-*). The 'scf' and 'nscf' have run successfully in Quantum Espresso.

Best
Md J Hasan
PhD student
Mechanical Engineering
University of Maine

[Daniele Varsano]

Dear Hasan,

the problem is that the setup cannot terminate correctly and generate the k/q points database. The reason is indicated in the log file:

Code: Select all

[RL indx] X & Total k-grids are not connected

Is there a particular reason to provide the k points by hands in the nscf run? Consider that Yambo needs a regular grid.

I suggest you to input an automatic MK grid in nscf and repeat the p2y procedure.

Best,
Daniele

[muhammadhasan]

Hi Professor,

Actually, there is no particular reason to provide the k points by hands in the nscf run. According to your suggestion, I considered regular grid (MK) and it works. Thank you so much.

Best
Md J Hasan
PhD student
Mechanical Engineering
University of Maine