[Pw_forum] Error in pw.x with QE 5.4 version

[Paolo Giannozzi]

Dear Pradip, the kind of error you got might be a problem of your
executable, or might be related to very specific cases of parallel
execution. It is impossible to do anything without (and sometimes even
with) an exact description of what happens under which conditions. I
made a quick test and it works for me. Note that your structure is
completely absurd. Also note that you should never use disk_io='high'
unless you have a good reason to.

Paolo

On Wed, Mar 29, 2017 at 12:25 AM, Pradip Shelke <shelke.pradip at gmail.com> wrote:
> Dear Sir/Madam
> We have recently installed QE-5.4 in the directory.
> we run a job for pw.x using following command:
> ==============================================================
> mpirun -np $NPROCS ${ESPRESSO_ROOT}/bin/pw.x  <zno.relax.in &> zno_relax.txt
> ==============================================================
>
> But it shows following error
>
> =============================================================================================
> Fatal error in MPI_Group_incl: Invalid rank, error stack:
> MPI_Group_incl(156)..............: MPI_Group_incl(group=0x88000002, n=64,
> ranks=0x2603400, new_group=0x7fffda873778) failed
> MPIR_Group_check_valid_ranks(257): Duplicate ranks in rank array at index 8,
> has value 0 which is also the value at index 0
> =============================================================================================
>
>
> =================
> The input file is
> ===============
> &CONTROL
> !calculation = "scf",
> calculation = 'vc-relax',
> restart_mode='from_scratch',
> !restart_mode='restart',
> prefix = "ZNO",
> pseudo_dir = "/home/external/aca/pradips/qmespresso/potentials/",
> outdir = "TMP_DIR",
> !nstep=1000,
> forc_conv_thr = 1.0D-3,
> disk_io='high',
> /
> &SYSTEM
> ibrav=0,
> !celldm(1) =6.330582528,
> nat=18, ntyp= 3, tot_charge=0.0,
> occupations='smearing', smearing='gauss', degauss=0.001,
> ecutwfc =140.0, !better 140 if too small convergence is not acheived and if
> too large have memory issue
> ecutrho =560.0, !better to keep 10 times ecutwfc
> !nr1=60,nr2=60,nr3=1,
> !vdw_corr='Grimme-D2'
> vdw_corr='ts-vdW'
> nspin=2,!1 ,2 or 4 non, spin along z,spin magnetizetion in generic direction
> starting_magnetization(1)= 0.003,
> /
> &ELECTRONS
> mixing_mode='plain',
> mixing_beta = 0.5,
> startingwfc='random',
> conv_thr = 1.0d-10,
> /
> &IONS
> ion_dynamics='bfgs',
> !ion_positions='from_input',
> ! trust_radius_ini=1.0d-1,
> !   trust_radius_max=1.0d-3,
> /
> &CELL
> cell_dynamics='bfgs',
> cell_factor=10,
> /
> ATOMIC_SPECIES
> Zn 65.409 Zn.pbe-d-hgh.UPF
> O 15.999 O.pbe-hgh.UPF
> Al 26.9815386 Al.pbe-hgh.UPF
> K_POINTS automatic
> 6 6 1 0 0 0
> CELL_PARAMETERS (angstrom)
>    3.842263623  -6.632836491  -0.000000000
>    3.841193017   6.632258570   0.000000000
>    0.000000000  -0.000000000   20.000000000
>
> ATOMIC_POSITIONS (angstrom)
> O       1.64316210       -0.93780403        2.19491689
> O       3.39586199        1.81576813        2.15414423
> O       4.90310690        4.71030578        2.10304279
> O       3.26801979       -3.77491426        2.24678444
> O       4.90312198       -0.79524830        2.20324294
> O       6.41037520        1.81577209        2.15412539
> O       4.90312792       -6.60906681        2.29269709
> O       6.53823127       -3.77492396        2.24677362
> O       8.16307066       -0.93781327        2.19488231
> Zn      1.63411410        0.94472754        2.16548083
> Zn      3.26801687        3.77652009        2.11943151
> Zn      4.90310105        6.60876713        2.06792438
> Zn      3.26892853       -1.88732667        2.21640255
> Zn      6.53818997        3.77651121        2.11941512
> Zn      4.90312260       -4.72045538        2.26207899
> Zn      6.53729795       -1.88732511        2.21638362
> Zn      8.17211183        0.94471853        2.16544718
> Al      4.90310193        0.94515020        2.17094412
> ===========================================
>
>
>
>
> Kindly help me in this regard,
>
> --
> Dr. Pradip B. Shelke
> Associate Professor,
> Department of Physics,
> Ahmednagar College, Ahmednagar, India.
> Contact: 9422226924



-- 
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222