Page 1 of 1

partial dos error on parallel computing

Posted: Fri Sep 12, 2014 1:26 pm
by spaz
hi all,
i am trying to calculate the partial DOS of the GaAs in parallel with mpirun. total dos, bandstructure.phonon calculations were succesfully run in parallel with 16 cores but partial dos crahes with the error below

The job will continue.
application called MPI_Abort(MPI_COMM_WORLD, 13) - process 6

===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 57930 RUNNING AT atos1
= EXIT CODE: 13
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
input file is as follows?
ndtset 2
jdtset 1 2
prtwf1 1
prtden1 1
tolvrs1 1.0d-11 # Strict tolerance on (squared) residual of the
#dataset 2 ldos and pdos calculation
#occopt2 5
iscf2 -3
getden2 1
getwfk2 1
prtdos2 3
kptrlatt2 4 -12 4 -12 4 4 -4 -4 -4

ratsph2 3.76 3.76 # This has been ajusted to give approximately 4 valenc electrons per atom
tolwfr2 1.0d-12
natsph2 2
iatsph2 1 2

###kpt-grid and basis set parameters############
ecut 30
ecutsm 5.00000000E-01 Hartree
kptopt 1 # Use symmetry and treat only inequivalent points
#pawecutdg 80
kptrlatt 4 -12 4 -12 4 4 -4 -4 -4
nshiftk 1
shiftk 0.5 0.5 0.5
nstep 80 # Maxiumum number of SCF iterations


##### space group Pm -3 (#200); Bravais cP (primitive cubic)-unit-cell contents ###
#Definition of the unit cell
acell 7.3953702155E+00 7.3953702155E+00 7.3953702155E+00 Bohr
rprim 0.0000000000E+00 7.0710678119E-01 7.0710678119E-01
7.0710678119E-01 0.0000000000E+00 7.0710678119E-01
7.0710678119E-01 7.0710678119E-01 0.0000000000E+00
xred 0.0000000000E+00 0.0000000000E+00 0.0000000000E+00
2.5000000000E-01 2.5000000000E-01 2.5000000000E-01
typat 1 2

angdeg 90 90 90

ntypat 2
znucl 31 33
natom 2
nband 30

can you help me to get ride of this problem thx very much......

Re: partial dos error on parallel computing

Posted: Fri Sep 12, 2014 9:44 pm
by Jordan
As a quick reply, you can not have both getwfk and getden defined for the same dataset.
So remove getden2 and keep only getwfk2 assuming your wfk are good enough

Cheers,

Jordan

Re: partial dos error on parallel computing

Posted: Fri Sep 12, 2014 11:15 pm
by spaz
thx Jordan but it didnt work, still same error appears...:(