[Pw_forum] pw.x always crashes at the end of relax calculations
Eduardo J. Lamas
ejl7240 at chemail.tamu.edu
Tue Apr 18 15:09:31 CEST 2006
Hi, pw.x always crashes to me during the last check of the relax
calculations
it happened in versions 2.1.3 and 3.0 and with different compilers.
In order to avoid this I commented the following lines in move_ions.f90
(then the program proceed
normally and ends normally but the check without electronic history is not
longer really done):
CALL hinit0()
CALL potinit()
CALL newd()
CALL wfcinit()
When I want to check that the forces are indeed low I take the coordinates
without the verifications
and run a scf calculation and print the forces.
I am wondering if it is something wrong in those subroutines and if somebody
else was having the same problem.
Maybe during the reinitialization a variable is deleted twice or something
or is accessed after it is deleted?.
Thanks,
Eduardo
This run for example crashes (but nearly all runs were having this problem):
&CONTROL
title='H20',
calculation='relax',
restart_mode='from_scratch',
tprnfor=.true.,
pseudo_dir='/home/lamas/espresso/pseudo/',
outdir='./tmp/',
max_seconds=90000,
/
&SYSTEM
ibrav = 1,
celldm(1)= 15,
nat = 3,
ntyp = 2,
nbnd = 5,
ecutwfc =50.0,
ecutrho =500,
occupations='from_input',
nspin=2,
starting_magnetization=0.0
/
&ELECTRONS
mixing_beta = 0.2,
conv_thr = 1.0d-6
/
&IONS
ion_dynamics='bfgs',
/
ATOMIC_SPECIES
O 16.00 O_PBE_v1.vdb
H 1.00 H_PBE_v1.vdb
ATOMIC_POSITIONS {angstrom}
O 2.021843520 1.760434290 0.28003039
H 2.286102819 0.815021260 0.08933101
H 2.811655420 2.359926539 0.15037213
OCCUPATIONS
1.0 1.0 1.0 1.0 0.0
1.0 1.0 1.0 1.0 0.0
More information about the users
mailing list