[Pw_forum] dE0s is positive which should never happen

[Guntram Schmidt]

I now ran the jobs from different directories and with different 
tmp-files, but I get the same error again and again (but some jobs seem 
to run fine).
Here is a sample input file:

&control
    calculation = 'scf',
    restart_mode='from_scratch',
    prefix='i2t0480n.xyz.201.804',
    tstress = .true.,
    tprnfor = .true.,
    pseudo_dir = '/gpfs/home/aasfu/espresso-4.3.1/pseudo/',
    outdir='/gpfs/home/aasfu/tmp/i2t0480n.xyz.201.804/',
    verbosity = 'high',
/
&system
  ibrav = 1,
  A = 20.0,
  B = 20.0,
  C = 20.0,
  cosAB = 0,
  cosAC = 0,
  cosBC = 0,
  nat = 49,
  ntyp = 5,
  occupations = 'fixed',
  ecutwfc = 201,
  ecutrho = 804,
  input_dft = 'vdW-DF',
/
&electrons
     conv_thr        = 1.0d-8,
     electron_maxstep= 1000,
/


ATOMIC_SPECIES
  O 1.0 O.rpb-nc.UPF
  P 1.0 P.rpb-nc.UPF
  H 1.0 H.rpb-nc.UPF
  N 1.0 N.rpb-nc.UPF
  C 1.0 C.rpb-nc.UPF


ATOMIC_POSITIONS angstrom
  C    2.53409    1.15758    5.90615
  C    3.15019   -0.07455    5.54807
  C    2.44131    2.12177    4.88712
  N    2.08837    1.31488    7.20825
  H    3.31810   -0.82944    6.31480
  C    3.53738   -0.33017    4.25447
  C    2.81937    1.85588    3.58926
  H    2.05732    3.11106    5.13159
  P    1.27411    2.59232    7.67785
  H    3.99492   -1.28895    4.01502
  C    3.35466    0.61367    3.23968
  H    2.70012    2.62397    2.82653
  C    2.32361    4.03989    7.96545
  C   -0.07967    3.08476    6.58551
  C    0.49920    2.23274    9.26472
  C    3.67968    0.32829    1.84558
  C    3.68044    3.95508    7.70220
  C    1.82454    5.21456    8.53441
  C   -0.99973    2.10073    6.24441
  C   -0.20939    4.36596    6.04202
  C    1.10848    1.38858   10.18043
  C   -0.66577    2.90557    9.63412
  O    3.33385    1.17929    0.95421
  O    4.28137   -0.76191    1.55162
  H    4.09160    3.03654    7.28608
  C    4.52529    5.02716    7.95979
  C    2.67299    6.27587    8.81889
  H    0.76162    5.29841    8.75594
  H   -0.88822    1.09226    6.63997
  C   -2.05731    2.39962    5.40230
  C   -1.25305    4.65424    5.19000
  H    0.51734    5.13775    6.29126
  H    2.01595    0.85160    9.90827
  C    0.56079    1.22802   11.44290
  C   -1.21257    2.72996   10.90225
  H   -1.15056    3.57429    8.92439
  H    3.53933    0.98640    0.00205
  H    5.58609    4.95170    7.72535
  C    4.02439    6.18320    8.51176
  H    2.27935    7.17994    9.28108
  H   -2.78394    1.63041    5.14492
  C   -2.19087    3.68406    4.88570
  H   -1.34098    5.64888    4.75538
  H    1.04561    0.56629   12.15913
  C   -0.59516    1.89743   11.80664
  H   -2.12784    3.25179   11.17774
  H    4.68666    7.02496    8.70859
  H   -3.03465    3.92651    4.24136
  H   -1.01367    1.76534   12.80330

K_POINTS automatic
1 1 1 0 0 0


and the errormessage in the loadleveler-logfile:

ATTENTION: 0031-408  4 tasks allocated by LoadLeveler, continuing...
ATTENTION: 0031-606 Unrecognized environment variable, MP_STACK_OVERFLOW.
    0:"recvec.f90", line 106: 1525-108 Error encountered while 
attempting to allocate a data object.  The program will stop.
    1:"recvec.f90", line 106: 1525-108 Error encountered while 
attempting to allocate a data object.  The program will stop.
    2:"recvec.f90", line 106: 1525-108 Error encountered while 
attempting to allocate a data object.  The program will stop.
    3:"recvec.f90", line 108: 1525-108 Error encountered while 
attempting to allocate a data object.  The program will stop.
ERROR: 0031-300  Forcing all remote tasks to exit due to exit code 1 in 
task 2


Any ideas?
Thank you,
Guntram


PS: last post was badly written - i didn't mean "without some things" 
but "with some things"


Am 08.07.2011 10:08, schrieb giannozz at democritos.it:
> Quoting Lorenzo Paulatto<Lorenzo.Paulatto at impmc.upmc.fr>:
>
>> It could be a small bug in the BFGS algorithm, or a compiler error or a
>> bug in some algebra library...
>
> or another job running in the same directory
>
> P.
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum