[Pw_forum] phcg.x segmentation fault

[Guntram Schmidt]

Dear All,

I'm trying to adopt the working SiH4-example for phonon calculations 
(http://www.fisica.uniud.it/~giannozz/QE-Tutorial/tutorial_advph.html) 
to my molecular crystal.

phcg.x crashes with

ERROR: 0031-250  task 0: Segmentation fault

using differing compiled versions of QE (with and without 
essl/lapack-combinations)...

The input file is:
-----------------------------------
Normal modes for SiH4
  &inputph
   tr2_ph=1.0d-14,
   prefix='i2t0480.xyz.42.168',
   outdir='/gpfs/home/aasfu/tmp/i2t0480.xyz.42.168/',
  amass(1)=30.973762,
  amass(2)=15.9994,
  amass(3)=14.00679,
  amass(4)=12.011,
  amass(5)=1.00794,
   epsil=.true.,
   trans=.true.,
   asr=.true.
   fildyn='i2t0480.xyz.scf.42.rpb-nc.UPF.dyn'
  /
  0.0 0.0 0.0
-------------------------------------
having done a successful scf-calculation on the referenced system (what 
are the files needed as input by phcg.x?).

The output file finished abruptly after:
-------------------------------------
       Stick Mesh
       ----------
       nst =  3369,  nstw =   846, nsts =  3369
                   n.st   n.stw   n.sts    n.g    n.gw   n.gs
       min        6737    1691    6737  495721   61939  495721
       max        6737    1691    6737  495721   61939  495721
           6737    1691    6737  495721   61939  495721


         Check: negative/imaginary core charge=   -0.000007    0.000000


 
---------------------------------------------------------------------------------
    Carrying out vdW-DF run using the following parameters:

    Nqs =      20    Nr_points =    1024   r_max =   100.000
    q_mesh =       0.00001000
         0.04494208
         0.09755937
         0.15916263
         0.23128650
         0.31572767
         0.41458969
         0.53033537
         0.66584808
         0.82450364
         1.01025438
         1.22772762
         1.48234092
         1.78043706
         2.12944203
         2.53805004
         3.01644009
         3.57652955
         4.23227104
         5.00000000

    Gradients computed in Reciprocal space

 
---------------------------------------------------------------------------------


Any ideas what's wrong?
I gave the calculation 12GB of memory for 196 atoms - this should be 
enough, shoudn't it be?

Thanks,
Guntram