[Pw_forum] Errors in PW.x calculation

[John B Burdette]

Dear Users,

     I am performing a pw.x calculation for potassium oxide.  
When I am running the pw.x program for certain values of 
celldm(1), since ibrav=2, I am receiving the following error:

     iteration #  2     ecut=    70.00 ryd     beta=0.70
     Conjugate-gradient style diagonalization

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%
     from cdiaghg : error #        12
     info =/= 0
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%

But only for celldm(1) < 5.00.  For values greater than 5.00, 
I receive a steep curve and not the parabola I expect.  The 
energy curve drops steeply down as celldm(1) approaches < 
5.00.  Here is the input file for pw.x:

K2O
Potassium Oxide
 &control
    calculation = 'relax'
    restart_mode='from_scratch',
    prefix='Potassium Oxide',
    tstress = .true.,
    tprnfor = .true.,
    pseudo_dir = './',
    outdir='./',
 /
 &system
    ibrav=2,
    celldm(1) = 7.000,
    nat=3,
    ntyp= 2,
    ecutwfc =70.0,
 /
 &electrons
    diagonalization='cg'
    mixing_mode = 'plain'
    mixing_beta = 0.7 
    conv_thr =  1.0d-8
 /
 &ions
    ion_dynamics = 'bfgs'
 /
ATOMIC_SPECIES
 K   39.0983  K+.US.Burdette.UPF
 O   15.9994  O.US.LDA.RRKJ3.UPF
ATOMIC_POSITIONS crystal
 O  0.00 0.00 0.00
 K  0.25 0.25 0.25
 K  -.25 -.25 -.25
 K_POINTS automatic
  4 4 4  1 1 1


The pseudopotentials are of neutral Oxygen and K with one 
extra electron.  Each was used in other pw.x experiments and 
worked fine.  This is why I am so confused.  Hopefully, one 
or more of you can help me.

Best regards,

John Burdette