<div dir="ltr"><div><br></div>Yes, as Mtthias observed, this is an error message due to the ELPA diagonalizer.<div>it can happen with ELPA, if you have too many processors for a too small matrix.<div>you can also try to reduce the number of nodes</div><div><br></div><div>ciao</div><div>Marcella</div><div><br><br>On Thursday, June 26, 2014 11:04:56 AM UTC+2, c.pignedoli wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;">Dear all, I am having a problem on the cray machine at CSCS,
<br>when printing MO_CUBES, after filled orbitals are printed the program
<br>exit with the following error:
<br>
<br>ERROR: Problem contains processor column with zero width
<br>
<br>IS anybody familiar with this problem?
<br>
<br>here is a cut from the input file (enclosed in the tar)
<br>
<br>Kind regards
<br>
<br>Carlo
<br>
<br> UKS
<br> MULTIPLICITY 1
<br> &PRINT
<br> &MO
<br> FILENAME EIG
<br> ADD_LAST NUMERIC
<br> &EACH
<br> QS_SCF 100
<br> &END
<br> EIGENVALUES
<br> OCCUPATION_NUMBERS
<br> &END
<br> &MO_CUBES
<br> NHOMO 200
<br> NLUMO 200
<br> STRIDE 2 2 2
<br> WRITE_CUBE TRUE
<br> &END
<br> &V_HARTREE_CUBE
<br> FILENAME ./V_HARTREE
<br> STRIDE 2 2 2
<br> &END
<br> &E_DENSITY_CUBE
<br> FILENAME RHO
<br> STRIDE 2 2 2
<br> &END
<br> &END
<br> &QS
<br> METHOD GPW
<br> EXTRAPOLATION ASPC
<br> EXTRAPOLATION_ORDER 3
<br> EPS_DEFAULT 1.0E-14
<br> MAP_CONSISTENT
<br> &END QS
<br> &MGRID
<br> CUTOFF 600
<br> NGRIDS 5
<br> &END
<br> &SCF
<br> MAX_SCF 500
<br> SCF_GUESS ATOMIC
<br> EPS_SCF 1.0E-7
<br> ADDED_MOS 500
<br> &SMEAR ON
<br> METHOD FERMI_DIRAC
<br> ELECTRONIC_TEMPERATURE [K] 200
<br> &END SMEAR
<br> &DIAGONALIZATION ON
<br> ALGORITHM STANDARD
<br> &END DIAGONALIZATION
<br> &MIXING ON
<br> METHOD BROYDEN_MIXING
<br> ALPHA 0.020
<br> BETA 0.5
<br> NBROYDEN 8
<br> &END MIXING
<br> &END SCF
<br></blockquote></div></div></div>