[QE-users] problems computing cholesky

Giuseppe Mattioli giuseppe.mattioli at ism.cnr.it
Wed Sep 12 10:42:34 CEST 2018


Dear Punit

>       problems computing cholesky

The first source of such error is generally in the structure itself.  
If you try to reach the ground state of an unphysical structure the  
diagonalization may be affected by strong instabilities. I'm not 100%  
sure that this is your case: the structure is strange, but after all  
I've no experience of La2O3. However, at first sight (I opened your  
input file with xcrysden) there are two inequivalent O sites in your  
structure. A fourfold-coordinated O which forms three short (2.32 A)  
La-O bonds and one longer (2.54 A). This is apparently reasonable.  
There is a different sixfold-(barely)coordinated O site, almost an  
interstitial O atom in an octahedral La cage, having longer La-O  
distances (2.75 A). O in oxides has generally no chemical tendency to  
stay in such an interstitial position, and this is likely the source  
of your scf instability. Hence I recommend that you double-check the  
structure. As a minor point, the vacuum space between periodically  
replicated images along the z axis is a bit on the side of "scarce".  
You should at least double the space occupied by your system  
(including the main part of the tail of charge density). You might  
also want to use a special solver to decouple the electrostatic  
potential along the z axis (see for example assume_isolated='esm' in  
the vacuum-slab-vacuum geometry).
HTH
Giuseppe

Quoting Punit Kumar <ip.punit.2016 at gmail.com>:

> Hello QE users
> Recently I have started working on QE. Currently I am trying to relax La2O3
> slab which is 4 layers deep. But while doing the calculation I have
> encountered an error which is "Problem computing cholesky". I have searched
> this issue on QE mailing list and found that to remove this problem one has
> to use norm-conserving PP. I am also using norm-conserving PP but error
> still remains the same. Since I am new user I have no clue how to tackle
> this problem. Any help to resolve this error is highly appreciated. Below I
> have given the crash report and my input file
>
> *CRASH*
>      task #         0
>      from  cdiaghg  : error #       151
>       problems computing cholesky
>
> *Input file*
>
> &CONTROL
>                        title = 'lao' ,
>                       prefix = 'lao'
>                  calculation = 'relax' ,
>                 restart_mode = 'from_scratch'
>                       outdir = './out' ,
>                   pseudo_dir =
> '/home/punit/Documents/punit/sg15_oncv_upf_2015-10-07/' ,
> /
> &SYSTEM
>                        ibrav = 0,
>                     celldm(1)= 1.89 ,
>                          nat = 20,
>                         ntyp = 2,
>                      ecutwfc = 50 ,
>                      ecutrho = 500 ,
>                  occupations = 'smearing' ,
>                     smearing = 'gaussian' ,
>                      degauss =  0.020
> /
> &ELECTRONS
>             electron_maxstep = 500 ,
>                     conv_thr = 1.0d-5 ,
>                  mixing_mode = 'plain' ,
>                  mixing_beta = 0.7 ,
>              diagonalization = 'david' ,
> /
> &IONS
> /
> &CELL
> /
> CELL_PARAMETERS alat
>         3.843000000   0.000000000   0.000000000
>        -1.921500000   3.328100000   0.000000000
>         0.000000000  -0.000000000  48.262800000
>
> ATOMIC_SPECIES
>     La  138.90   La_ONCV_PBE-1.0.upf
>     O   15.999   O_ONCV_PBE-1.0.upf
> ATOMIC_POSITIONS alat
> La      0.00000000       2.21880000      13.62680000    1  1  1
> La      1.92150000       1.10940000      16.84560000    1  1  1
> O       0.00000000       0.00000000      12.00000000    1  1  1
> O       0.00000000       2.21880000      16.16450000    1  1  1
> O       1.92150000       1.10940000      14.30790000    1  1  1
> La      0.00000000       2.21880000      20.09920000    1  1  1
> La      1.92150000       1.10940000      23.31800000    1  1  1
> O       0.00000000       0.00000000      18.47240000    1  1  1
> O       0.00000000       2.21880000      22.63690000    1  1  1
> O       1.92150000       1.10940000      20.78030000    1  1  1
> La      0.00000000       2.21880000      26.57160000    1  1  1
> La      1.92150000       1.10940000      29.79040000    1  1  1
> O       0.00000000       0.00000000      24.94480000    1  1  1
> O       0.00000000       2.21880000      29.10930000    1  1  1
> O       1.92150000       1.10940000      27.25270000    1  1  1
> La      0.00000000       2.21880000      33.04400000    0  0  0
> La      1.92150000       1.10940000      36.26280000    0  0  0
> O       0.00000000       0.00000000      31.41720000    0  0  0
> O       0.00000000       2.21880000      35.58170000    0  0  0
> O       1.92150000       1.10940000      33.72510000    0  0  0
> K_POINTS automatic
> 4 4 1  0 0 0
>
> Regards
> Punit (PhD)
> IIT Bombay



GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>



More information about the users mailing list