[QE-users] Problem with PBE0 calculation

[Giuseppe Mattioli]

Dear Rameswar Bhattacharjee
I don't know if this is the actual source of error, but I strongly  
recommend to use norm-conserving pseudopotentials when performing EXX  
calculations. You have generally no speed-up with US because ecutrho  
is more or less often the same, and in your case you are using  
tremendously high (and likely useless) values for ecutwfc and ecutrho.  
Moreover, using NC you can finely tune the density basis set used to  
build Fock integrals with ecutfock, otherwise stuck to the ecutrho  
value. This combination of values is probably responsible for a very  
high memory request (the code prints an estimate at the beginning of  
the output). For example, *using C and H ONCV pseudopotentials*, you  
should be on the reasonably accurate side using

ecutwfc=90.0
ecutfock=135.0~180.0
ecutrho is automatically fixed at ecutwfc*4=360.0 in the case of NC

HTH
Giuseppe

Quoting Rameswar Bhattacharjee <rb1820 at georgetown.edu>:

> Hi Everyone,
> I am trying to get a band structure using a hybrid functional. As suggested
> previously in this forum I choose PBE0 functional with pbe-ultrasoft
> pseudopotential. The calculation starts fine and scf iteration converges
> smoothly. But the job is stuck and does not complete with the last line "ACE
> projected onto    60 (nbndproj) and applied to    60 (nbnd) bands". I am
> not sure what mistake I am making here or if it is just a memory issue as
> no error message is printing. Any suggestion would be highly appreciated. I
> am providing last few lines of the output and my input str for reference.
> Thank you.
>
> Last few lines of the scf output
> -------------------------------
>
>     highest occupied, lowest unoccupied level (ev):    -3.5627   -3.2294
>
>
> !    total energy              =    -312.15802972 Ry
>
>      estimated scf accuracy    <          3.2E-09 Ry
>
>
>      convergence has been achieved in  12 iterations
>
>
>      Using ACE for calculation of exact exchange
>
>
>      EXX grid:  4188907 G-vectors     FFT dimensions: ( 100, 288, 288)
>
>      ACE projected onto    60 (nbndproj) and applied to    60 (nbnd) bands
> ----------------------------------
>
> Input str
> ------------------------------
>
> &control
>
>     restart_mode='from_scratch'
>
>     calculation='scf'
>
>     prefix='Test',
>
>     pseudo_dir = '/home/QE-PP',
>
>     outdir='./scratch'
>
> /
>
> &system
>
>   ibrav = 0,
>
>    tot_charge=0,
>
>   nat=36,
>
>   ntyp=2,
>
>   ecutwfc = 140,
>
>   ecutrho = 1400.0,
>
>   input_dft='pbe0'
>
>   vdw_corr='grimme-d2',
>
>   nbnd = 60
>
> /
>
> &electrons
>
> electron_maxstep = 1000
>
> mixing_beta=0.7
>
> diagonalization='david'
>
> diago_david_ndim = 2
>
> conv_thr=1.0d-8
>
> /
>
> &ions
>
> /
>
> &CELL
>
> cell_dynamics='bfgs',
>
>  press = 0.0
>
> cell_dofree = 'x',
>
> /
>
> ATOMIC_SPECIES
>
> H   1.008     H.pbe-rrkjus_psl.1.0.0.UPF
>
> C   12.0107   C.pbe-n-rrkjus_psl.1.0.0.UPF
>
>
> ATOMIC_POSITIONS {angstrom}
>
>
> K_POINTS automatic
>
> 12 1 1 0 0 0
>
>
> CELL_PARAMETERS {angstrom}
>
>    6.933462797   0.000000000   0.000000000
>
>    0.000000000  20.000000000   0.000000000
>
>    0.000000000   0.000000000  20.000000000
>
> ----------------
>
>
> Rameswar Bhattacharjee
>
> Georgetown University



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>