[QE-users] Problem with PBE0 calculation

Giuseppe Mattioli giuseppe.mattioli at ism.cnr.it
Thu Jan 12 09:03:20 CET 2023


AFAIK, nscf calculations are not compatible with EXX functionals.  
There are a few workarounds, like manually adding the band path as  
"fake" k-points with weight=0 to the regular grid obtained by the  
automatic distribution used in the scf calculation.
I generally perform EXX calculations of isolated or very large systems  
@Gamma. More experienced users may add more accurate information.
HTH
Giuseppe

Quoting Rameswar Bhattacharjee <rb1820 at georgetown.edu>:

> Hi Giuseppe,
> Thank you very much for your detailed reply. I followed the instructions
> and was able to complete the SCF calculation. Also, I did an NSCF
> calculation using the same combination you suggested as a pre-step to
> perform the band calculation. My ultimate aim is to get the band gap using
> a hybrid functional such as PBE0.
>
> However,  I am not able to run the band calculation with this combination
> got an error like below:
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
>      Error in routine exx_grid_init (1):
>
>      wrong EXX q grid
>
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
> It seems the k-points grid I am using is not suitable for EXX calculation
> although I used the same grid for PBE and it was perfect. I am sure someone
> in the forum had performed band structure using hybrid functional and would
> be able to help me resolve this. Any help would be highly appreciated.
>
> Here is my input for band calculation using PBE0
>
> ------------------------------------------
>
> &control
>
>     restart_mode='from_scratch'
>
>     calculation='bands'
>
>     prefix='Ethyne-Anthracene',
>
>     pseudo_dir = '/home/rb1820/QE-PP/NC',
>
>     outdir='./scratch'
>
> /
>
> &system
>
>   ibrav = 0,
>
>    tot_charge=0,
>
>   nat=36,
>
>   ntyp=2,
>
>   ecutwfc = 90,
>
>   ecutfock= 150,
>
>   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_ONCV_PBE-1.2.upf
>
> C   12.0107   C_ONCV_PBE-1.2.upf
>
>
> ATOMIC_POSITIONS {angstrom}
>
> K_POINTS { crystal }
>
> 11
>
> -0.5  0.0  0.0   1.0
>
> -0.4  0.0  0.0   1.0
>
> -0.3  0.0  0.0   1.0
>
> -0.2  0.0  0.0   1.0
>
> -0.1  0.0  0.0   1.0
>
>  0.0  0.0  0.0   1.0
>
>  0.1  0.0  0.0   1.0
>
>  0.2  0.0  0.0   1.0
>
>  0.3  0.0  0.0   1.0
>
>  0.4  0.0  0.0   1.0
>
>  0.5  0.0  0.0   1.0
>
>
> CELL_PARAMETERS {angstrom}
>
>    6.933462797   0.000000000   0.000000000
>
>    0.000000000  20.000000000   0.000000000
>
>    0.000000000   0.000000000  20.000000000
> ---------------------------------------------------------------
>
>
> Thanks
>
> Rameswar Bhattacharjee
> Georgetown University
>
> On Wed, Jan 11, 2023 at 3:21 AM Giuseppe Mattioli <
> giuseppe.mattioli at ism.cnr.it> wrote:
>
>>
>> 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>
>>
>> _______________________________________________
>> The Quantum ESPRESSO community stands by the Ukrainian
>> people and expresses its concerns about the devastating
>> effects that the Russian military offensive has on their
>> country and on the free and peaceful scientific, cultural,
>> and economic cooperation amongst peoples
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>



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