[Q-e-developers] any example for Hybrid functionals with PAW

[Lorenzo Paulatto]

On Thursday, 4 May 2017 18:35:56 CEST Ching-Liang Huang wrote:
> Hi all,
> Since the QE6.1 provde Hybrid functionals with PAW as release notes, I would
> like to do some test cases to  practice. Dose anyone test it? Could you
> share an example?
> thanks
> James

There is no special option, just take any input with a PAW pseudopotential and 
exact exchange functional. If you really insist, here is the test input I used 
during developement.

You can get the pseudopotentials here: <https://drive.google.com/open?
id=0Bzgwv8VZEl9SWkJTZk9ZNkhWdDg> (If google drive is blocked in China, tell me 
and I'll send you the file by email).

hth

! === beginning of input
! HF + nspin = 2 + no div threatement 
!   ==>  E_fock = E_H
! HF + nspin = 1 + no div threatment
!   ==> E_fock = 1/2 E_H
! NOTE: the PAW contribution to E_H goes in the paw_contribution term
!
 &control
    calculation = 'scf'
    restart_mode='from_scratch',
    prefix='hydrogen',
    pseudo_dir = './UPF'
    outdir='/home/paulatto/tmp/'
    verbosity = 'high'
 /
 &system    
    ibrav=  0, celldm(1) =30, nat=  1, ntyp= 1,
    ecutwfc = 48
    !ecutfock = 96 !192 
    ! test with and without ecutrho, especially in combination 
    ! with/without tqr
    ecutrho = 384
    !
    ! --- BANDS TEST --- 
    ! result should not change with more bands
    nbnd = 4,
    ! --- SPIN TEST ---
    ! result does change with spin, see top of the file
    nspin = 2
    !
    ! --- FUNCTIONAL TEST ---
    ! For testing, we use 100% Hartree-Fock
    occupations='from_input'
    input_dft='hf+noc+nogx+nogc'
    exx_fraction=1.0
    !
    ! All divergence treatments mus work,
    ! they will converge for HUGE cells to same as above:
    x_gamma_extrapolation = .false.
    exxdiv_treatment='none'
    ! exxdiv_treatment='gygi-baldereschi'
    ! exxdiv_treatment='vcut_spherical'
    !
    ! -- SYMMETRY TEST --- 
    !nosym=.true.
    !noinv=.true.
    !
    ! --- QGRID TEST ---
    ! not very useful in this case, check for troubles
    ! nqx1 = 4, nqx2 = 4, nqx3 = 4, 
    ! nqx1 = 2, nqx1 = 2, nqx3 = 2, 
    ! nqx1 = 1, nqx1 = 1, nqx3 = 1, 
      nqx1 = 0, nqx2 = 0, nqx3 = 0, 

 /
 &electrons
    adaptive_thr=.false.
    conv_thr = 1e-10
    mixing_beta = 0.7 
   !  --- TQR TEST ---
    ! Set real space augmentation, it is much faster, 
    ! check that result is the same in both cases
    tqr=.true.
 /
! All of the following choices of pseudopotential must work:
ATOMIC_SPECIES
 H  1.  H.pbe-pawh2.UPF ! PAW, 4 projector (2x s + 2x p)
 H  1.  H.pbe-ush.UPF   ! US, 2 projectors
 H  1.  H.pbe-nc.UPF    ! NC
 H  1.  H.pbe-us.UPF    ! US, single projector
 H  1.  H.pbe-pawh.UPF  ! PAW, 2 projectors (s channel)
 H  1.  H.pbe-paw.UPF   ! PAW, single projector
 H  1.  H.pbe-pawnc.UPF ! PAW on top of NC
CELL_PARAMETERS
1. 0. 0.
0. 1. 0.
0. 0. 1.
! --- POSITION TEST ---
! alternative positions: result must be the same
ATOMIC_POSITIONS alat
  H  0.33333333333333 0.3333333333333 0.3333333333333
  H  0.5 0.5 0.5
  H  0.25 0.25 0.25
  H  0.0 0.0 0.0

! --- OCCUPATIONS TEST ---
! when using nspin=1 only the first line is read
OCCUPATIONS
  1.0  0.0 0.0 0.0
  0.0  0.0 0.0 0.0

! --- KPOINTS TEST ---
! three+ ways to do the same thing, we expect the same result
! 1. one-point grid, finer grids should also not make a difference (for large 
enough cells)
!K_POINTS automatic
  1 1 1  0 0 0
  2 2 2  0 0 0
  4 4 4  0 0 0
  4 4 4  1 1 1
! 2. manually, as a single point or several copies of Gamma,
!    with weight zero or not zero
!K_POINTS tpiba
 1
  0  0  0   1
  0.25 0.25 0.25 1
  1  1  1   1
  0  0  0   1
! 3. gamma-only
K_POINTS gamma



-- 
Dr. Lorenzo Paulatto 
IdR @ IMPMC -- CNRS & Université Paris 6
phone: +33 (0)1 442 79822 / skype: paulatz
www:   http://www-int.impmc.upmc.fr/~paulatto/
mail:  23-24/423 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05