[QE-users] Convergence of dexx using hybrid functionals

[Prasenjit Ghosh]

Dear all,

I am using qe-6.3 to perform hybrid calculations using gau-PBE.

My system is magnetic and metallic.
The problem that I am facing is that the "dexx" is not converging. The
smallest value it goes to is 0.00011633 Ry. The variation of "dexx" is
given below:

     est. exchange err (dexx)  =       0.00208322 Ry
     est. exchange err (dexx)  =       0.00056558 Ry
     est. exchange err (dexx)  =       0.00025778 Ry
     est. exchange err (dexx)  =       0.00015787 Ry
     est. exchange err (dexx)  =       0.00012726 Ry
     est. exchange err (dexx)  =       0.00012170 Ry
     est. exchange err (dexx)  =       0.00011830 Ry
     est. exchange err (dexx)  =       0.00011673 Ry
     est. exchange err (dexx)  =       0.00011645 Ry
     est. exchange err (dexx)  =       0.00011667 Ry
     est. exchange err (dexx)  =       0.00011639 Ry
     est. exchange err (dexx)  =       0.00011646 Ry
     est. exchange err (dexx)  =       0.00011750 Ry
     est. exchange err (dexx)  =       0.00011779 Ry
     est. exchange err (dexx)  =       0.00011697 Ry
     est. exchange err (dexx)  =       0.00011673 Ry
     est. exchange err (dexx)  =       0.00011683 Ry
     est. exchange err (dexx)  =       0.00011692 Ry
     est. exchange err (dexx)  =       0.00011699 Ry
     est. exchange err (dexx)  =       0.00011544 Ry
     est. exchange err (dexx)  =       0.00011540 Ry
     est. exchange err (dexx)  =       0.00011640 Ry
     est. exchange err (dexx)  =       0.00011707 Ry
     est. exchange err (dexx)  =       0.00011741 Ry
     est. exchange err (dexx)  =       0.00011642 Ry
     est. exchange err (dexx)  =       0.00011633 Ry
     est. exchange err (dexx)  =       0.00011662 Ry
     est. exchange err (dexx)  =       0.00011663 Ry
     est. exchange err (dexx)  =       0.00011808 Ry
     est. exchange err (dexx)  =       0.00011663 Ry
     est. exchange err (dexx)  =       0.00011557 Ry
     est. exchange err (dexx)  =       0.00011635 Ry
     est. exchange err (dexx)  =       0.00011650 Ry
     est. exchange err (dexx)  =       0.00011681 Ry
     est. exchange err (dexx)  =       0.00011631 Ry
     est. exchange err (dexx)  =       0.00011668 Ry
     est. exchange err (dexx)  =       0.00011657 Ry

Can you please let me know what parameters (apart from those related to
smearing) in the input file can be used to control the convergence of the
fock integral ?

The following is my input file:


   &control
    calculation = 'relax'
    prefix='titanium_${a}'
    pseudo_dir = './'
    outdir='./tmp'
/
 &system
    ibrav=0, nat=  3, ntyp= 3, nspin= 2,
    ecutwfc =80, occupations= 'smearing',
    starting_magnetization(1)= 1.0d0,
    starting_magnetization(2)=-1.0d0,
    smearing= 'mv', degauss= 0.007
    exxdiv_treatment='none'
    input_dft='gaupbe'
    nqx1=5,nqx2=5,nqx3=1
    x_gamma_extrapolation=.false.
    ecutfock=90.0d0

 /
 &electrons
    electron_maxstep = 100
    diagonalization='cg'
    mixing_mode = 'local-TF'
    mixing_beta = 0.2
    conv_thr =  1.0d-8
    adaptive_thr=.true.
 /
&ions
ion_dynamics='damp'
/
ATOMIC_SPECIES
 Ti1 47.867 Ti.upf
 Ti2 47.867 Ti.upf
 C  12.01  C.upf
CELL_PARAMETERS bohr
   ${a}   0.000000000   0.000000000
  -${bx}  ${by}  0.000000000
   0.000000000   0.000000000  34.117402123
ATOMIC_POSITIONS (crystal)
Ti1      0.333333333   0.666666667   0.568980897
Ti2      0.333333333   0.666666667   0.431019103
C       -0.000000000  -0.000000000   0.500000000
K_POINTS {automatic}
15 15 1 0 0 0


With regards,
Prasenjit

-- 
PRASENJIT GHOSH,
IISER Pune,
Dr. Homi Bhabha Road, Pashan
Pune, Maharashtra 411008, India

Phone: +91 (20) 2590 8203
Fax: +91 (20) 2589 9790
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190404/e1f77655/attachment.html>