[QE-users] bfgs failed after 25 scf cycles and 23 bfgs steps, convergence not achieved

[Barsha Pal]

The error I have got is:

Message from routine bfgs:
     history already reset at previous step: exiting
     Energy error            =      3.4E-09 Ry
     Gradient error          =      4.0E-05 Ry/Bohr
     Cell gradient error     =      2.8E-03 kbar

     bfgs failed after  25 scf cycles and  23 bfgs steps, convergence not
achieved
     (criteria: energy <  1.0E-06 Ry, force <  1.0E-05 Ry/Bohr, cell <
 5.0E-03 kbar)

     End of BFGS Geometry Optimization

     Final enthalpy           =    - **  Ry

     File ./vc/g.bfgs deleted, as requested
Begin final coordinates
     new unit-cell volume =     **  a.u.^3 (   **Ang^3 )
     density =    **   g/cm^3

CELL_PARAMETERS (alat= )


ATOMIC_POSITIONS (crystal)

End final coordinates
---------

I get this message for a spin orbit coupling calculations, I had used all
relativistic pseudos. Input file description is as follows:


&CONTROL
    calculation   = "vc-relax"
    etot_conv_thr =  1.00e-06
    forc_conv_thr =  1.00e-05
    nstep         =  100
    outdir        = "./vc"
    prefix        = "g"
    pseudo_dir    = "./"
    verbosity     = "high"
/

&SYSTEM
    celldm(1)   =
    degauss     =  0.000735
    ecutrho     =  440
    ecutwfc     =  110
    ibrav       =  0
    nat         =
    ntyp        =  2
    occupations = "smearing"
    smearing    = "gaussian"
    starting_magnetization(1) =  0.00000
    starting_magnetization(2) =  1.00000
    noncolin = .true.
    lspinorb = .true.
/

&ELECTRONS
    conv_thr         =  1.00000e-08
    electron_maxstep =  200
    mixing_beta      =  7.00000e-01
/

&IONS
    ion_dynamics = "bfgs"
/

&CELL
    cell_dofree    = "2Dxy"
    cell_dynamics  = "bfgs"
    press          =  0.00000e+00
    press_conv_thr =  5.00000e-03
/

K_POINTS {automatic}
 8  8  1  0 0 0

CELL_PARAMETERS {alat}


ATOMIC_SPECIES
A      *******      A.upf   !using norm-conserving and full relativistic
with pbe
B       *******     B.upf   !using norm-conserving and full relativistic
with pbe

ATOMIC_POSITIONS {crystal}


What should I do next for optimization to be converged?
Also, please provide some guidelines for applying the magnetization value.
How much should it be used?
If I have two types of atoms and put zero for starting_magnetization for
both, how is it taken? If magnetization is given for one type of atom then
how is it calculated? And if startine_magnetization is given for both, then
what does the code consider?


Thank you.
Ashley Cooper
Phd
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20250426/eaaed2d7/attachment.html>