[QE-users] Error with SSSP Efficiency (version 1.0) Pseudopotentials

[Karmodak, Naiwrit]

Hello,

I am trying to relax the bulk structure of CoI2 with SSSP Efficiency (version 1.0) Pseudopotentials, but met with the following error:
"S matrix not positive definite". This might be due to the input structures, but the structure is taken from the COD.cif. I am attaching the input file.

However, the same input file runs and convergence is achieved when the pseudopotential is changed to SSSP Efficiency (version 0.7).

I am not able to understand the problem with the version of pseudopotential. Thank you very much.

<input_file>

&control
calculation = 'vc-relax',
prefix = 'CoI2',
tprnfor = .true.,
verbosity = "high"
forc_conv_thr = 0.001,
pseudo_dir = "./"
outdir = './',
wfcdir = './'
/
&system
nspin = 1,
ecutwfc = 55,
nat = 3
occupations = 'smearing',
ibrav = 0,
degauss = 0.02,
smearing = 'marzari-vanderbilt',
tot_charge = -0.00,
ntyp = 2
ecutrho = 440,
input_dft = "vdw-df2-c09"
/
&electrons
electron_maxstep = 200,
conv_thr = 5e-09,
mixing_beta =   3.0000000000d-01
/
&ions
ion_dynamics = 'bfgs',
/
&cell
cell_dynamics = 'bfgs',
/
ATOMIC_SPECIES
Co  58.93 Co_pbe_v1.2.uspp.F.UPF
I   126.90 I.pbe-n-kjpaw_psl.0.2.UPF
CELL_PARAMETERS {angstrom}
  3.96000   0.00000   0.00000
 -1.98000   3.42946   0.00000
  0.00000   0.00000   6.65000
ATOMIC_POSITIONS {crystal}
Co   0.00000   0.00000   0.00000
I   0.33333   0.66667   0.25000
I   0.66667   0.33333   0.75000
K_POINTS {automatic}
10 10 10   0 0 0



---

Postdoctoral Associate

Andreussi Group

Physics Dept.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190424/bd593b70/attachment.html>