The ABINIT Users Mailing List ( CLOSED )

[Damien Caliste <damien.caliste@cea.fr>]

Hello,

   We've run several calculations in PAW with lobpcg on parallel
computer using 64 procs with success for neutral systems.

   We now try to add a charge just adding the keyword of the same name
and we put namely "charge +1" (or -1) in our input file without any
other change.

   But on contrary to the neutral case, the calculation goes wrong from
the first iteration, with warnings for the computation of the density :

mkdenpos : WARNING -
  Density went < 0 at       305 points
  and was set to  1.00E-14.  Lowest was  -0.20E+00.

   The cut-off is 10Ht for silicon PAW pseudo, and especially such input
file works well for the neutral case. Does anyone has an idea of which
issue this may come from ? btw, we run on a 5.6 version of Abinit,
which is not the last production version but which works well for
neutral cases.

Regards,

Damien.