Author Topic: Subsequent Escf runs - How to?  (Read 3806 times)

marcen

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Subsequent Escf runs - How to?
« on: March 14, 2018, 10:56:15 AM »
Hello Everyone,

I'm pretty new to the escf module of Turbomole, so I am trying to get familiar with it.
But I'm not really sure whether my calculation setup is correct and i hope someone can help me out.

Right now, I am calculating excited states to obtain CD and UV/VIS spectra. Afterwards, the calculated spectra are compared to experimentally derived spectra in order to identify the absolute configuration of a molecule.  Since my molecules have about 140k states to calculate, i split up the calculation in portions of 4k states (calculation of 140k states at once seems not bearable).
In the manual it is said: "In subsequent runs add more excitations until a converged result is reached. escf will keep the converged roots, so not much time is lost using this restart approach."

So now my question is:
Does this mean, that i need to change the number of states in the control file in each subsequent run, adding another 4k on top of that previous number, or is Turbomole able to recognize the already calculated states.
So as an example, i have calcualted the first 4k states in a first escf run with the following $soes-flag:
$soes
 a         4000

Can i leave the control file with the given $soes flag untouched for a second escf run, or do i need to change it to
$soes
 a         8000
in order to give Turbomole the information that is has to calculate another 4k states.


Thanks in advance,

Marc