The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

Two cases:
- your potential converge: at the end of your 100 SCF cycle you get a small valu for tolvrs
this is the best case. Likely  increasing nline will  solve your problem. 
For instance put "nline 15" in your input file.
- your potential is hardly converged after 100 SCF cycles. You will have to find how to make it converge. At the same time increase also nline if tolwfr do not decrease.

Best

PMA

On Jan 29, 2008 3:11 PM, hyello25 Qgrui <hyello25@gmail.com> wrote:

Dear abinit users,

In the document file of abinit of the parameter 'tolwfr':
"Note that the preparatory GS calculations before a RF calculations must be highly converged.
Typical values for these preparatory runs are tolwfr between 1.0d-16 and 1.0d-22."

While in my calculation of a complex system "Ba2ZnNb2O9", I could not to reach such stringent precision.
After about 100 steps to run a GS calculation,  the output file gives:
-----------------------------------------------------------------------------------------------------------------------
 scprqt:  WARNING -
  nstep=  100 was not enough SCF cycles to converge;
  maximum residual=  2.826E-07 exceeds tolwfr=  1.000E-18

------------------------------------------------------------------------------------------------------------------------

So can anyone give me some suggestions how to treat this situation. Need I to do a more accurate structure optimize?

Thanks in advance.
Bests.


G.R. Qian
Physics Dept.
NJU  CN

-- 
Pierre-Matthieu Anglade