not enough SCF cycles to converge

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
sdwang
Posts: 5
Joined: Thu Apr 22, 2010 5:05 am

not enough SCF cycles to converge

Post by sdwang » Fri Sep 24, 2010 8:47 am

Dear all:
I have did the test t11 in tutorial/Input, it appears:
scprqt: WARNING -
< nstep= 10 was not enough SCF cycles to converge;

But in the Refs, nstep=6 is converged.
Where is my problem?
My version and system:
.Version 6.2.3 of ABINIT
< .(MPI version, prepared for a x86_64_linux_intel9.0 computer)
Thanks!
sdwang
Southeast University in Nanjing, China

mverstra
Posts: 655
Joined: Wed Aug 19, 2009 12:01 pm

Re: not enough SCF cycles to converge

Post by mverstra » Mon Oct 11, 2010 12:16 pm

Usually the test results are very portable. You have probably changed something in the input, or are using a different pseudopotential, etc... You can also check earlier on that you are getting the same total energy forces etc...

The physical criterion is whether the end result is the same for the same required tolerances etc... The tests are not meant to be physically meaningful! They are meant to be quick and to port the same results on all platforms


Matthieu
Matthieu Verstraete
University of Liege, Belgium

Locked