forum@abinit.org

Subject: The ABINIT Users Mailing List ( CLOSED )

List archive

[abinit-forum] problem with electronic convergence

From: Rye Terrell <ryeterrell@ryeterrell.net>
To: forum <forum@abinit.org>
Subject: [abinit-forum] problem with electronic convergence
Date: Mon, 12 Oct 2009 00:56:48 -0500
Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:date:x-google-sender-auth:message-id:subject :from:to:content-type; b=QacdGCpDyrPMmUlkanf0LCIAdOtgmbO2Sp5XQfOij42sMTClHQjmsx41GyXADaYzaG kd6Cit7aksGKSp+wctLNkUBHQJTN/9YM/iluXgAC2KpNmutPoi7A3DqDMCYzN9Nwa5co S8k9BQEC9jTRQki4yS/wa31B/xsP1gyW+11GM=

Hi everyone,

I submit the following input file to abinit:

ecut 368 eV
nstep 512
toldfe 1.0d-3 eV

acell 3.843479 3.843479 3.843479 angstroms
rprim 1 0 0 0 1 0 0 0 1
ntypat 2
znucl   1 77
natom 6
typat 2 2 2 2 1 1
xangst 0.192173926625 0.192173926625 0.192173926625
        0.192173926625   2.1139131929   2.1139131929
          2.1139131929 0.192173926625   2.1139131929
          2.1139131929   2.1139131929 0.192173926625
        0.330133922984 1.97581998601 0.330339691348
         3.05034388421 3.09920533477 3.05037017781

For some reason, my calculation does not converge very well:

...
ETOT509 -99.967600677647     7.946E-03 2.238E-05 1.743E+01 1.348E-02 5.832E-02
ETOT510 -99.971348913873    -3.748E-03 3.132E-05 1.371E+01 6.212E-03 6.023E-02
ETOT511 -99.970856919498     4.920E-04 9.470E-06 1.497E+01 1.438E-02 5.768E-02
ETOT512 -99.962207567125     8.649E-03 9.493E-05 2.549E+01 9.176E-02 5.033E-02

Cartesian components of stress tensor (hartree/bohr^3)
sigma(1 1)= -1.48092765E-03 sigma(3 2)= 4.93782918E-05
sigma(2 2)= -1.19629690E-03 sigma(3 1)= -9.22122242E-06
sigma(3 3)= -6.46521337E-04 sigma(2 1)= -7.09301595E-04

scprqt: WARNING -
nstep= 512 was not enough SCF cycles to converge;
maximum energy difference= 8.649E-03 exceeds toldfe= 3.675E-05

Can anyone tell me where I need to start looking to correct this?

Thanks for any advice!

Cheers,
Rye

[abinit-forum] problem with electronic convergence, Rye Terrell, 10/12/2009
- Re: [abinit-forum] problem with electronic convergence, pascal boulet, 10/12/2009