[Wien] Re: min force
WANG YX
wang.yuanxu at nims.go.jp
Thu Aug 26 01:45:03 CEST 2004
Dear Dr. L. D. Marks,
Thank you for your reply!
I agree your suggestion.
But in somr case (such as PbTiO3), even if I use "PORT 1.0", it still have some warning(such as S.Y WAS 0.251222294751182701E-008).
In order to get well converged and high-quality calculation, how should I do? Increase k-points? or increase R-MT*K-max?or some other option.
For example PbTiO3 with a=3.92 c =4.24 (P4mm), optimize structure.
Best
WANG YX
wang.yuanxu at nims.go.jp
2004-08-26
======= 2004-08-24 04:16:20 you wrote=======
>When I look with a grep at your case.outputM, this is what I get is shown
>below. In the first step, you'll see that the energy does not follow a
>smooth, linear sequence as the step is reduced but instead is oscillating.
>You will also see two warnings, one that the curvature condition failed,
>and another that you might have inconsistent forces/energy.
>
>In order to minimize the forces to 0.1 (! a very small number) you have to
>have a high-quality, well converged calculation. It looks like you do not.
>Without inspecting how you did your calculation (which is your job to do),
>I cannot say what has gone wrong, but I can see that it is the runXYZ_lapw
>is not "good enough" to warrant minimization to 0.1.
>
> grep -e :WARN -e :MIN -e :ENE -e DD qqq
> :MIN Minimizing for 24 parameters
> :ENE Retrived energy -67.7541190000000029
>:MIN IT NF E RELDF PRELDF RELDX STPPAR D*STEP
>NPRELDF
>:DD7DOG Newton Step, radius 0.125E-02
> :ENE Retrived energy -67.7541140000000013
>:DD7DOG Cauchy Step 0.276 Radius 0.312E-03
> :ENE Retrived energy -67.7541039999999981
>:DD7DOG Cauchy Step 0.028 Radius 0.312E-04
> :ENE Retrived energy -67.7541209999999978
> :WARNING, CURVATURE CONDITION FAILED
> :WARNING, S.Y WAS 0.251222294751182701E-008
>:MIN 1 4-0.5000D+00 0.4D-05 0.3D-06 0.6D-06 0.4D+02 0.3D-04
>0.5D-05
>:DD7DOG Cauchy Step 0.040 Radius 0.625E-04
> :ENE Retrived energy -67.7541159999999962
>:DD7DOG Cauchy Step 0.004 Radius 0.625E-05
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.002 Radius 0.312E-05
> :ENE Retrived energy -67.7541200000014214
>:DD7DOG Cauchy Step 0.000 Radius 0.312E-06
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.156E-06
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.781E-07
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.390E-07
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.195E-07
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.976E-08
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.488E-08
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.244E-08
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.122E-08
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.610E-09
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.305E-09
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.152E-09
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.762E-10
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.381E-10
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.191E-10
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.953E-11
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.476E-11
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.238E-11
> :ENE Retrived energy -67.7541209999999978
>:DD7DOG Cauchy Step 0.000 Radius 0.119E-11
> :ENE Retrived energy -67.7541209999999978
>:MIN 2 26-0.5000D+00 0.0D+00 0.1D-13 0.2D-13 0.1D+10 0.1D-11
>0.7D-05
>:MIN I FINAL X(I) D(I) G(I)
> :MIN Probably converged to numerical accuracy
> :WARNING you might have inconsistent Forces/Energies
> :MIN BFGS minimizer drmng terminating
> :MIN Final function value -0.500001999999994950
>
More information about the Wien
mailing list