failed convergance for a single atom
Clicked A Few Times
12:48:18 PM PDT - Fri, Mar 28th 2014
Hi,

I'm trying to simulate single uranium atom (spherically symmetric and spin-restricted). I failed to get a converged result, although I tried to tune convergence parameters. the part of the input is like follows:


geometry
U 0.0 0.0 0.0
symmetry oh
end

charge +2

basis spherical
U library SARC-ZORA
end

relativistic
   ZORA
end

dft
 RODFT

 CGMIN

 xc rpbe

 iterations 300

 convergence damp 40

 convergence lshift 0.6

 convergence rabuck 20

 grid fine

 smear 1d-6

 mult 1
end

task dft



Can anyone give some suggestions to improve the convergence?

Forum Regular
5:27:27 PM PDT - Tue, Apr 8th 2014
Hi Alexx,

With the current development version of the code, running in serial, your calculation converges in 7 iterations:

 Setting level-shift to   2.00 to force positive preconditioner

Warning: symmetry breaking in RKS:   0.164620275346105

                6   -25645.0431550529  5.33D-03  1.54D-03    148.6



*** WARNING: SMALL OCC-VIRT ORBITAL ENERGY DIFFERENCE DETECTED

(Including frequency factor for response)

This is going to lead to diverging respose properties

I have adjusted the denominator to a value of 1E-5



*** WARNING: SMALL OCC-VIRT ORBITAL ENERGY DIFFERENCE DETECTED

(Including frequency factor for response)

This is going to lead to diverging respose properties

I have adjusted the denominator to a value of 1E-5



*** WARNING: SMALL OCC-VIRT ORBITAL ENERGY DIFFERENCE DETECTED

(Including frequency factor for response)

This is going to lead to diverging respose properties

I have adjusted the denominator to a value of 1E-5



*** WARNING: SMALL OCC-VIRT ORBITAL ENERGY DIFFERENCE DETECTED

(Including frequency factor for response)

This is going to lead to diverging respose properties

I have adjusted the denominator to a value of 1E-5

Warning: symmetry breaking in RKS:   0.164681101004830

                7   -25645.0431560512  2.64D-04  9.22D-05    150.2




        Total DFT energy =   -25645.043156051197

     One electron energy =   -35173.158088234217

          Coulomb energy =     9957.539244877560

   Exchange-Corr. energy =     -429.424312694540

Nuclear repulsion energy =        0.000000000000



Numeric. integr. density =       90.000000001906

So maybe there is something else in your input that gets in the way (your post seemed to suggest that the Uranium atom input was just a part of a larger input file), maybe some issue in the code has been fixed in the meantime. Can you provide some more details on what you are trying to do, please?

Huub


Forum >> NWChem's corner >> General Topics