Reevaluating energy in geometry optimization

[e khodayar]

Hi,

In the main output file geometry optimization,  at the end of  geometry optimisation step, I have the following information:

 --------  Informations at step =    53 ------------

  Optimization Method        =                   CG

  Total Energy               =       -94.0112708604

  Real energy change         =        -0.0000036472

  Decrease in energy         =                  YES

  Used time                  =              343.953

  Convergence check :

  Max. step size             =         0.0006162087

  Conv. limit for step size  =         0.0010000000

  Convergence in step size   =                  YES

  RMS step size              =         0.0002737159

  Conv. limit for RMS step   =         0.0010000000

  Convergence in RMS step    =                  YES

  Max. gradient              =         0.0005305121

  Conv. limit for gradients  =         0.0010000000

  Conv. in gradients         =                  YES

  RMS gradient               =         0.0002356500

  Conv. limit for RMS grad.  =         0.0010000000

  Conv. in RMS gradients     =                  YES

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

 *******************************************************************************

 ***                    GEOMETRY OPTIMIZATION COMPLETED                      ***

 *******************************************************************************

                    Reevaluating energy at the minimum

 Number of electrons:                                                         68

 Number of occupied orbitals:                                                 34

 Number of molecular orbitals:                                                34

 Number of orbital functions:                                                232

 Number of independent orbital functions:                                    232

 Extrapolation method: ASPC

 SCF WAVEFUNCTION OPTIMIZATION

  Step     Update method      Time    Convergence         Total energy    Change

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

     1 Pulay/Diag. 0.50E+00    3.7     0.00001087       -94.0112708604 -9.40E+01

     2 Pulay/Diag. 0.50E+00    4.7     0.00001294       -94.0112951749 -2.43E-05

     3 Pulay/Diag. 0.50E+00    4.3     0.00000783       -94.0112680643  2.71E-05

  *** SCF run converged in     3 steps ***

  Electronic density on regular grids:        -67.9999999983        0.0000000017

  Core density on regular grids:               68.0000000001        0.0000000001

  Total charge density on r-space grids:        0.0000000017

  Total charge density g-space grids:           0.0000000017

  Overlap energy of the core charge distribution:               0.00000565490189

  Self energy of the core charge distribution:               -223.82607238936723

  Core Hamiltonian energy:                                     69.37181702794591

  Hartree energy:                                              89.86177654697752

  Exchange-correlation energy:                                -29.41879490480421

  Total energy:                                               -94.01126806434611

 ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.):              -94.011268064346112

Now my question is:

Which total energy i have to use for my data?and why this two total energy that with yellow color is different?

Thanks in advance.

[Patrick Gono]

Dear Khodayar,

The reevaluation of the energy at the end of the optimization might yield a different total energy value if your calculation is not converged. If that is the case, you might choose more stringent values for EPS_SCF inside the SCF subsection, which governs the convergence of the SCF loop, and EPS_DEFAULT, which is tied to the numerical accuracy of your calculation. From the excerpt you provide, I see that the energy changed by 2.71E-05 Hartree, or nearly one meV, in the last iteration before convergence. That's rather a lot, and hence I'd advise you to decrease EPS_SCF to, say, 1.0E-6.

However, in the end, the two values of total energy differ by less than 0.08 meV. For most intents and purposes, that's identical, and you can choose whichever.

Yours sincerely,

Patrick Gono

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/0996cb3f-acf7-4e34-b948-7be5936dba1f%40googlegroups.com.

[e khodayar]

Thanks for your reply. 

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAPL1_v8JenHMs7NHOX1%3DWbZQwHwF7-%2Bu2Pbt1mYBbfAQ3gT3Gw%40mail.gmail.com.