Hi Saulius and Torbjoern, What about RMSD of forces? Usually one uses two parameters for the convergence, Max Force and RMS Force.

Cheers, Linas

On Thu, Nov 6, 2014 at 12:43 PM, Björkman Torbjörn < torbjorn.bjorkman@aalto.fi> wrote:

Hi Saulius,

...

"These data items describe residual forces on atoms in the final structure. For a converged computation of a stable structure these should be close to 0.

A residual force on a given atom should be computed with all other parameters, including all other atom positions, fixed. The force components for each atom are given as partial derivatives \d E/\d x, \d E/\d y and \d E/\d z, where E is the total energy of the system and x, y, and z are coordinates of the atom for which residual force is computed (the \d is the partial derivative sign)."

Is this OK? Or are other definitions used in fact?

This looks good and correct. Specifically, the convergence parameter is usually the maximal residual force on any one atom, this is what I think we should demand should be given.

Regards, Torbjörn

---

Torbjörn Björkman, PhD COMP, Aalto University School of Science Espoo, Finland