[Pw_forum] BFGS convergence

[Paolo Giannozzi]

On Monday 28 August 2006 05:21, PRASENJIT GHOSH wrote:

> I'm keeping the bottom 2 layers fixed & relaxing rest of the layers along
> all directions. I have set forc_conv_thr as 5.0D-4. When the BFGS
> optimization is complete the total force printed out is 0.001184 Ry/au.,
> which is higher than the threshold.

this is not the quantity that is checked for convergence, though:
---
forc_conv_thr  REAL ( default = 1.0D-3 )
               convergence threshold on forces (a.u) for ionic
               minimization: the convergence criterion is satisfied
               when all components of all forces are smaller than
               forc_conv_thr.
---

> My question is when the code calculates the force (grad) & their
> differencec between 2 consecutive BFGS iter. (grad_error) & is it only for
> the atoms which are allowed to relax or for the full system or something
> is going wrong in the calculation? I'm using version 3.0.

forces on atoms that are kept fixed are not considered in
the convergence criterion, or else you will never achieve
convergence in presence of a constraint

Paolo
-- 
Paolo Giannozzi             Phone:   +39/050-509876
DEMOCRITOS and SNS          Fax:     +39/050-563513 
Piazza dei Cavalieri 7      I-56126 Pisa, Italy