[Pw_forum] Relaxation doesn't converge

[Lorenzo Paulatto]

In data 02 settembre 2009 alle ore 13:57:25, Pablo Aguado  
<paguado at gmail.com> ha scritto:
> Only z is allowed to relax (cell_dofree = 'z') so:
> sigma_xx & sigma_yy should be set to 0 at some point
> sigma = 0.87 Kbar press_conv_thr = epsp = 1.0 Kbar
> press = 0.0


I can understand your concern: you would expect that if only z can move  
the pressure along x and y would not be considered for convergence. I  
think you are right, and it should be implemented that way. In order to do  
that you can modify PW/vcsmd.f90 in the following way:

diff -r1.36 vcsmd.f90
215c215
<                     ( ABS( sigma(i,i) - press) * uakbar < epsp )
---
>                     ( ABS( sigma(i,i) - press) * uakbar * iforceh(i,i) <  
> epsp )
219c219
<                        ( ABS( sigma(i,j) ) * uakbar < epsp )
---
>                        ( ABS( sigma(i,j) ) * uakbar * iforceh(i,j) <  
> epsp )

i.e. adding iforceh(*,*) where necessary. Could you please test this patch?

best regards

-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
phone: +39 040 3787 511
skype: paulatz
www:   http://people.sissa.it/~paulatto/

     *** save italian brains ***
  http://saveitalianbrains.wordpress.com/