[Pw_forum] md calculation and constraints in 4.3.2

[Paolo Giannozzi]

On Mon, 2012-01-09 at 15:53 +0100, Jean-Roch Huntzinger wrote:

> in PW/input.f90 there is, line 1436 :
>   IF ( lconstrain ) THEN
>      IF ( .NOT. ldamped .OR. lmovecell ) CALL errore( 'iosys', &
>               'constraints only with damped dynamics and fixed cell', 1 )
> 
> which seems to forbid constraints with verlet ?

may I suggest experiment instead of theory? replace the above check 
with a weaker one

     IF ( lbfgs .OR. lmovecell ) CALL errore( 'iosys', &
                'constraints only with fixed-cell dynamics', 1 )

and please report if it works or not.

> Note: changing 'verlet' to 'langevin' yields the same error, 

"langevin" is a poorly documented unmaintained option

> And actually ldamped is not defined for 'md' calculations 

because it is used for minimization, not MD. There are two algorithms
for minimization, BFGS (no constraints) and damped Verlet (constraints
implemented)

P.
-- 
Paolo Giannozzi, IOM-Democritos and University of Udine, Italy