problems of improper dihedral calculation with Fist

[姚懿]

And some modification to the gradients calculation is needed.

    ! compute the gradients
    gt1 =  ( s32 * ism * ism )  * SIGN(1.0_dp , DOT_PRODUCT ( t12, tn)) * tm
    gt4 = -( s32 * isn * isn )  * SIGN(1.0_dp , DOT_PRODUCT ( t12, tn)) * tn

On Friday, August 14, 2015 at 11:28:30 AM UTC-4, 姚懿 wrote:
>
> In the mol_force.F file. there is one line in force_imp_torsions subroutine
>
> phi = SIGN( ACOS ( cosphi ), DOT_PRODUCT ( t12, tn ))
>
> I suggest to change it to 
>
> phi = ACOS( cosphi )
>
> since improper potential is not periodic. For a dihedral close to 180 
> degree it can varies between -180.0 and 180.0 that will influence the 
> result a lot.
>
> YY
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150814/93d2d33e/attachment.htm>