On Fri, Sep 14, 2007, Ilyas Yildirim wrote:
> noshakemask = '.%DH | (:4,12.H6) | (:4,12.%H)',
>
> Shake option is selected; and noshakemask is used. When I look at the .out
> file; I see the following:
>
> ----
> Noshake mask .%DH | (:4,12.H6) | (:4,12.%H); matches 4 atoms
> Removing shake constraints from N6 RA 4 -- H61 RA 4
> Removing shake constraints from N6 RA 4 -- H62 RA 4
> Removing shake constraints from N6 RA 12 -- H61 RA 12
> Removing shake constraints from N6 RA 12 -- H62 RA 12
> Setting ntf to 1
>
> I thought that when shake and noshakemask is defined in the input file,
> the constraints will be applied on both states explicitly (for instance, I
> was expecting the above sander output file to show not just 4 atoms, but
> 10 atoms). Thanks.
The mask is probably parsed using the first state's prmtop file, although you
should check the code to make sure. If there is any question about which
atoms the noshakemask applies to, use absolute atom numbers (not atom types
and so on) to be sure that it is doing what you want. [In Amber9, the atom
numbers have to be the same in the two states.]
We need to put a comment in the documentation pointing out this potential
problem.
...dac
Received on Sun Sep 16 2007 - 06:07:51 PDT
