On Thu, Jan 10, 2013, Fabian Zeller wrote:
>
> This refers to all nmr group restraints, excluding distance restraints,
> that have a zero energy area in the middle of the parabola (r2=r3). The
> restraint forces seem to be not correctly set to zero in this area.
>
> For example, see subroutine angnrg. In case (thet <= r3), the energy is
> set to zero, but not the dfr force array. Then the values are returned
> (omitting the later assignment of dfr):
>
> // else if (thet <= r3) then//
> // e = 0.0d0//
> // return//
> // else if (thet < r4) then//
I agree: the simplest (and safest, I think) solution is to set df=0 and remove
the return statement. There is a negligible performance advantange to
returning early, and removing the return statement is a protection against
future code changes. This makes the "thet <= r3" option parallel to all the
other options.
There are lots of these places in nmr.F90. I've created in a patch to sander
in the developers' version, and will be testing it. (pmemd updates to
follow).
Thanks for the report!....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jan 11 2013 - 07:30:04 PST
