Dac,
We will do the implementation as you suggest to allow the middle scheme to
be used for parallel MD as well. It should be completed soon. We will
update the change.
We will also implement the middle scheme into AMBER for GPU, but this will
be done for the next version of AMBER.
All the best
Jian
On Mon, May 21, 2018 at 9:16 PM, David A Case <david.case.rutgers.edu>
wrote:
> On Mon, May 21, 2018, Jian Liu wrote:
>
> > The "middle" scheme was implemented in sander for MD and in sander.mpi
> for
> > PIMD for the NVT ensemble.
>
> Sounds like we need to prepare an update to trap attempts to use the
> middle scheme in sander.MPI if PIMD is not activated, exiting with a
> useful error message. (Or: allow the middle scheme to be used for
> parallel MD--it that a simple task?)
>
> Jian: if your group can suggest changes, I can arrange to get those
> added to the bugfix list.
>
> ...thx...dac
>
> > On Mon, May 21, 2018 at 12:03 PM, Pengfei Li <lipengfei_mail.126.com>
> wrote:
> >
> > > Recently, I just test the “middle” scheme, which can be obtained in
> > > AmberTools 18.
> > > But I find that when the sander engine is used, heating simulation can
> > > continue to end.
> > > However, when the sander.MPI engine is used, heating simulation will
> stop
> > > at initial step of simulation and give no messages about errors.
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue May 22 2018 - 02:30:02 PDT
