[gmx-users] AWH with multiple parallel simulations

[Berk Hess]

Hi,

Indeed, the former can be very beneficial, even in the case all replicas start at the same point.
The latter is an interesting idea (even better when using REST instead of T-REMD), but would require adaption of the exchange criterion. Just exchanging ignoring the bias will likely give completely incorrect results.

Cheers,

Berk

________________________________
From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of Mark Abraham <mark.j.abraham at gmail.com>
Sent: Thursday, March 28, 2019 3:02 PM
To: gmx-users at gromacs.org
Cc: gromacs.org_gmx-users at maillist.sys.kth.se
Subject: Re: [gmx-users] AWH with multiple parallel simulations

Hi,

The former is definitely an intended use case. We observe great efficiency
improvements - even super-linear for adding just a second replica. While
the latter is likely useful in principle, I'm not sure offhand that it is
implemented, and I agree that in any case bias exchange would only be
physical between replicas at a common T.

Mark

On Thu, 28 Mar 2019 at 14:28 Piotr Setny <p.setny at cent.uw.edu.pl> wrote:

> Dear GMX users and developers,
>
> I'm evaluating free energy profile along certain reaction coordinate. In a
> simplified test run I used the accelerated weight histogram method (AWH)
> and it worked pretty well. Now, for more complex system I intend to further
> enhance sampling by using multiple parallel simulations. I consider two
> options:
>
> - to use multiple AWH runs at single target temperature (T), each started
> at different point of the reaction coordinate and allow bias exchange,
> - to use multiple AWH combined with T-REMD, i.e. each AWH run at different
> T, allowing for T exchange but no bias exchange (I believe the latter would
> by unphysical in this case).
>
> Do you have any thoughts/recommendations concerning the efficiency and
> suitability of the two options?
>
> Best,
> Piotr
>
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.
>
--
Gromacs Users mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.