On 6/26/18 10:37 AM, Eric Smoll wrote:
Justin,
In a previous thread on core/shell optimization (
https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-users/2014-August/091101.html)
you state:
"No, I've made wholesale changes to the code in the way that energies and
forces
are computed. The current code has bugs. If you want a modified version
(beta
software, so caveat emptor), contact me off-list."
Is this still true? Is core/shell optimization broken in GROMACS 2018.1?
If so, would you mind sharing your Drude implementation?
Check out git master and switch to the drude branch. Don't try to use
domain decomposition, only OpenMP for parallelization. Documentation is
unfortunately somewhat sparse but the code is reasonably commented.
Note that everything I have done is for extended Lagrangian dynamics, I
haven't tested much with massless shells or use of [polarization]
directives.
-Justin
Best,
Eric
On Mon, Jun 18, 2018 at 1:14 PM, Justin Lemkul <jalem...@vt.edu> wrote:
On 6/18/18 4:05 PM, Eric Smoll wrote:
Justin,
Thank you so much for the rapid and clear reply! Sorry to ask for a bit
more clarification.
The thole_polarization isn't in the manual at all. Is it structured the
same way as the [ polarization ] directive in the manual:
[ thole_polarization ]
; Atom i j type alpha
1 2 1 0.001
If I want Thole corrections, am I correct in assuming that I should list
*all shells* in the system under this thole_polarization directive with
(as
you pointed out) "i" or "j" as the shell? If "i" is the shell, "j" is the
core. If "j" is the shell, "i" is the core.
You have to list everything explicitly, including the shielding factor,
and between dipole pairs (Thole isn't between an atom and its shell, it's
between neighboring dipoles). I honestly don't know what the format is; I
completely re-wrote the Thole code for our Drude implementation (still not
officially incorporated into a release due to DD issues, but we're close to
a fix...)
The code for "init_shell_flexcon" was very helpful. Thank you!
nstcalcenergy must be set to 1. The code says that domain decomposition
is not supported so multi-node MPI calculations are not allowed. I can
still use an MPI-enabled GROMACS executable on a single node for shell MD,
correct? Thread parallelization is still permitted, correct?
Presumably you're limited to OpenMPI, but again I have no idea about this
code. I've never actually used it.
-Justin
--
==================================================
Justin A. Lemkul, Ph.D.
Assistant Professor
Virginia Tech Department of Biochemistry
303 Engel Hall
340 West Campus Dr.
Blacksburg, VA 24061
jalem...@vt.edu | (540) 231-3129
http://www.thelemkullab.com
==================================================
--
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-requ...@gromacs.org.
--
==================================================
Justin A. Lemkul, Ph.D.
Assistant Professor
Virginia Tech Department of Biochemistry
303 Engel Hall
340 West Campus Dr.
Blacksburg, VA 24061
jalem...@vt.edu | (540) 231-3129
http://www.thelemkullab.com
==================================================
--
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-requ...@gromacs.org.
