VMD-L Mailing List

From: Tom (mooreth42_at_gmail.com)
Date: Fri May 31 2024 - 15:29:37 CDT

Hi Josh,

Thank you for the suggestion. It led to the solution! When you wrote that
"it takes a while for a small tug on a single atom to manifest itself...."
I decided to change the trate in the LAMMPS input file from 1 (default) to
10, so I believe LAMMPS communicates with VMD 10 times slower. I will have
to double check that in the documentation. However the IMD force applying
works now even with a single atom.

Thanks again, GO SPARTANS (I'm an alumnus)!

Tom

On Fri, May 31, 2024 at 3:51 PM Josh Vermaas <vermaasj_at_msu.edu> wrote:

> Hi Tom,
>
> I'm less familiar on the LAMMPS side, but if you see the trajectory steps,
> I *think* that means that VMD is emitting forces every step by default.
> Have you tried forcing entire residues? It takes a while for a small tug on
> a single atom to manifest as a movement that you will be able to see. If
> you yank on more atoms, you should see a result more quickly just to debug
> if LAMMPS and VMD are interoperating correctly.
>
> -Josh
>
> On 5/31/24 1:55 PM, Tom wrote:
>
> Greetings,
>
> I am using VMD version 1.9.3 and LAMMPS version 2 Aug 2023 - Update 1 On
> Windows 10 Home. I downloaded and installed binaries of both.
>
> My LAMMPS input script is below. I am able to connect VMD to a LAMMPS IMD
> simulation on my local machine (127.0.0.1, port 5678), and VMD displays the
> trajectory steps just fine. However I cannot seem to apply any forces to
> atoms using “Mouse -> Force -> Atom”. I do see the red force arrows on the
> atoms, but I don’t believe any forces are being applied based on:
>
> 1. The structure does not bend as it does when I use the addforce fix with
> a force equal to or less than I have set for the IMD fscale. I’m using real
> units in LAMMPS and a fscale of 10 would be 10*1(kcal/mol)/Angstrom, and
> that should produce a nice bend in the structure. I’ve also tried very
> large numbers like 20,000.
>
> 2. The LAMMPS IMD fix documentation says “If IMD control messages are
> received, a line of text describing the message and its effect will be
> printed to the LAMMPS output screen, if screen output is active.”. I do not
> see anything like this go by in the terminal while LAMMPS is running. This
> makes me think the force IMD commands from VMD is not getting to LAMMPS.
> Though I can start and stop the simulation from the VMD IMD Connection
> window.
>
> I’ve checked that both LAMMPS and VMD have firewall permissions. They
> appear to.
>
> Could anyone please recommend something else to try? Thanks in advance.
>
>
>
> Tom
>
>
> #LAMMPS input file
>
> # Initialization
>
> variable myfile string DNW
>
>
> log ${myfile}.init.log
>
> units real
>
> atom_style molecular
>
> boundary f f f #fixed non-periodic
>
>
> bond_style morse
>
> angle_style harmonic
>
> pair_style lj/cut 14
>
> special_bonds lj 0.0 0.0 0.5
>
>
> # System definition
>
> read_data DNW.data
>
> include paramV04.lammps
>
>
> # set groups
>
> group lonsfree type 1 2
>
> group anchorgroup type 4 5
>
> group forcegroup type 3
>
>
> # Dynamics
>
> velocity lonsfree create 6 48455 mom yes rot yes
>
> fix 1 all nve
>
>
> #freeze the anchor atoms
>
> fix anchor_fix anchorgroup setforce 0.0 0.0 0.0
>
> velocity anchorgroup set 0.0 0.0 0.0
>
>
> #make IMD connection
>
> fix comm all imd 5678 unwrap off fscale 20.0 trate 1 nowait off
>
>
> timestep 1.0 #femtoseconds
>
> dump mydmp1 all atom 100 DNW.lammpstrj
>
> thermo 1000
>
>
> --
> Josh Vermaas
> vermaasj_at_msu.edu
> Assistant Professor, Plant Research Laboratory and Biochemistry and Molecular Biology
> Michigan State Universityvermaaslab.github.io
>
>