ABF error with colvars module

User discussions
1

GROMACS version:2022.5-Colvars_2023_05_01
GROMACS modification: Yes
I am attempting to use ABF in the colvars module for my simulation. I have tried installing several versions of Gromacs (gromacs2021.6-colvars, fresh download of gromacs2022.5 that I patched with colvars). I have tried installing them on my own laptop in Ubuntu22.04 on WSL2 as well as in a folder I have write permission to in our HPC environment. The error I’m getting is “colvars: Error: total forces are currently not implemented.”
I have learned that if I delete the abf block within the colvars.dat file and keep the colvar and harmonicwalls block the simulation will run just fine.

Any help is greatly appreciated.

Thanks,

Robert Coffman

2

Hi Robert,
Indeed, the projected force calculation that is the basis of “classic ABF” is not available so far in Gromacs. However, you can run the extended-system variant eABF by adding the extended Lagrangian options to the collective variables involved.
http://colvars.github.io/colvars-refman-gromacs/colvars-refman-gromacs.html#sec:eABF
http://colvars.github.io/colvars-refman-gromacs/colvars-refman-gromacs.html#sec:colvar_extended

Mostly, you need to add to the colvar block:

extendedLagrangian on
extendedFluctuation <positive decimal, a small fluctuation of the variable>

The Colvars / Gromacs manual needs to be updated to reflect this. Thanks for your report.

Best,
Jérôme

3

Jérôme,
Thank you! This is very helpful to know.

Sincerely,

Robert

4

I am running into a different issue now. I added extendedLangrangian and extendedFluctuation to my colvar block but I’m getting the error that the keyword isn’t supported or not recognized in this context. I have pasted in my colvar block below.
colvar {
name dist1
distance {
group1 { #add in grouping information here.
indexGroup Transmembrane_helix
}
group2 { #add in grouping information here.
indexGroup Headgroup_residues
}
}
extendedLangrangian on
extendedFluctuation 0.03
lowerBoundary 5.0
upperBoundary 5.5
width 0.05
}

Thanks,

Robert

5

Hi Robert, try “extendedLagrangian”, not “extendedLangrangian” :-)

6

Haha, Thanks for you help. It is working great now.