Quip potential

LAMMPS LAMMPS Development
, ,
1

Hi,
I am trying to run quip pair_style in lammps ( 9 Nov 21) to perform MD using my trained gap potential. I trained GAP potential for sodium data using distance_2b and soap as descriptors. But when I run the input file, I get the following warnings and no data are given.

WARNING: No pairwise cutoff or binsize set. Atom sorting therefore disabled. (src/atom.cpp:2179)
WARNING: Communication cutoff is 0.0. No ghost atoms will be generated. Atoms may get lost. (src/comm_brick.cpp:209)

The relevant part of LAMMPS input script:

pair_style quip
pair_coeff * * gap_example.Na.xml “Potential xml_label=GAP_2021_12_24_210_5_59_39_424” 11
neighbor 0.2 bin
neigh_modify delay 10
fix 1 all nvt temp 300.0 300.0 50.0
thermo 1
timestep 0.001

while the example part in lammps package runs very well.
I have no idea how to fix this and would appreciate any help.
Thank you

2

The cutoff must be set somewhere in the XML file.
The LAMMPS pair style code just contains a rather minimal wrapper for the quip library code, so for more details on this, you will be better off to contact the libatoms/quip/gap developers.

3

If you look at the source code for pair style quip, there is simply a call to the function quip_lammps_potential_initialise() which passes the information from the pair_coeff command and receives the cutoff in return.

4

Thank you.
So you mean those warnings have nothing to do with input file of lammps and I have to go through my xml file.

5

Didn’t I state that in my first response already?

6

yes you did and thank you for reply.