[lammps-users] Coordination Number

Hi,

Why is there a restriction on the cut-off for the compute coordination number per atom?

Pooja

Hi,
Why is there a restriction on the cut-off for the compute coordination
number per atom?

because information beyond that distance is not guaranteed to be
available on a single node/domain in a parallel run.

axel.

So what happens if I specify the following:

I use 1.1225 as the cut-off for LJ and 2.5 for pair coeff… can I then use 2.5 as a cut-off for n-coord. I have tried and it seems to run but I would like to know how do things change? The potential remains the same?

pair_style lj/cut 1.1225
pair_coeff 1 1 1.0 1.0 2.5
pair_modify shift yes

Pooja

So what happens if I specify the following:

I use 1.1225 as the cut-off for LJ and 2.5 for pair coeff.. can I then use
2.5 as a cut-off for n-coord. I have tried and it seems to run but I would
like to know how do things change? The potential remains the same?

you can check the latter _extremely_ easily, by comparing the energy output.

pair_style lj/cut 1.1225
pair_coeff 1 1 1.0 1.0 2.5
pair_modify shift yes

this will probably not do what you want.
1.1225 is the global default value and it will be
overridded by the 2.5 in the pair_coeff statement.

you probably want this to be the other way around.

axel.

The compte coord/atom command will tell
you if the cutoff you specify with it is longer
than can be used.

Steve

So I need to write a new compute which calculates an order parameter based on the nearest neighbour list. I had build upon the compute_coord code earlier but I need the Rcutoff for this calculation to be greater than that for pair-wise potential. Would it be a good idea to look at the compute_rdf code instead? Is there another way around this?

Pooja

The cutoff used for pairwise computation (any pair style) and the
associated neighbor list is the maximum cutoff that any compute
can use. Atoms simply don't know any info about their neighbors
further away than that.

Steve