[lammps-users] atom_style molecular for worm-like filaments

Dear all,

I am simulating a system with 1428 number of wormlike filaments where each filament consist of 61 beads connected by harmonic bonds and angle potential. The system is in 2D. In x-direction I have PBC and in y direction I have fixed boundaries where I have placed repulsive wall particles.

For such a system can one use atom_style molecular instead of atom_style template. From the documentation I understand that for such systems template style is more effective in terms of memory usage. Other than using comparatively larger memory can atom_style molecular lead to erroneous results for such systems. Specifically, I want to ask if atom_style molecular produce correct bond topology? Does anyone have any experience?

Best regards,
Chandana

you seem to have a severe case of “premature optimization syndrome”.

the system size you are mentioning is far too small to make memory consumption a concern.
the savings from atom style template also depend on the number of bonds/angles per atom. for your kind of system they will be among the smallest possible.

the bond topology for computing interactions will be the same in both cases, if you provide correct input.
the atom style has nothing to do with that. the GI-GO principle always applies.

axel.

Dear Axel,

Thank you for your reply.
One more question: I see in the examples/template directory in “in.mol.data.mix” (file attached) molecule command has been used along with atom_style molecular

molecule cychex cyclohexane.mol
atom_style molecular

Is it necessary to define the molecule command here? The input script runs and produces the same results with/without the molecule definition. I am asking because I did not define it in my simulations.

Actually I have defined a new compute style which calculates local forces on each atom (in a filament) depending on the bond length. The force flips direction in a certain interval. This works fine
for a single filament at any time. However, it seems to stop flipping direction for a system of 1428 polymers after some time of simulations.

Best regards,
Chandana

in.mol-data-mix (465 Bytes)

Dear Axel,

Thank you for your reply.
One more question: I see in the examples/template directory in “in.mol.data.mix” (file attached) molecule command has been used along with atom_style molecular

molecule cychex cyclohexane.mol
atom_style molecular

Is it necessary to define the molecule command here?

this is a superfluous question since you have answered already by yourself.

The input script runs and produces the same results with/without the molecule definition. I am asking because I did not define it in my simulations.

Actually I have defined a new compute style which calculates local forces on each atom (in a filament) depending on the bond length. The force flips direction in a certain interval. This works fine
for a single filament at any time. However, it seems to stop flipping direction for a system of 1428 polymers after some time of simulations.

i don’t see what this has to do with your question.

axel.