npt integrator

[Tommy Waltmann]

Hi all,

I have read the documentation on the npt integrator, but I cannot make it work for a simple system consisting of a Toluene solvent. When I set the integrator, I supply a group of rigid Toluene centers, tau=1.0, kT=.03877, P=6.32e-7 and the error reads "Incompatible function arguments. The following types are supported: (a big list of scary looking stuff).". I have no idea what this means is wrong with my script. Can someone please help me find the problem?

[Jens Glaser]

Are you also supplying tauP (the barostat time scale)?

Jens

On Jun 19, 2018, at 4:49 PM, Tommy Waltmann <twalt...@gmail.com> wrote:

Hi all,

I have read the documentation on the npt integrator, but I cannot make it work for a simple system consisting of a Toluene solvent. When I set the integrator, I supply a group of rigid Toluene centers, tau=1.0, kT=.03877, P=6.32e-7 and the error reads "Incompatible function arguments. The following types are supported: (a big list of scary looking stuff).". I have no idea what this means is wrong with my script. Can someone please help me find the problem?

--
You received this message because you are subscribed to the Google Groups "hoomd-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hoomd-users...@googlegroups.com.
To post to this group, send email to hoomd...@googlegroups.com.
Visit this group at https://groups.google.com/group/hoomd-users.
For more options, visit https://groups.google.com/d/optout.

[Tommy Waltmann]

No I am not. The argument is not required in the hoomd.md.integrate.npt function. If I need to set it, what would be a good value to set it at?

[Jens Glaser]

According to the source code, it should produce an error if tauP is not supplied, so I am a little confused why you wouldn’t see that.

In any case, similar to tau, it sets the time scale of the damping of pressure fluctuations.It may be anything between 0.1*tau, for example (tau = LJ time) and a multiple of tau, for very strongly fluctuating boxes (such as initially dilute boxes). The value only affects the dynamics of the box, not equilibrium of course.

- Jens

[Tommy Waltmann]

My simulation is working now after I supplied a value of tauP. Is it better to have a large tauP if I think my box will have large fluctuations during the simulation?

[Jens Glaser]

It makes sense to experiment a bit, a too large value will slow down equilibration of the box degree of freedom, but a finite value is required to not get out of box errors because particles crashing into each other because of too fast box size changes.

[Tommy Waltmann]

Ok, thank you!

[Tommy Waltmann]

Now that I have gotten my system to equilibrate successfully, I have one more question about the output data. The pressure unit that corresponds to 1atm in my simulation is ~10^-7 (the value that I set my simulation at), and as I look at my simulation data I see that the fluctuations in pressure are ~10^-4. can I still trust that my other simulation data is correct even though the pressure fluctuations are larger than the actual pressure itself?

[Jens Glaser]

That is very typical. The pressure fluctuations are determined by the compressibility of your system (look up fluctuation-response relations), which can be quite large, depending on density e.g.

Jens