Force mixing potential

[stenczelt]

There are other ways to construct a Potential that accept pot1 and pot2, e.g. a QM/MM-style ForceMixingPotential. Might be worth checking if this code works correctly for that use case too.

Originally posted by @jameskermode in https://github.com/libAtoms/QUIP/pull/185

[stenczelt]

I took a look at the py2 code and the currently wrapped one as well. we have quippy.potential_module.Potential_FM as well, so should probably use that directly. I think this needs more insight and time to develop a reliable wrapper for, which we could do together with the DynamicalSystems, that we have started but left unfinished at the end of my last visit

Originally posted by @stenczelt in https://github.com/libAtoms/QUIP/pull/185

[gabor1]

I thought that DS wouldn't be wrapped because it is superseded by ASE. What functionality does it offer that ASE doesn't have?

[jameskermode]

Principally the fancy thermostats, e.g. massive Nose-Hoover-Langevin

[gabor1]

Hm.. yes, Noam (and others?) did a big chunk of work on the thermostat that's only in quip. How much of that stuff made it into lammps in the end? I thought Leimkuhler worked with them on that.

[bernstei]

On Apr 27, 2020, at 5:01 PM, gabor1 notifications@github.com wrote:

Hm.. yes, Noam did a big chunk of work on the thermostat that's only in quip. How much of that stuff made it into lammps in the end? I thought Leimkuhler worked with them on that.

I see no evidence of it when I search for lammps thermostats. I suppose I could consider refactoring out a useful subset and re-implementing it in ASE and/or LAMMPS?

                                Noam

[bernstei]

And I could perhaps do ABOBA (position Verlet) at the same time, which Leimkuhler said is magically accurate.

                        Noam

---

|| |U.S. NAVAL| |RESEARCH| LABORATORY Noam Bernstein, Ph.D. Center for Materials Physics and Technology U.S. Naval Research Laboratory T +1 202 404 8628 F +1 202 404 7546 https://www.nrl.navy.mil https://www.nrl.navy.mil/

[gabor1]

Maybe worth asking BL first if they have something in the works?

-- Gábor

Gábor Csányi Professor of Molecular Modelling Engineering Laboratory, University of Cambridge Pembroke College Cambridge

Pembroke College supports CARA. A Lifeline to Academics at Risk. http://www.cara.ngo/

On 27 Apr 2020, at 22:33, bernstei notifications@github.com wrote:

On Apr 27, 2020, at 5:01 PM, gabor1 notifications@github.com wrote:

Hm.. yes, Noam did a big chunk of work on the thermostat that's only in quip. How much of that stuff made it into lammps in the end? I thought Leimkuhler worked with them on that.

I see no evidence of it when I search for lammps thermostats. I suppose I could consider refactoring out a useful subset and re-implementing it in ASE and/or LAMMPS?

Noam

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[bernstei]

On Apr 27, 2020, at 5:37 PM, gabor1 notifications@github.com wrote:

Maybe worth asking BL first if they have something in the works?

Probably. Also, do we think that being in LAMMPS is sufficient, or do we really want a separate implementation in ASE also?