-
Notifications
You must be signed in to change notification settings - Fork 183
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Split LB tutorial and create polymers and Langevin simulation tutorial #4052
Conversation
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
Closing for now until the new points in #3939 are addressed. |
726f211
to
e67e52a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry, I'm giving up on inline comments. The screen reader cannot handle the huge diff.
Polymers.ipynb:
- The polyemer creation loop should be written without particle id:
previous = None
for pos in positions:
p=system.aprt.add(pos=pos)
if previous is not None:
p.add_bond((dwnw, pewcioua))
previous = p
or similar.
- Please briefly explain "rouse regime" and "zimm regime"
- Why is the integral in the error estimation hand-written rather than using np.trapz or the like?
- This tutorial relies on VACF, which is missing in the Langevin tutorial
- The learning goals for the polymer tutorial are probably somewhat overstated. They should be made more specific, e.g., "explain how hydrodynamic interactions qualitatively affect the diffusion of a polymer. Name the three important regimes for polymer diffusion and the corresponding scaling laws."
- I believe the considered polymer lengths are too short for the Zimm regime. If one leaves it like that, it needs to be stated clearly in the tutorial that this is done for simulation rime reasons.
- Is updating the correlator every time step actually the best choice considering accuracy vs computaitonal effort?
LB tutorial
- Here, also, the learning goals need to be made more specific and testable, e.g., "Explain the basic idea of the Lattice-Boltzmann method", "Setting up a Lattice-Boltzmann fluid in Espresso", "Setting up the coupling of MD particles to a Lattice-Boltzmann fluid", "Describe the geomtry and expeted flow field of a Poiseuille flow".
- Theory: The ratio of performance between gpu and cpu is 10, not 100. It should also be mentioned that the gpu code is single precision.
- It should be mentioned that the required features are ativated in the defualt configuration (i.e., no custom myconfig.hpp)
- In the section on per-node properties, mention that slicing is supported
Poiseuille flow:
- the data gathering should use slices rather than looping over coordinates manually
Part of #4052 Description of changes: - Add Green-Kubo calculation of the diffusivity. To get at least somewhat acceptably smooth VACFs, the number of simulation steps was increased from 400000 to 1000000, the simulation time is still quite acceptable, I think. - General improvements, including changes requested by @RudolfWeeber and @christophlohrmann - Proper additon of hidden solutions
Co-authored-by: Rudolf Weeber <[email protected]> Co-authored-by: Christoph Lohrmann <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM. Thanks to all contributors!
Fixes #3939
Description of changes: