Skip to content
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.

Sign up for GitHub

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

Jump to bottom

WIP: Stokesian dynamics #3241

Closed
wants to merge 179 commits into from
Closed

WIP: Stokesian dynamics #3241

wants to merge 179 commits into from

Conversation

hmenke
Copy link
Member

@hmenke hmenke commented Oct 8, 2019 •
edited
Loading

TODO:

  • Test case
  • Communication of rotations
  • Thermalization
  • Docs
  • Lubrication corrections (units)

hmenke added 16 commits June 30, 2019 12:53
@hmenke
[SD] Starting with the new implementation
4b40b4f
@hmenke
[SD] Stokesian dynamics in the integration loop
650e940
@hmenke
[SD] Unequal radius, variable viscosity
1a50e1c
@hmenke
[SD] Communicate new positions and velocities
d365bce
@hmenke
[SD] Interface
773042d
@hmenke
[SD] Sample for dancing spheres
134b83f
@hmenke
[SD] Adapt to the new MPI communication scheme
87d061b
@hmenke
[SD] Rebase onto latest origin/python
6bbb4e8
@hmenke
[SD] Lubrication corrections and feature bitflags
4e706ab
@hmenke
Merge remote-tracking branch 'origin/python' into sd
1efccea 
# Conflicts:
#	maintainer/configs/nocheck-maxset.hpp
#	src/core/integrate.cpp
#	src/python/espressomd/thermostat.pxd
#	src/python/espressomd/thermostat.pyx
@hmenke
[SD] Adapt to new integrators
95aa778
@hmenke
[SD] Associate radius with particle type
a7682ba
@hmenke
[SD] Expose bit flags for feature selection
d06b989
@hmenke
[SD] Move implementation to libs/stokesian_dynamics
1e4c236
@hmenke
[SD] Test against reference data from Brady's paper
f036395
@hmenke
Formatting suggested by fix_style
75f147f
@RudolfWeeber
Copy link
Contributor

I think, this is a good addition to Espresso.

As long as it's marked WIP, just a few comments:

  • Collecting the particles via partCfg does not work in parallel, neither do the partiel property setters defined in particle_data.hpp. You need some MPI-gather and scatter based approach, instead.
  • Do I understand correctly that this requires CUDA for thrust, even with the CPU implementation?
  • Can you please add some doxygen to device_matrix.hpp and similar files?

@mkuron mkuron force-pushed the sd branch 3 times, most recently from 213f517 to 2086b34 Compare October 15, 2019 08:01
@RudolfWeeber RudolfWeeber self-assigned this Oct 21, 2019
@RudolfWeeber
Copy link
Contributor

My understanding of the current state of affairs is the following:
This PR still needs more work in some areas. These are:

  • more extensive testing
  • collection of particle pos/vel/omega and back-distribution of vel and omega that do not break the synchronized parallelization scheme
  • making the CPu solver independent from Cuda components (thrust)
  • docs

@mkuron explained in the ESpresso offline meeting that he aand @hmenke currently do not have the time to address these issues.

Currently, most student assistants are involved with teaching. We will try to find someone to work on this PR in the term break. Since the interface to the Espresso core is quite narrow, there should not be major merge conflicts.

jngrad added 27 commits June 15, 2020 19:28
@jngrad
Merge remote-tracking branch 'origin/cmake-deps-inversion' into sd
cd3232d
@jngrad
cmake: Dependency inversion
3726d54
@jngrad
Merge branch 'python' into sd
27333d4
@jngrad
Formatting
d5e0413
@jngrad
cmake: Fix operator precedence
b5cb4fd 
AND has no precedence over OR.
@jngrad
SD: Update documentation
9be4433
@jngrad
SD: Add missing license headers
8b54548
@jngrad
SD: Simplify interface code
fc496f9
@jngrad
SD: Add CMake version guard for CUDA 10
cad813a
@jngrad
Formatting
3092aae
@jngrad
Merge branch 'python' into sd
b47d368
@jngrad
SD: Test checkpointing mechanism
8c42877
@jngrad
testsuite: Fix MAX_NUM_PROC
5422567
@jngrad
SD: Split GPU and CPU test
e0a4a4e
@jngrad
SD: Disable GPU test
811f3de
@jngrad
SD: Commit missing files...
e768253
@jngrad
Merge branch 'python' into sd
23ff1cf
@jngrad
SD: Test CPU version on multiple OSse
7054743
@jngrad
SD: Integrate SD docs into the integrator and thermostat pages
ccbaa82
@jngrad
SD: Improve sample
ae909ca
@jngrad
SD: Test sample
f5bc594
@jngrad
SD: Accelerate tests
efb34cf
@jngrad
SD: Style fix
721c145
@jngrad
Formatting
7ef68a8
@jngrad
SD: Mark GPU implementation as experimental
543245a 
The GPU implementation is failing on ROCm 3.3 and produces the wrong
trajectories on a GeForce RTX 2080 with CUDA 9.1.
@jngrad
Fix build system
53b31bf
@jngrad
SD: Rename integrator and thermostat
4a352a9 
Disambiguate Stokesian Dynamics from Steepest Descent.
@jngrad jngrad mentioned this pull request Jul 3, 2020
Merged
@jngrad
Copy link
Member

jngrad commented Jul 20, 2020

The rebased SD was merged. Closing.

@jngrad jngrad closed this Jul 20, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@biermanncarl biermanncarl biermanncarl left review comments

Assignees

@RudolfWeeber RudolfWeeber

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
5 participants
@hmenke @RudolfWeeber @jngrad @biermanncarl @mkuron