Skip to content

Adding implicit time integrators for general use in Spheral #339

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
Merged
merged 103 commits into from
Apr 24, 2025
Merged

Adding implicit time integrators for general use in Spheral #339

merged 103 commits into from
Apr 24, 2025

Conversation

jmikeowen
Copy link
Collaborator

@jmikeowen jmikeowen commented Apr 9, 2025 •
edited
Loading

Summary

  • This PR is a feature
  • It does the following:
    • Adds new implicit time integration options to Spheral.
      • This requires additions to the Physics interface: Physics::dtImplicit and Physics::maxResidual. Packages that we want to use with implicit time advancement must provide these methods.
    • Also added Sundials as an optional TPL for Spheral, and a Spheral interface to the KINSOL solver in Sundials.
    • There are several bugfixes in this commit as well, such as fixing the deviatoric evolution in 1D/2D for FSISPH.

ToDo :

  • Annotate RELEASE_NOTES.md with notable changes.
  • Create LLNLSpheral PR pointing at this branch. (PR#132)
  • LLNLSpheral PR has passed all tests.

@jmikeowen
Introducing a new ImplicitIntegrator base class for implicit time adv…
1bcf6bd 
…ancement. Also cleaned up some stuff in the Integrator package.
@jmikeowen
Fixing a few references to the now renamed ForwardEuler integrator
b7d0611
@jmikeowen
Initial whack at adding a rudimentary implicit integrator
d51fe67
@jmikeowen
Adding more implementation for BackwardEuler implicit integrator. Also
5724865 
cleaned up a few details (int->size_t for sizes of arrays handed back
a few places, and DataBase Python interface for getting NodeList
arrays), and these had wide ranging impact in terms of touching a lot
of files.  No substantive changes to most of these files however.
@jmikeowen
Merge branch 'develop' into feature/ImplicitTimeIntegration
4398bff
@jmikeowen
Fixes for clang insanity with structured bindings
a125f8a
@jmikeowen
Adding more to our residuals interface in Physics, which is a minor
7c19a1d 
change. In the process I removed some inline headers which resulted in
including fewer headers across other classes.  This cascaded into
touching a lot of files to include those where necessary, which is
most of the files touched in this commit. It's still worthwhile
getting unnecessary includes out of headers so I went with it.
@jmikeowen
Adding residual calculation for smoothing scales
a78049d
@jmikeowen
Interim work on implicit integration
687f190
@jmikeowen
Making copyState and assingment with State objects fail when we hit u…
f458deb 
…nknown types
@jmikeowen
Simplifying the Integrator interfaces a bit, and more work on the Bac…
af89a77 
…kwardEuler implicit integrator
@jmikeowen
Cleaned up some unitialized variables pointed out by valgrind
2dc49ef
@jmikeowen
Fixed graphics on this test
78f2333
@jmikeowen
Got around our SEGV's, but implicit integration results not good yet
1e33a8a
@jmikeowen
Merge branch 'develop' into feature/ImplicitTimeIntegration
eda9d40
@jmikeowen
Compiling but untested
c75f6f1
@jmikeowen
Using DeprecationWarning
8d28d05
@jmikeowen
First stage of adding serialization logic to state and policies for
b5dc7a2 
use with implicit integration and packing the solution vector. Our
logic for picking the independent state to be packed though needs to
be improved.
@jmikeowen
Fiddling with how we serialize the State to vector<double>
f3654d5
@jmikeowen
Serializing state and derivs to vector<double> for implicit integration
39a1ef6
@jmikeowen
Adding option for State::update to only update dependent state
c70dc7e
@jmikeowen
Checkpoint (moving BackwardEuler to LLNLSpheral for now)
a274832
@jmikeowen
Different approach to dependent/independent state update for implicit
2ee7e17
@jmikeowen
Python binding update
fb60831
@jmikeowen
Adding Sundials as a TPL
e2bfa2a
@jmikeowen
Got KINSOL from Sundials working in C++ wrappers
b054ab7
@jmikeowen
Fixed overloading of virtual operator() <->__call__ mapping in
024f41b 
PYB11Generator, and updated our use of this in Spheral accordingly
@jmikeowen
Adding KINSOL unit test
01e8749
@jmikeowen
Checkpoint testing backward Euler
cdecd35
@jmikeowen
Exposing access to KINSOL tolerance parameters. Also using input data
12e4354 
memory rather than copying.
@jmikeowen jmikeowen marked this pull request as draft April 10, 2025 00:01
@jmikeowen jmikeowen marked this pull request as ready for review April 16, 2025 21:24
@jmikeowen
Jason pointed out that we really only want to do our explicit zeroing
2fb58e2 
of the deviatoric stress trace in 3D -- fixed
ldowen
ldowen reviewed Apr 23, 2025
View reviewed changes
jmpearl
jmpearl previously approved these changes Apr 23, 2025
View reviewed changes
Copy link
Collaborator

@jmpearl jmpearl left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Everything I looked at looks good to me, I got through ~ 1/3 of the files focusing on the hydros, and update policies

@jmikeowen jmikeowen merged commit 8103e15 into develop Apr 24, 2025
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@mdavis36 mdavis36 mdavis36 approved these changes

@ldowen ldowen ldowen approved these changes

@jmpearl jmpearl jmpearl left review comments

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@jmikeowen @mdavis36 @ldowen @jmpearl