Nonorthogonal with minimal differential operators MMS test. #285

mrhardman · 2025-01-24T09:49:01Z

Fixes to the Div_a_Grad_perp_nonorthog(a, f) operator (commit 3d11cb7, commit 2e5e827), merge with master, and new automatic tests integrated with ctest showing a correct implementation. Theses tests are hermes-3/tests/mms/orthogonal_test.py and hermes-3/tests/mms/nonorthogonal_test.py, and may be run interactively by, e.g.,

cd hermes-3/build-path/tests/mms/
python3 orthogonal_test.py

or

cd hermes-3/build-path/
ctest -I 12

These tests can be readily extended to test new nonorthogonal operators and the other existing operators in the Hermes-3 code base.

See /hermes-3/tests/mmsREADME.md for some documentation. Comments and requests appreciated.

Thanks to @dschwoerer for providing an initial MMS test to develop. Thanks to @johnomotani for numerous suggestions and bugfixes.

Below is the plot output from the tests, showing the error (blue), the expected (orange), and the fit of the error (green). Increase the range of $\Delta$ by increasing ntest in the test_input dictionary in hermes-3/tests/mms/orthogonal_test.py and hermes-3/tests/mms/nonorthogonal_test.py. Note that since the number of grid points $N_x = N_y = N_z$, the cost of the test rapidly increases with ntest.

FV::Div_a_Grad_perp(a, f) with orthogonal metric

Div_a_Grad_perp_nonorthog(a, f) with orthogonal metric

Div_a_Grad_perp_nonorthog(a, f) with nonorthogonal metric

A version of FV::Div_a_Grad_perp that includes all metric components, including g12 (X-Y non-orthogonal) and g13 (X-Z nonorthogonal) terms. Aim to test and then merge upstream into BOUT++

Testing, using operator that includes nonorthogonal mesh corrections

…ile as this causes problems when loading from xbout.

…operator tests.

…MMS test.

…re descriptive name.

… order with fit function, add comments.

… functions.

… x y z coordinates instead of R Z.

…sults yet provided.

…g sympy for maximum flexibility.

…loading.

…ement with try, except block.

… of convergence for orthogonal grid operator with a simple metric constant in x, y, z.

…erators implemented for orthogonal and nonorthogonal grids.

…st master version of xBOUT.

dschwoerer · 2025-01-24T13:13:44Z

Can you either change this PR to go into master or open a new one that goes into master:
https://github.com/bendudson/hermes-3/compare/bendudson:master...bendudson:nonorthogonal-divops-mms-test-minimal?expand=1

Like this it is hard to see the changes you introduced.

Unfortunately I cannot do that, as I am not a collaborator :-(

.pip_install_for_tests.sh

mrhardman · 2025-01-24T13:41:29Z

Changed base to master in response to @dschwoerer's review. Making master the base obscures slightly the changes made in div_ops.cxx. The changes to this with respect to #88 are simply commit 3d11cb7 and commit 2e5e827.

bendudson

Amazing work. Thanks @mrhardman !

CMakeLists.txt

dschwoerer · 2025-01-24T13:46:29Z

src/div_ops.cxx

+BoutReal BOUTMIN(const BoutReal& a, const BoutReal& b, const BoutReal& c,
+                 const BoutReal& d) {


Suggested change

BoutReal BOUTMIN(const BoutReal& a, const BoutReal& b, const BoutReal& c,

const BoutReal& d) {

BoutReal BOUTMIN(const BoutReal a, const BoutReal b, const BoutReal c,

const BoutReal d) {

This seems orthogonal to this PR -- can you please explain what the suggested change does here? I don't know for sure whether or not this line is covered by a test.

dschwoerer · 2025-01-24T13:47:32Z

src/div_ops.cxx

  BoutReal r1 = (a < b) ? a : b;
  BoutReal r2 = (c < d) ? c : d;
  return (r1 < r2) ? r1 : r2;


Which I think should be equivalent to

Suggested change

BoutReal r1 = (a < b) ? a : b;

BoutReal r2 = (c < d) ? c : d;

return (r1 < r2) ? r1 : r2;

return std::min({a, b, c, d});

This seems orthogonal to this PR -- can you please explain what the suggested change does here? I don't know for sure whether or not this line is covered by a test.

dschwoerer · 2025-01-27T10:02:52Z

src/div_ops.cxx

+    aup = a.yup();
+    adown = a.ydown();
+
+    mesh->communicate(fddx_xhigh);


Note that this does not work for fci.
for fci the right thing would be to do the above calculation also for the parallel slices.

Suggested change

mesh->communicate(fddx_xhigh);

if (f.isFci()) {

throw BoutException("This is not implemented for FCI.\n"

"The above calculation needs to be done for the parallel slices as well");

}

mesh->communicate(fddx_xhigh);

f.isFci() appears not be be a valid call in this branch, errors appear on compilation. Could you please make a PR into this PR with valid changes?

src/div_ops.cxx

dschwoerer · 2025-01-27T10:07:08Z

src/div_ops.cxx

+  for (int i = mesh->xstart - 1; i <= mesh->xend; i++) {
+    for (int j = mesh->ystart; j <= mesh->yend; j++) {
+      for (int k = 0; k < mesh->LocalNz; k++) {
+        // Calculate flux from i to i+1
+
+        const int kp = (k + 1) % mesh->LocalNz;
+        const int km = (k - 1 + mesh->LocalNz) % mesh->LocalNz;


same as above

Can you please be more specific with a code suggestion? I do not understand the code-base well enough to guess what you mean here.

David Bold correction making sure appropriate error message is present if FCI is used. Co-authored-by: David Bold <[email protected]>

Co-authored-by: David Bold <[email protected]>

…b.com/bendudson/hermes-3 into nonorthogonal-divops-mms-test-minimal

This reverts commit ca23fb9.

mrhardman · 2025-01-27T14:07:00Z

@dschwoerer Could you please make a PR with your suggestions into this PR? It looks like some of the functions that you expect to be available are not.

…t at the current time.

bendudson and others added 30 commits November 28, 2022 22:54

Div_a_Grad_perp_nonorthog function for testing

18ac63e

A version of FV::Div_a_Grad_perp that includes all metric components, including g12 (X-Y non-orthogonal) and g13 (X-Z nonorthogonal) terms. Aim to test and then merge upstream into BOUT++

Clang formatting

809ed4b

neutral_mixed: Use Div_a_Grad_perp_nonorthog

8156075

Testing, using operator that includes nonorthogonal mesh corrections

Merge branch 'master' into nonorthogonal

910a0c4

Attempt to port David Bold`s MMS test to Hermes-3.

aadc272

Read Rxy and Zxy from mesh file (hypnotoad), do not write 3D R Z to f…

8f847a9

…ile as this causes problems when loading from xbout.

Merge branch 'master' into divops-mms-test

b0f380f

Initial commit of files needed to port David Bold's MMS differential …

f43cc54

…operator tests.

Changes to makes mms.py script run.

fe91bd7

Reorganise and comment the mms.py test script.

cdb7be3

Reorganise and comment the mms.py test script, remove unused code.

fdc7d1e

Reorganise the mms.py test script, rationalise plotting.

0b6b066

Add script to generate circular closed-field-line equilibria for the …

427e76e

…MMS test.

Make test BOUT.inp file use a third circular equilibrium.

a05eeed

Restore variable path input functionality, rename main function to mo…

4685453

…re descriptive name.

Rename variables to clarify meaning.

73e5c49

Modify plot to show expected convergence order, calculate convergence…

fd9192d

… order with fit function, add comments.

Refactor plotting so that each test is a unique plot, experiment with…

12488c3

… functions.

Comment out Poisson bracket tests for now from main.cxx.

86aaa19

Make modifications to test the feasibility of MMS test in a slab with…

e020cd4

… x y z coordinates instead of R Z.

Reduce main.cxx for MMS test to the minimal code for a single operator

7fa43c7

Initial python script for running minimal MMS test, no plotting of re…

11ba89e

…sults yet provided.

Start adding i/o result testing functionality.

7f55ffe

Bare minimum script to compare analytical and numerical results, usin…

90677d8

…g sympy for maximum flexibility.

Change syntax of open_boutdataset() to use xBOUT branch fix-gridless-…

071850d

…loading.

Make sure g** are taken from a single source, accomodate trivial agre…

ef7aeeb

…ement with try, except block.

Cut down seemingly unused options in BOUT.input.template.

e674d04

Support the ability to deal with polynomial test functions.

4b5c896

Changes needed to get minimal MMS test to work, showing expected rate…

b828e7f

… of convergence for orthogonal grid operator with a simple metric constant in x, y, z.

Include C language to permit compiling of C dependencies.

a4a2da4

mrhardman added 9 commits January 23, 2025 14:53

Refactored MMS test to allow for different metrics to be tested on op…

8fcd604

…erators implemented for orthogonal and nonorthogonal grids.

Remove older revisions of test for clarity in ongoing development.

8aae139

Make sure to make root test directory.

101697e

Correct minor typos.

31e9af0

Include differential operator tests in CMakelists.txt correctly

17238a6

Make interactive plotting optional.

4698c47

Try to change .pip_install_for_tests.sh to make sure we have the late…

da10095

…st master version of xBOUT.

Some documentation for the tests.

62b9f4a

Make sure plots saved in appropriate directory.

607780a

mrhardman requested review from bendudson and mikekryjak January 24, 2025 09:49

mrhardman self-assigned this Jan 24, 2025

mrhardman mentioned this pull request Jan 24, 2025

Reduce main.cxx for MMS test to the minimal code for a single operator #283

Closed

dschwoerer reviewed Jan 24, 2025

View reviewed changes

.pip_install_for_tests.sh Outdated Show resolved Hide resolved

mrhardman changed the base branch from nonorthogonal to master January 24, 2025 13:41

mrhardman requested a review from dschwoerer January 24, 2025 13:42

bendudson approved these changes Jan 24, 2025

View reviewed changes

bendudson mentioned this pull request Jan 24, 2025

WIP: Operators for non-orthogonal meshes #88

Closed

dschwoerer reviewed Jan 27, 2025

View reviewed changes

mrhardman and others added 5 commits January 27, 2025 13:50

Remove unnecessary code for MMS tests in CMakelists.txt.

cd679c8

Update src/div_ops.cxx

ca23fb9

David Bold correction making sure appropriate error message is present if FCI is used. Co-authored-by: David Bold <[email protected]>

Update src/div_ops.cxx

dd95973

Co-authored-by: David Bold <[email protected]>

Merge branch 'nonorthogonal-divops-mms-test-minimal' of https://githu…

ab2831e

…b.com/bendudson/hermes-3 into nonorthogonal-divops-mms-test-minimal

Revert "Update src/div_ops.cxx "

6b11132

This reverts commit ca23fb9.

mrhardman added 3 commits January 28, 2025 08:49

Use David Bold suggestion to streamline .pip_install_for_tests.sh.

7bd1012

Update docs to make it clear that xBOUT needs to be installed from gi…

70a15e1

…t at the current time.

Note on debug flags for Spack notes.

05ef0a9

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Nonorthogonal with minimal differential operators MMS test. #285

Nonorthogonal with minimal differential operators MMS test. #285

mrhardman commented Jan 24, 2025 •

edited

Loading

dschwoerer commented Jan 24, 2025

mrhardman commented Jan 24, 2025 •

edited

Loading

bendudson left a comment

dschwoerer Jan 24, 2025

mrhardman Jan 27, 2025 •

edited

Loading

dschwoerer Jan 24, 2025

mrhardman Jan 27, 2025

dschwoerer Jan 27, 2025

mrhardman Jan 27, 2025

dschwoerer Jan 27, 2025

mrhardman Jan 27, 2025

mrhardman commented Jan 27, 2025

		BoutReal BOUTMIN(const BoutReal& a, const BoutReal& b, const BoutReal& c,
		const BoutReal& d) {

-    mesh->communicate(fddx_xhigh);
+    if (f.isFci()) {
+        throw BoutException("This is not implemented for FCI.\n"
+        "The above calculation needs to be done for the parallel slices as well");
+    }
+    mesh->communicate(fddx_xhigh);

Nonorthogonal with minimal differential operators MMS test. #285

Are you sure you want to change the base?

Nonorthogonal with minimal differential operators MMS test. #285

Conversation

mrhardman commented Jan 24, 2025 • edited Loading

dschwoerer commented Jan 24, 2025

mrhardman commented Jan 24, 2025 • edited Loading

bendudson left a comment

Choose a reason for hiding this comment

dschwoerer Jan 24, 2025

Choose a reason for hiding this comment

mrhardman Jan 27, 2025 • edited Loading

Choose a reason for hiding this comment

dschwoerer Jan 24, 2025

Choose a reason for hiding this comment

mrhardman Jan 27, 2025

Choose a reason for hiding this comment

dschwoerer Jan 27, 2025

Choose a reason for hiding this comment

mrhardman Jan 27, 2025

Choose a reason for hiding this comment

dschwoerer Jan 27, 2025

Choose a reason for hiding this comment

mrhardman Jan 27, 2025

Choose a reason for hiding this comment

mrhardman commented Jan 27, 2025

mrhardman commented Jan 24, 2025 •

edited

Loading

mrhardman commented Jan 24, 2025 •

edited

Loading

mrhardman Jan 27, 2025 •

edited

Loading