Changes for FCI with hermes-3 #3079

dschwoerer · 2025-03-03T10:28:24Z

Opened a PR to get the documentation build.

The PR needs likely be split up, in e.g. #2651 #2887 #3004 and once they are merged likely more ...

mesh may not be initialised or be a different mesh

Use the new track feature (better name required) to dump the different components of the ddt() as well as the residuum for the evolved fields.

This keeps track of all the changes done to the field and stores them to a OptionsObject.

This reverts commit 8485353. The parallel metric components are loaded, thus we can take meaningful derivatives in y-direction.

Directly iterate over the points

github-actions

clang-tidy made some suggestions

There were too many comments to post at once. Showing the first 25 out of 381. Check the log or trigger a new build to see more.

github-actions · 2025-03-03T10:42:20Z

externalpackages/PVODE/include/pvode/band.h

@@ -138,7 +138,7 @@ typedef struct bandmat_type {
 *                                                                *
 ******************************************************************/

-#define PVODE_BAND_ELEM(A,i,j) ((A->data)[j][i-j+(A->smu)])
+#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])


warning: function-like macro 'PVODE_BAND_ELEM' used; consider a 'constexpr' template function [cppcoreguidelines-macro-usage]

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)]) ^

github-actions · 2025-03-03T10:42:20Z

externalpackages/PVODE/include/pvode/band.h

@@ -138,7 +138,7 @@
 *                                                                *
 ******************************************************************/

-#define PVODE_BAND_ELEM(A,i,j) ((A->data)[j][i-j+(A->smu)])
+#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])


warning: macro argument should be enclosed in parentheses [bugprone-macro-parentheses]

Suggested change

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])

#define PVODE_BAND_ELEM(A, i, j) (((A)->data)[j][i - j + (A->smu)])

github-actions · 2025-03-03T10:42:20Z

externalpackages/PVODE/include/pvode/band.h

@@ -138,7 +138,7 @@
 *                                                                *
 ******************************************************************/

-#define PVODE_BAND_ELEM(A,i,j) ((A->data)[j][i-j+(A->smu)])
+#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])


warning: macro argument should be enclosed in parentheses [bugprone-macro-parentheses]

Suggested change

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][(i) - j + (A->smu)])

github-actions · 2025-03-03T10:42:20Z

externalpackages/PVODE/include/pvode/band.h

@@ -138,7 +138,7 @@
 *                                                                *
 ******************************************************************/

-#define PVODE_BAND_ELEM(A,i,j) ((A->data)[j][i-j+(A->smu)])
+#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])


warning: macro argument should be enclosed in parentheses [bugprone-macro-parentheses]

Suggested change

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - (j) + (A->smu)])

github-actions · 2025-03-03T10:42:20Z

externalpackages/PVODE/include/pvode/band.h

@@ -138,7 +138,7 @@
 *                                                                *
 ******************************************************************/

-#define PVODE_BAND_ELEM(A,i,j) ((A->data)[j][i-j+(A->smu)])
+#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])


warning: macro argument should be enclosed in parentheses [bugprone-macro-parentheses]

Suggested change

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])

#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + ((A)->smu)])

github-actions · 2025-03-03T10:42:22Z

include/bout/boundary_iterator.hxx

+  int abs_offset() const { return 1; }
+
+#if BOUT_USE_METRIC_3D == 0
+  BoutReal& ynext(Field2D& f) const { return f[ind().yp(by).xp(bx)]; }


warning: no header providing "Field2D" is directly included [misc-include-cleaner]

include/bout/boundary_iterator.hxx:2:

- #include "bout/mesh.hxx" + #include "bout/field2d.hxx" + #include "bout/mesh.hxx"

github-actions · 2025-03-03T10:42:22Z

include/bout/boundary_iterator.hxx

+  const BoutReal& yprev(const Field2D& f) const { return f[ind().yp(-by).xp(-bx)]; }
+#endif
+
+  const int dir;


warning: member 'dir' of type 'const int' is const qualified [cppcoreguidelines-avoid-const-or-ref-data-members]

const int dir; ^

github-actions · 2025-03-03T10:42:22Z

include/bout/boundary_iterator.hxx

+  const BoutReal& yprev(const Field2D& f) const { return f[ind().yp(-by).xp(-bx)]; }
+#endif
+
+  const int dir;


warning: member variable 'dir' has public visibility [cppcoreguidelines-non-private-member-variables-in-classes]

const int dir; ^

github-actions · 2025-03-03T10:42:22Z

include/bout/boundary_iterator.hxx

+  const int dir;
+
+protected:
+  int z{0};


warning: member variable 'z' has protected visibility [cppcoreguidelines-non-private-member-variables-in-classes]

int z{0}; ^

github-actions · 2025-03-03T10:42:23Z

include/bout/boundary_iterator.hxx

+
+protected:
+  int z{0};
+  int x;


warning: member variable 'x' has protected visibility [cppcoreguidelines-non-private-member-variables-in-classes]

int x; ^

This avoids errors in the MMS tests

This partially reverts 0bcc047 It seems the achieved accuracy depends on some factors that are not well controlled.

Likely this is not needed.

…ci-auto-with-debug-higher-order

That might format more code at once, but should avoid a CI loop.

bendudson · 2025-03-15T18:33:59Z

src/field/gen_fieldops.jinja

@@ -12,10 +12,41 @@
    {% if lhs == rhs == "Field3D" %}
      {{out.name}}.setRegion({{lhs.name}}.getMesh()->getCommonRegion({{lhs.name}}.getRegionID(),
                                                                     {{rhs.name}}.getRegionID()));
+#if BOUT_USE_FCI_AUTOMAGIC


I think BOUT_USE_FCI_AUTOMAGIC may be useful to get something working, but I don't think it's a good way to go: We will have code that either breaks when this is turned off, or performs too many operations when it is turned on. I think the options are:

a) Have two functions / operators: One that discards yup/ydown, and the other that keeps them.

b) Always operate on yup/down if they are present, but add a method to fields e.g withoutParallelSlices() that returns a copy of the field without the yup/down slices. That could be used e.g in assignment

Field3D f = g; // f has copies of g.yup and g.ydown Field3D f = g.withoutParallelSlices(); // f has no parallel slices

I suspect that operating on two fields, one with slices and one without, should be an error: Either they should both have slices, or neither have.

…igher-order

github-actions

clang-tidy made some suggestions

There were too many comments to post at once. Showing the first 25 out of 417. Check the log or trigger a new build to see more.

github-actions · 2025-07-21T08:24:21Z

include/bout/boundary_iterator.hxx

+    return (f[ind()] * 3 - yprev(f)) * 0.5;
+  }
+
+  BoutReal extrapolate_next_o2(const Field3D& f) const { return 2 * f[ind()] - yprev(f); }


warning: '*' has higher precedence than '-'; add parentheses to explicitly specify the order of operations [readability-math-missing-parentheses]

Suggested change

BoutReal extrapolate_next_o2(const Field3D& f) const { return 2 * f[ind()] - yprev(f); }

BoutReal extrapolate_next_o2(const Field3D& f) const { return (2 * f[ind()]) - yprev(f); }

github-actions · 2025-07-21T08:24:22Z

include/bout/boundary_iterator.hxx

+
+  BoutReal
+  extrapolate_next_o2(const std::function<BoutReal(int yoffset, Ind3D ind)>& f) const {
+    return 2 * f(0, ind()) - f(0, ind().yp(-by).xp(-bx));


warning: '*' has higher precedence than '-'; add parentheses to explicitly specify the order of operations [readability-math-missing-parentheses]

Suggested change

return 2 * f(0, ind()) - f(0, ind().yp(-by).xp(-bx));

return (2 * f(0, ind())) - f(0, ind().yp(-by).xp(-bx));

github-actions · 2025-07-21T08:24:22Z

include/bout/boundary_iterator.hxx

+protected:
+  int z{0};
+  int x;
+  int y;


warning: member variable 'y' has protected visibility [cppcoreguidelines-non-private-member-variables-in-classes]

int y; ^

github-actions · 2025-07-21T08:24:22Z

include/bout/boundary_iterator.hxx

+  int z{0};
+  int x;
+  int y;
+  const int bx;


warning: member 'bx' of type 'const int' is const qualified [cppcoreguidelines-avoid-const-or-ref-data-members]

const int bx; ^

github-actions · 2025-07-21T08:24:22Z

include/bout/boundary_iterator.hxx

+  int z{0};
+  int x;
+  int y;
+  const int bx;


warning: member variable 'bx' has protected visibility [cppcoreguidelines-non-private-member-variables-in-classes]

const int bx; ^

github-actions · 2025-07-21T08:24:24Z

include/bout/field.hxx

  return result;
 }

 #undef FIELD_FUNC

+template <typename T, typename = bout::utils::EnableIfField<T>, class... Types>
+inline void setName(T& f, const std::string& name, Types... args) {
+#if BOUT_USE_TRACK


warning: no header providing "BOUT_USE_TRACK" is directly included [misc-include-cleaner]

{ ^

github-actions · 2025-07-21T08:24:24Z

include/bout/field.hxx

+}
+
+template <typename T, typename = bout::utils::EnableIfField<T>, class... Types>
+inline T setName(T&& f, const std::string& name, Types... args) {


warning: forwarding reference parameter 'f' is never forwarded inside the function body [cppcoreguidelines-missing-std-forward]

> ^

github-actions · 2025-07-21T08:24:24Z

include/bout/field2d.hxx

@@ -66,7 +66,8 @@ public:
   */
  Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE,


warning: function 'Field2D::Field2D' has a definition with different parameter names [readability-inconsistent-declaration-parameter-name]

Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE, ^

Additional context

src/field/field2d.cxx:50: the definition seen here

Field2D::Field2D(Mesh* localmesh, CELL_LOC location_in, DirectionTypes directions_in, ^

include/bout/field2d.hxx:66: differing parameters are named here: ('region'), in definition: ('UNUSED_regionID')

Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE, ^

github-actions · 2025-07-21T08:24:24Z

include/bout/field2d.hxx

@@ -66,7 +66,8 @@
   */
  Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE,
          DirectionTypes directions_in = {YDirectionType::Standard,
-                                          ZDirectionType::Average});
+                                          ZDirectionType::Average},


warning: no header providing "ZDirectionType" is directly included [misc-include-cleaner]

ZDirectionType::Average}, ^

github-actions · 2025-07-21T08:24:24Z

include/bout/field2d.hxx

@@ -66,7 +66,8 @@
   */
  Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE,
          DirectionTypes directions_in = {YDirectionType::Standard,
-                                          ZDirectionType::Average});
+                                          ZDirectionType::Average},
+          std::optional<size_t> region = {});


warning: no header providing "size_t" is directly included [misc-include-cleaner]

include/bout/field2d.hxx:26:

- class Field2D; + #include <cstddef> + class Field2D;

dschwoerer and others added 30 commits February 4, 2024 20:27

Apply black changes

4566405

CI: install wget

4935742

Merge remote-tracking branch 'origin/next' into fci-splitting

01a436c

Apply clang-format changes

147a872

Add asserts to serial methods to avoid using in parallel

cee68f2

Apply clang-format changes

15aaa05

Merge branch 'par-bc-cleanup' into fci-splitting

eba722b

Remove accidentially added files

30ea7e3

Merge remote-tracking branch 'origin/next' into fci-splitting

66ff716

Use parallel_neumann as BC

10d320f

fix usage of f-string

4b05708

Use localmesh

60224e4

mesh may not be initialised or be a different mesh

add PETSc requirement for MPI test

608bb5d

Update header location

88741c5

More const correctness

09f609b

Merge branch 'origin/next' into fci-splitting

d46efba

Add some asserts for non parallelised XZ interpolation

1fb921e

enable openmp for sundials if it is enabled for BOUT++

4a79fb4

Add required interfaces

b5bd5f0

Add maskFromRegion

6819718

Dump field before and after rhs evaluation for debugging

fee27c9

Add setName function

96da2e9

Set div_par and grad_par names

e56981c

Dump debug file if PVODE fails

6b2c132

Use the new track feature (better name required) to dump the different components of the ddt() as well as the residuum for the evolved fields.

Add tracking to Field3D

df2d661

This keeps track of all the changes done to the field and stores them to a OptionsObject.

Add tracking to Field3D

263f9fe

This keeps track of all the changes done to the field and stores them to a OptionsObject.

cvode: Add option to use Adams Moulton solver instead of BDF

bac4ca9

Expose more pvode option to user

708bdcb

Fix bad cherry-pick

d88b454

Update to new API

023bc41

dschwoerer added 5 commits February 24, 2025 13:41

Allow to overwrite MYG with options.

f6106a2

Revert "Disable metric components that require y-derivatives for fci"

8665a67

This reverts commit 8485353. The parallel metric components are loaded, thus we can take meaningful derivatives in y-direction.

Add iter_pnts function

ff71182

Directly iterate over the points

Add example on how to replace RangeIterator with YBoundary

d3bc5cc

Move documentation to sphinx

8337531

github-actions bot reviewed Mar 3, 2025

View reviewed changes

dschwoerer added 15 commits March 3, 2025 13:40

Only read MYG if it set or mesh:MYG is not set

fcc3af6

This avoids errors in the MMS tests

Do not set MYG/MXG if it is not needed

f925c94

Do not set MYG/MXG if it is not needed

51e7b58

Fix iter_pnts

1c8fb47

Add more documentation on YBoundary

bfaf986

Ensure the field has parallel slices

f4acdb0

Lower check level

23f5992

Loosen tolereances again

d05e733

This partially reverts 0bcc047 It seems the achieved accuracy depends on some factors that are not well controlled.

Remove broken code

35206d4

Likely this is not needed.

Merge branch 'next' of https://github.com/boutproject/BOUT-dev into f…

261442e

…ci-auto-with-debug-higher-order

CI: Avoid issues with special characters

df490b9

CI: run git-clang-format until there are no more changes

eef32f9

That might format more code at once, but should avoid a CI loop.

CI: use one line

9fd76bf

CI: stage before we run git-clang-format again

ee9dc99

fix bad merge

be5e285

dschwoerer force-pushed the fci-auto-with-debug-higher-order branch from b1ee735 to be5e285 Compare March 14, 2025 09:31

dschwoerer and others added 4 commits March 14, 2025 09:31

Apply clang-format changes

092c578

Add option to copy parallel slices at setBoundaryTo

2fb1b34

Print timings for petsc solver in test

7a1d061

Add Field2D::splitParallelSlices() for writing FCI aware code

a6be0ab

bendudson reviewed Mar 15, 2025

View reviewed changes

dschwoerer added 2 commits July 17, 2025 10:19

Merge remote-tracking branch 'origin/next' into fci-auto-with-debug-h…

58811c8

…igher-order

Merge remote-tracking branch 'origin/next' into fci-auto-with-debug-h…

6e19edb

…igher-order

github-actions bot reviewed Jul 21, 2025

View reviewed changes

	#define PVODE_BAND_ELEM(A, i, j) ((A->data)[j][i - j + (A->smu)])
	#define PVODE_BAND_ELEM(A, i, j) (((A)->data)[j][i - j + (A->smu)])

	BoutReal extrapolate_next_o2(const Field3D& f) const { return 2 * f[ind()] - yprev(f); }
	BoutReal extrapolate_next_o2(const Field3D& f) const { return (2 * f[ind()]) - yprev(f); }

	return 2 * f(0, ind()) - f(0, ind().yp(-by).xp(-bx));
	return (2 * f(0, ind())) - f(0, ind().yp(-by).xp(-bx));

		@@ -66,7 +66,8 @@ public:
		*/
		Field2D(Mesh* localmesh = nullptr, CELL_LOC location_in = CELL_CENTRE,

Changes for FCI with hermes-3 #3079

Are you sure you want to change the base?

Changes for FCI with hermes-3 #3079

Uh oh!

Conversation

dschwoerer commented Mar 3, 2025

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Mar 3, 2025

Choose a reason for hiding this comment

Uh oh!

bendudson Mar 15, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!