Make get_cell_range_category() more consistent with get_cell_category() #14015
Conversation
| * maximum catergory of any cell batch. Only in the hp case, it is | ||
| * guaranteed that all cells and as a consquence all cell batches have | ||
| * the same categroy. |
There was a problem hiding this comment.
There are a few typos, plus maybe something that can be clarified
| * maximum catergory of any cell batch. Only in the hp case, it is | |
| * guaranteed that all cells and as a consquence all cell batches have | |
| * the same categroy. | |
| * maximum category of any cell batch. In the hp case, it is | |
| * guaranteed that all cells and as a consequence all cell batches in a range have | |
| * the same category. Otherwise, there may be different categories in different | |
| * cell batches. |
| @@ -1912,6 +1917,11 @@ class MatrixFree : public Subscriptor | |||
| * run between the given values in the field | |||
| * AdditionalData::cell_vectorization_category for non-hp-DoFHandler types | |||
| * and return the active FE index in the hp-adaptive case. | |||
| * | |||
| * @note In the non-hp-DoFHandler case, a categroy of a cell batch is given | |||
There was a problem hiding this comment.
| * @note In the non-hp-DoFHandler case, a categroy of a cell batch is given | |
| * @note In the non-hp case, a category of a cell batch is given |
| * @note In the non-hp-DoFHandler case, a categroy of a cell batch is given | ||
| * as the maximum category of any of its cell. In the hp case or the case that | ||
| * MatrixFree::AdditionalData::cell_vectorization_categories_strict was | ||
| * enabled, it is guaranteed that all cells have the same categroy. |
There was a problem hiding this comment.
You should run a spellchecker on this particular word. 😉
| * enabled, it is guaranteed that all cells have the same categroy. | |
| * enabled, it is guaranteed that all cells have the same category. |
peterrum
force-pushed
the
get_cell_range_category_max
branch
from
bcd2903
to
f2d0e65
Compare
| * AdditionalData::cell_vectorization_category for non-hp-DoFHandler types | ||
| * AdditionalData::cell_vectorization_category for the non-hp case |
There was a problem hiding this comment.
@kronbichler There have been other places with references to non-hp-DoFHandler ;)
|
@kronbichler I have added a test! It is a modified version of the test in #14013. |
| { | ||
| const auto category = cell_vectorization_category | ||
| [matrix_free.get_cell_iterator(cell, v)->active_cell_index()]; | ||
| std::cout << category << " "; |
There was a problem hiding this comment.
Is this output on purpose? It won't be compared, and one will only see it when manually running the test (which might be ok).
There was a problem hiding this comment.
I would keep it this way. I introduced these for visualization purposes. It looks like this:
0 0 0 0 0 0 0 1
1 1 1 1 1 1 2 2
2 2 2 2 2 3 3 3
3 3 3 3 4 4 4 4
4 4 4 5 5 5 5 5
5 5 6 6 6 6 6 6
6 7 7 7 7 7 7 7
8 8 8 8 8 8 8 9
0 0 0 0 0 0 0
1 1 1 1 1 1 1
2 2 2 2 2 2 2
3 3 3 3 3 3 3
4 4 4 4 4 4 4
5 5 5 5 5 5 5
6 6 6 6 6 6 6
7 7 7 7 7 7 7
8 8 8 8 8 8 8
9
0 0 0 0 0 0 0
1 1 1 1 1 1 1
2 2 2 2 2 2 2
3 3 3 3 3 3 3
4 4 4 4 4 4 4
5 5 5 5 5 5 5
6 6 6 6 6 6 6
7 7 7 7 7 7 7
8 8 8 8 8 8 8
9
In the test, I am only checking the properties, since outputting such an output would require that we create an output for each vectorization type.
There was a problem hiding this comment.
I agree that we should not print this with vectorization. Maybe add a short comment in the code that explains why we have the cout in the test, to guide the reader of test on how to use it?
peterrum
force-pushed
the
get_cell_range_category_max
branch
from
f2d0e65
to
291768b
Compare
| // note: here and below we are using std::cout to output | ||
| // information that is useful for debugging; however, since the | ||
| // output depends on the hardware and might change if we | ||
| // internally modify the way we loop over cells, we don't write | ||
| // the data to the log but instead only check that the | ||
| // properties that are described in the documentation are | ||
| // fulfilled |
There was a problem hiding this comment.
@kronbichler What do you think?
…akeover [9.4] Take over #14015: Make get_cell_range_category() more consistent with get_cell_category()
…_max Make get_cell_range_category() more consistent with get_cell_category()
If possible, it would be good to integrate this also into the 9.4 release. The function is newly added in the last months.