Convert to MATSELL or its subclass when calling MatSetType()

This simplifies user code so that users do not need to convert the matrix themselves.

Add SELLHIP

- The HIP kernels are converted directly from their CUDA version
- AMD GPUs and NVIDIA GPUs use different warp sizes. We set the warp size to 64 by default for AMD GPUs to faciliate comp

Add SELLHIP

- The HIP kernels are converted directly from their CUDA version
- AMD GPUs and NVIDIA GPUs use different warp sizes. We set the warp size to 64 by default for AMD GPUs to faciliate compile-time code optimization

show more ...

Docs: fix missing or extra dash in right-hand side

Merge branch 'jolivet/feature-consolidate-issymmetric' into 'main'

Mat: consolidate MatIs[Symmetric|Hermitian]() implementations

See merge request petsc/petsc!7380

MatView: relax requirement on same communicator between matrix and viewer

Mat: consolidate MatIs[Symmetric|Hermitian]() implementations

MATDENSE: support NULL in MatSetValues

Remove carriage return in Petsc(Assert|Check)()

git grep -l -E "Petsc(Assert|Check)\(.*\\\n\"[,)]" | xargs sed -r -i'' 's#Petsc(Assert|Check)\((.*)\\n"([,)])#Petsc\1\(\2\"\3#g'

Suggested-by: Stefan

Remove carriage return in Petsc(Assert|Check)()

git grep -l -E "Petsc(Assert|Check)\(.*\\\n\"[,)]" | xargs sed -r -i'' 's#Petsc(Assert|Check)\((.*)\\n"([,)])#Petsc\1\(\2\"\3#g'

Suggested-by: Stefano Zampini

show more ...

Remove multiple parentheses and extra semicolon

git grep -l -E "\(\([a-zA-Z0-9_]+\)\(\*[a-zA-Z0-9_]+\)[, )]" | xargs sed -r -i'' 's#\(\(([a-zA-Z0-9_]+)\)\(\*([a-zA-Z0-9_]+)\)([, )])#\(\(\1\)\*\2\3#g

Remove multiple parentheses and extra semicolon

git grep -l -E "\(\([a-zA-Z0-9_]+\)\(\*[a-zA-Z0-9_]+\)[, )]" | xargs sed -r -i'' 's#\(\(([a-zA-Z0-9_]+)\)\(\*([a-zA-Z0-9_]+)\)([, )])#\(\(\1\)\*\2\3#g'
git grep -l -E "Petsc(.)*\(\(\*[a-zA-Z0-9_]*\)," | xargs sed -r -i'' 's#Petsc([a-zA-Z0-9_]*)\(\(\*([a-zA-Z0-9_]*)\), #Petsc\1(*\2, #g'
git grep -l -E "([\!\&\~\*\(]|\)\)|\([^,\*\(]+\**\))\(([a-zA-Z0-9_]+((\.|->)[a-zA-Z0-9_]+|\[[a-zA-Z0-9_ \%\+\*\-]+\])+)\)" | xargs sed -r -i'' 's#([\!\&\~\*\(]|\)\)|\([^,\*\(]+\**\))\(([a-zA-Z0-9_]+((\.|->)[a-zA-Z0-9_]+|\[[a-zA-Z0-9_ \%\+\*\-]+\])+)\)#\1\2#g'

show more ...

Merge branch 'barry/2024-02-23/doc-trivial-mat' into 'main'

Bug fix, calling MatXXXSetPreallocation on MPI matrices after MatSetOption() loses the option value

See merge request petsc/petsc!7319

Typo

Trivial fix to MatSetValuesBlocked() manual page, spelling and spacing

Keep any (near) null space attached to submatrices in PCFIELDSPLIT even if the nonzero structure of the original matrix changes.

Also applied to PCBJACOBI and PCASM

Reported-by: "Jeremy Theler (Ext

Keep any (near) null space attached to submatrices in PCFIELDSPLIT even if the nonzero structure of the original matrix changes.

Also applied to PCBJACOBI and PCASM

Reported-by: "Jeremy Theler (External)" <jeremy.theler-ext@ansys.com>

The long-term way to handle this is to have MatGetSubMatr*() reuse the same submatrix objects even if the nonzero structure changes. This would just require a bit of recoding in multiple places.

show more ...

Fix words appearing twice in a row for no good reason

fix error message when MatProductSymbolic() is not supported for the matrix types
Reported-by: Hana Honnerová <hhornik@kma.zcu.cz>
add 'bool setfromoptions_called' in Mat_Product for checking if MatP

fix error message when MatProductSymbolic() is not supported for the matrix types
Reported-by: Hana Honnerová <hhornik@kma.zcu.cz>
add 'bool setfromoptions_called' in Mat_Product for checking if MatProductSetFromOptions() is called

show more ...

Merge branch 'barry/2024-02-09/fix-matzerorowscolumns' into 'main'

Fix bugs in MatZeroRowsColumns() both in docs and handling of no-nonzero on diagonal

See merge request petsc/petsc!7279

Fix bugs in MatZeroRowsColumns() both in docs and handling of no-nonzero on diagonal

Bug 1 - documentation
The documentation statement "Unlike `MatZeroRows()` this does not change the nonzero struct

Fix bugs in MatZeroRowsColumns() both in docs and handling of no-nonzero on diagonal

Bug 1 - documentation
The documentation statement "Unlike `MatZeroRows()` this does not change the nonzero structure of the matrix, it merely zeros those entries in the matrix." appears to be incorrect for matrices missing diagonal entries since new nonzeros are added (to fill in the diagonal). The documentation should really say

"Unlike `MatZeroRows()` this routine cannot remove the zeroed entries from the nonzero structure of the matrix; in other words setting the option `MAT_KEEP_NONZERO_PATTERN to PETSC_FALSE has no effect on this routine.

Bug 2 -
The short circuit if (!((Mat_SeqAIJ *)(l->A->data))->keepnonzeropattern) is wrong because that flag is meaningless for this operation. The check should be changed (I think) to nonew instead of keepnonzeropattern.

Reported-by: Jeremy Theler <jeremy.theler-ext@ansys.com>

show more ...

Merge remote-tracking branch 'origin/release'

MatProduct_Private(): code factorization to ensure proper error checking

Mat: GetAbsSum (l1) row norms; PCJacobi: add scaled L1 diagonal approximation

Merge branch 'barry/2024-01-19/typedef-snes-functions' into 'main'

add typedef for functions passed to SNESSetFunction() etc to match approach in TS

See merge request petsc/petsc!7206

Change the use of the _Fn suffix to indicate a typedef of a function to just Fn

As determined on a vote on GitLab