MatSeqAIJMKL: warnings --with-scalar-type=complex

Merge branch 'stefanozampini/gamg-gpu-setup' into 'master'

MatMat operations with CUSPARSE

See merge request petsc/petsc!3391

MatAIJMKL: enable parallel PtaP

Wondering if it ever worked?

Minor

Merge remote-tracking branch 'origin/release'

Merge branch 'jose/fix-aijmkl-complex' into 'release'

AIJMKL: cannot print a scalar in case of complex

See merge request petsc/petsc!3494

AIJMKL: cannot print a scalar in case of complex

Merge branch 'barry/2020-08-16/fix-source-formatting' into 'master'

Turn on checkbadSource test to generate an error when found; fix all source code that causes errors

See merge request petsc/petsc

Merge branch 'barry/2020-08-16/fix-source-formatting' into 'master'

Turn on checkbadSource test to generate an error when found; fix all source code that causes errors

See merge request petsc/petsc!3064

show more ...

Turn on checkbadSource test to generate an error when found; fix all source code that causes errors

Commit-type: portability-fix, testing-fix, style-fix, feature, maintainability
/spend 1.5h

Merge branch 'rmills/aijmkl-mark-not-optimized-after-sp2m' into 'master'

Mark MKL matrix handle as not optimized after MatProduct operations.

See merge request petsc/petsc!3030

Mark MKL matrix handle as not optimized after MatProduct operations.

I had previously incorrectly assumed that the matrix optimization step is
carried out by MKL at the end of a matrix-matrix produc

Mark MKL matrix handle as not optimized after MatProduct operations.

I had previously incorrectly assumed that the matrix optimization step is
carried out by MKL at the end of a matrix-matrix product operation, but
a conversation with the MKL team confirms that this is not the case.
This did not cause a correctness issue, but might have led to reduced
performance if the matrix was then used for repeated operations that could
benefit from the optimization step.

Thanks-to: Maria Zhukova <maria.zhukova@intel.com>

show more ...

Merge branch 'rmills/feature-aijmkl-matproduct' into 'master'

Rmills/feature aijmkl matproduct

See merge request petsc/petsc!2841

Eliminate trailing whitespaces in aijmkl.c.

Thanks-to: Satish Balay <balay@mcs.anl.gov>

Add a few comments to indicate what some #endif directives match.

Use PETSC_ERR_PLIB, not PETSC_ERR_LIB, in MatPtAPNumeric_SeqAIJMKL_SeqAIJMKL_SymmetricReal() when input matrix not marked as symmetric.

Thanks-to: Stefano Zampini <stefano.zampini@gmail.com>

Define MatSeqAIJMKL_setup_structure_from_mkl_handle() and MatSeqAIJMKL_update_from_mkl_handle() only if PETSC_HAVE_MKL_SPARSE_SP2M_FEATURE is defined.

This change prevents warnings about functions t

Define MatSeqAIJMKL_setup_structure_from_mkl_handle() and MatSeqAIJMKL_update_from_mkl_handle() only if PETSC_HAVE_MKL_SPARSE_SP2M_FEATURE is defined.

This change prevents warnings about functions that are defined but unused
when building against an older version of MKL that has mkl_sparse_optimize()
but does not have the two stage matrix-multiplication API.

show more ...

Make error messages for MKL failures more consistent in MATSEQAIJMKL.

Rename MatPtAPNumeric_SeqAIJMKL_SeqAIJMKL() to MatPtAPNumeric_SeqAIJMKL_SeqAIJMKL_SymmetricReal() for clarity and consistency.

Style and formatting changes suggested by Barry, plus a few other minor edits.

Thanks-to: Barry Smith <bsmith@mcs.anl.gov>

Add MatSeqAIJMKL_view_mkl_handle().

Add support for new MatProductXXX() API in MATSEQAIJMKL.

This commit also:
Removes MatSeqAIJMKL_create_from_mkl_handle(), as it is no longer needed.
Adds code to handle empty sparse matrices in AIJM

Add support for new MatProductXXX() API in MATSEQAIJMKL.

This commit also:
Removes MatSeqAIJMKL_create_from_mkl_handle(), as it is no longer needed.
Adds code to handle empty sparse matrices in AIJMKL; mostly needed for MatProductXXX() routines, but also for MatMatMultNumeric.
Makes various cleanup, formatting, and typo fixes.

For helpful comments and suggestions,
Thanks-to: Stefano Zampini <stefano.zampini@gmail.com>
Thanks-to: Barry Smith <bsmith@mcs.anl.gov>
Thanks-to: Hong Zhang <hzhang@mcs.anl.gov>

show more ...

Add MatMatMultSymbolic_SeqAIJMKL_SeqAIJMKL_SpMV2().

Merge remote-tracking branch 'origin/maint'

Merge branch 'rmills/fix-aijmkl-create-mkl-handle/maint' into 'maint'

Rmills/fix aijmkl create mkl handle/maint

See merge request petsc/petsc!2942

Fix logic error in MatSeqAIJMKL_create_mkl_handle().