Change how solves with multiple RHS work

* Deprecate KSPHPDDMMatSolve
* Add KSPMatSolve, previously known as KSPHPDDMMatSolve
* New option -ksp_matsolve_block_size
* New routines MatDense[Get|Restor

Change how solves with multiple RHS work

* Deprecate KSPHPDDMMatSolve
* Add KSPMatSolve, previously known as KSPHPDDMMatSolve
* New option -ksp_matsolve_block_size
* New routines MatDense[Get|Restore]SubMatrix

show more ...

Merge branch 'stefanozampini/feature-add-hpackages-rebased' into 'master'

Fixes to MatProduct + many bug fixes + new hierarchical matrix support

See merge request petsc/petsc!2800

MatProduct: proper cleanup of additional data structure

Do not mess with view/destroy routine, add an extra slot to MatProduct struct to stash the data, destroy/view routines
Deprecate MatFreeInterm

MatProduct: proper cleanup of additional data structure

Do not mess with view/destroy routine, add an extra slot to MatProduct struct to stash the data, destroy/view routines
Deprecate MatFreeIntermediateStructures in favor of MatProductClear
Clean up base headers for matrix classes
Update tests (remove some todos from ex70 because now the tests work properly)
MatProductSetType should not set the setfromoptions operation!
MatProductClear should reset operations
Remove no longer needed function composition
Add few more PetscInfo messages
Single dispatch routine
MatProductSetFromOptions routines are now allowed to fail
MatProductSetFromOptions_SeqXBAIJ_SeqDense: support for MATPRODUCT_AtB
MatTransposeMatMultNumeric_SeqAIJ_SeqDense: fix bug with leading dimensions
Make internal functions PETSC_INTERN
Fix a couple of places where the destroy routine of the matrix was changed
Products should all use internal data (or composed data attached to the internal data)
Fix few bugs in MPIAIJ_MPIDENSE product with stages

show more ...

MATDENSE: Add MatDenseSetLDA

clean up seqdense data from max and changelda entries

MATDENSE: add API to access columns as vectors

MATMPIDENSE: support CUDA

MatSeqDense: add support to check for symmetricity

Add array accessors routines

fix MatSetRandom for leading dimension

Use PetscSF for MatMult
- uses the right communica

MATMPIDENSE: support CUDA

MatSeqDense: add support to check for symmetricity

Add array accessors routines

fix MatSetRandom for leading dimension

Use PetscSF for MatMult
- uses the right communication pattern instead of point-to-point
- can take advantage of CUDA aware MPI

Fix bugs related with LDA
Fix direct usage of cpu memory: use MatDenseGetArray (and variants) instead
Fix MatZeroRows
Fix MatView in Binary for SeqDense

show more ...

Merge remote-tracking branch 'origin/hzhng/matproduct-api'

This branch reorganizes the API of mat-mat operations: A*B, At*B, A*Bt, PtAP, RARt, A*B*C
- See matproduct.c for detailed description of new API
- removed API function XXXSymbolic() and XXXNumeric()

This branch reorganizes the API of mat-mat operations: A*B, At*B, A*Bt, PtAP, RARt, A*B*C
- See matproduct.c for detailed description of new API
- removed API function XXXSymbolic() and XXXNumeric()
- removed subroutines MatMatMult_xxx, MatPtAP_xxx, ...
- modified MatxxxSymbolic_XXX() prototype (product is an input now)
- selection of algorithm is in MatProductSetFromOptions_xxx()

show more ...

Merge branch 'dalcinl/binary-io' into 'master'

Binary IO Enhancements

See merge request petsc/petsc!2594

MATDENSE: Add support for MPI-IO binary viewers

Merge branch 'jolivet/fix-upstream-hpddm' into 'master'

Multiple fixes to PCHPDDM + MatMatMult_SeqSBAIJ_SeqDense

See merge request petsc/petsc!2123

MatMatMult_SeqSBAIJ_SeqDense

Merge branch 'jolivet/matmatmult-nest-dense' into 'master'

Support for MatMatMult with Nest and MPIDense

See merge request petsc/petsc!2069

Support for MatMatMult with Nest and MPIDense

Merge branch 'stefanozampini/gpu-bddc' into 'master'

Improvements towards BDDC on GPUs

See merge request petsc/petsc!2067

MATSEQDENSECUDA: few improvements and bug fixes

- support standard CPU Mat and Vec inputs in solve routine
- support conversion to SeqDENSE
- MatFactorInvertSchurComplement: support DENSECUDA matric

MATSEQDENSECUDA: few improvements and bug fixes

- support standard CPU Mat and Vec inputs in solve routine
- support conversion to SeqDENSE
- MatFactorInvertSchurComplement: support DENSECUDA matrices

show more ...

Merge branch 'stefano_zampini/GPU-matdensecuda' [PR #1911]

* stefano_zampini/GPU-matdensecuda:
GPU: Initial implementation for SeqDense class on GPUs.

MATSEQDENSE: prepare the class for CUDA extension

expose (intern) few methods via dense.h
For consistency of the memory model, use MatDenseGetArray as much as possible
Remove MatSolveAdd routines as

MATSEQDENSE: prepare the class for CUDA extension

expose (intern) few methods via dense.h
For consistency of the memory model, use MatDenseGetArray as much as possible
Remove MatSolveAdd routines as they were redundant with the public interface
MatMatSolve_SeqDense: let PETSc complain only if MatDenseGetArray is not implemented

show more ...

Merge branch 'jolivet/matmatmult-seqbaij-seqdense' [PR #1887]

* jolivet/matmatmult-seqbaij-seqdense:
MatMatMult(SeqBAIJ,SeqDense,SeqDense)
Preliminary support for C=A*B with A of type SeqBAIJ and B

Merge branch 'jolivet/matmatmult-seqbaij-seqdense' [PR #1887]

* jolivet/matmatmult-seqbaij-seqdense:
MatMatMult(SeqBAIJ,SeqDense,SeqDense)
Preliminary support for C=A*B with A of type SeqBAIJ and B of type SeqDense.
There are (at least) four possible improvements (with respect to performance):

dispatch to specialized routines like mkl_?bsrmm ;
for small bs, use specialized BLAS like libxsmm ;
use batch BLAS ;
use MKL_DIRECT_CALL when possible.

show more ...

MatMatMult(SeqBAIJ,SeqDense,SeqDense)

Merged in haplav/feature-matload-dense-hdf5 (pull request #1795)

MatLoad_Dense_HDF5

Approved-by: BarryFSmith <bsmith@mcs.anl.gov>

MatLoad_Dense_HDF5 impl.

Merge branch 'master' into knepley/feature-plex-elasticity-test

* master: (1553 commits)
-options_table was changed to options_view in 2014 but documentation was not updated to reflect it
config

Merge branch 'master' into knepley/feature-plex-elasticity-test

* master: (1553 commits)
-options_table was changed to options_view in 2014 but documentation was not updated to reflect it
configure: fix F2003 test
Plex test ex5: Removed "future" test, as it resulted in issues in the nightly tests.
Some minor typos in TS
Add ADIOS2.py which I forgot to git add in the past
Plex ex5: Disable test
Doc: Added note for interface change
KSPLSQRMonitorDefault fixed for quad precision
Tao: Added alternative output for chwirut2 test.
TS ex45,ex46: Removed overly refined tests for faster nightly tests.
KSP ex72: Fixed needlessly long-running PCTFS tests.
SNES ex56: Reduced test size to speed up nightly test suite.
KSPSolve_LSQR: don't compute unused norm of rhs
SNES ex13: Reduced problem sizes for nightly tests.
PetscOptionsDeprecated(): fix for NULL prefix
PetscOptionsDeprecated(): add option to print additional info
SNES ex12: Reduced sizes for BDDC tests.
PetscFileRetrieve(): Suppress output to STDERR when downloading a file.
MatMatMult: Fixed unused variable pN.
DMDestroy: Guard call to dm->ops->destroy * Similar guards are in place in many other methods (IS,Vec,Mat) and permit the DM to be destroyed even if DMSetType() was never called.
...

show more ...

Merge branch 'master' into knepley/feature-plex-parallel-submesh

* master: (2931 commits)
remove N in docs for SNESQN
replaced test tolerance with PETSC_SQRT_MACHINE_EPSILON
fortran: PetscComp

Merge branch 'master' into knepley/feature-plex-parallel-submesh

* master: (2931 commits)
remove N in docs for SNESQN
replaced test tolerance with PETSC_SQRT_MACHINE_EPSILON
fortran: PetscComplex type is missing from fortran
initialize variables so that static analyzer doesn't mistakenly think they are used uninitialized.
Move access of prefix until it is actually used
Remove DMDASplitComm2d(), used in one example that was not tested
Functions that do not return PetscErrorCode should not use PetscFunctionBegin/End paradgm
remove unneeded reassignment of same value
Missing CHKERRQ(); Mostly in Tao code
Remove clang static analyser warning about passing NULL variable to strncat()
missing newline in -ts_monitor_extreme
minor fixes for integer counters
Configure: Deal with '__func__' in C89 builds with newer Clang and GCC.
Jenkinsfile: Increased timeout for tests to 5 minutes.
3.9 changes: further clarify removal of -pc_factor_mat_solver_package
Options deprecations: add "PETSc" to inline warning
removing unused variable
MatShift support added to MATSUBMAT. Mimics MATSHELL functionality. Needed for matrix-free Hessians in TAO Newton-Line-Search methods.
Add -ts_monitor_extreme
Dev manual: add note on deprecating options database keys
...

Conflicts:
src/dm/impls/plex/examples/tests/ex4.c

show more ...

Merge branch 'master' into knepley/feature-snes-ex13-spectral

* master: (2350 commits)
migrate warning fix from next-tmp
Revert "Remove dead code."
clarifying comment in the code
Remove dead

Merge branch 'master' into knepley/feature-snes-ex13-spectral

* master: (2350 commits)
migrate warning fix from next-tmp
Revert "Remove dead code."
clarifying comment in the code
Remove dead code.
Add JuMP citation
Copyedits and updates to developers manual
Add more protection to intrinsics
MPI: protect arguments in MPI macros
Add the missing conditionals for 64-bit build
Plex+VTK: Fix field type determination
Fix bug in name cleanup
Bib: Added ref
Plex+VTK: Fix field type determination
Plex: Updated test output - I am now hating this shit
forgot to save file after moving prototypes to petscsys.h
update Apple example project settings per Xcode
Two functions should actually be public, used, for example from Apple MacOS and iOS programs
includes should always be included relative to include directory
Apple removed old locations of docsets
SNES ex12: Updated test output
...

Conflicts:
src/snes/examples/tutorials/ex13.c

show more ...