PCFieldSplit: add support for MATNEST

PCBDDC: added a couple of events for approximate solvers

Some typos.

Merge branch 'barry/2019-09-01/rm-default-language-from-dm' into 'master'

Update public API, and internal naming conventions for DM to eliminate use of...

See merge request petsc/petsc!1998

Update public API, and internal naming conventions for DM to eliminate use of "default" (previously refered to default sections)

Commit-type: style-fix, documentation

Use PetscErrorCode as return for TFS functions that return PETSc error codes, not PetscInt

Commit-type: style-fix

Commit-type: bug-fix
Reported-by: gitlab ci

Fixes for testing

added p4est to valgrind supression file
add missing error checking in xxt and xyt
mark as TODO system examples for error checking that produce inconsistent output
single precision

Fixes for testing

added p4est to valgrind supression file
add missing error checking in xxt and xyt
mark as TODO system examples for error checking that produce inconsistent output
single precision alt files
filters for yaml test

Commit-type: portability-fix, testing-fix

Commit-type: bug-fix

Commit-type: testing-fix
Reported-by: gitlab ci

show more ...

Typos.

Merge branch 'jolivet/redundant-Cholesky-sbaij' [PR #1968]

* jolivet/redundant-Cholesky-sbaij:
PCCHOLESKY in PCREDUNDANT with SBAIJ matrices.

PCREDUNDANT does not currently play nice with SBAIJ mat

Merge branch 'jolivet/redundant-Cholesky-sbaij' [PR #1968]

* jolivet/redundant-Cholesky-sbaij:
PCCHOLESKY in PCREDUNDANT with SBAIJ matrices.

PCREDUNDANT does not currently play nice with SBAIJ matrices out-of-the-box and you end up with something like the following.

[0]PETSC ERROR: #1 MatGetFactor() line 4411 in src/mat/interface/matrix.c
[0]PETSC ERROR: #2 PCSetUp_LU() line 88 in src/ksp/pc/impls/factor/lu/lu.c
[0]PETSC ERROR: #3 PCSetUp() line 894 in src/ksp/pc/interface/precon.c
[0]PETSC ERROR: #4 KSPSetUp() line 377 in src/ksp/ksp/interface/itfunc.c
[0]PETSC ERROR: #5 PCSetUp_Redundant() line 179 in src/ksp/pc/impls/redundant/redundant.c

This is an attempt to fix this by switching to PCCHOLESKY instead of PCLU for symmetric matrices.

show more ...

DMGetSection: replace all internal usage with DMGetLocalSection

Same for DMSetSection -> DMSetLocalSection. We retain the interface
for DMGetSection/DMSetSection without deprecation warnings (for n

DMGetSection: replace all internal usage with DMGetLocalSection

Same for DMSetSection -> DMSetLocalSection. We retain the interface
for DMGetSection/DMSetSection without deprecation warnings (for now).

Suggested-by: Barry Smith <bsmith@mcs.anl.gov>

show more ...

DMGetDefaultSection et al: add PETSC_DEPRECATED_FUNCTION

These functions were deprecated in v3.9, but no
PETSC_DEPRECATED_FUNCTION was added at that time so many uses crept
back into the source code

DMGetDefaultSection et al: add PETSC_DEPRECATED_FUNCTION

These functions were deprecated in v3.9, but no
PETSC_DEPRECATED_FUNCTION was added at that time so many uses crept
back into the source code. Here we mark it deprecated and fix all
existing uses in PETSc.

show more ...

Check the assembled matrix.

PCCHOLESKY with SBAIJ matrices.

Merge branch 'karlrupp/change-gamg-reduction-logic-a-bit' [PR #1952]

* karlrupp/change-gamg-reduction-logic-a-bit:
GAMG: Changed reduction logic a bit and cleaned up format

Merge branch 'Fande-Kong/feature-maij-hmg' [PR #1933]

* Fande-Kong/feature-maij-hmg:
Using MatMAIJ to store interpolations in PCHMG
(1) Rewrote MatGetBrowsOfAcols_MPIXAIJ to extract the remote part

Merge branch 'Fande-Kong/feature-maij-hmg' [PR #1933]

* Fande-Kong/feature-maij-hmg:
Using MatMAIJ to store interpolations in PCHMG
(1) Rewrote MatGetBrowsOfAcols_MPIXAIJ to extract the remote part of interpolations
(2) Refactored MatPtAP__allatonce and MatPtAPXXX__allatonce_merged to take care of MatMAIJ
(3) Added MatMAIJ support to PCHMG

show more ...

changed reduction logic a little and cleaned up format

docs: PCShell man pages: fix argument name inconsistencies

Display the right PC type name

HMG is able to use MATMAIJ for interpolations for multicomponent problems

Merge branch 'rmills/mat-kaij' [PR #1621]

* rmills/mat-kaij:
Getting the KAIJ matrix class integrated into petsc-master.

Merge branch 'jczhang/sf-more-opts' [PR #1567]

* jczhang/sf-more-opts:
Add more optimizations in SF and use it as the default for VecScatter.

Removed duplicate PCApply_PBJacobi_N() introduced in bad automerge by kdiff3.

PCPBJacobi: add PCApplyTranspose (not optimized)

PCPBJacobi: add PCApply_PBJacobi_N (not unrolled/optimized)

Merge branch 'pr1819/jkruzik/feature-pcdeflation/master' [PR #1819]

* pr1819/jkruzik/feature-pcdeflation/master:
PCDeflation.
Implements multilevel deflation preconditioner that can shift part of th

Merge branch 'pr1819/jkruzik/feature-pcdeflation/master' [PR #1819]

* pr1819/jkruzik/feature-pcdeflation/master:
PCDeflation.
Implements multilevel deflation preconditioner that can shift part of the spectrum to a predefined value.

Given a full rank deflation matrix W, the deflation preconditioner uses projections Q = W*(W'*A*W)^{-1}*W' and P = I - Q*A.
It computes computes initial guess x0 = x_{-1} - Q*r_{-1} which is the solution on the deflation space.
The preconditioner application consists of P*M^{-1} + factor*Q, where M^{-1} is an additional preconditioner.

See the documentation for PCDEFLATION above PCCreate_Deflation for more detailed description.

show more ...