Merge remote-tracking branch 'origin/master' into feature-dmmoab

Merged petsc/petsc into master

Merge branch 'knepley/feature-plex-hdf5-parallel-load'

* knepley/feature-plex-hdf5-parallel-load:
HDF5: Fixed Xdmf script
Plex+HDF5: Do not use vectors for periodic visualization
Plex+HDF5: Ig

Merge branch 'knepley/feature-plex-hdf5-parallel-load'

* knepley/feature-plex-hdf5-parallel-load:
HDF5: Fixed Xdmf script
Plex+HDF5: Do not use vectors for periodic visualization
Plex+HDF5: Ignore negative sequence numbers
DMPlex: Now Plex output is parallel - We now write visualization specific topology in /viz/topology - We output a point reordering, coneSizes, cones, and orientations - Now longer need to interpolate on load
DMPlex: Use the presence of faceGeometry in the DM to signal we are using FVM - This will become a PetscFVM object soon
DMPlex: Preserve the block size of the coordinate vector after distribution
Plex+HDF5: Moved all HDF5 to a separate file, and mapped 2D periodic mesh to the cylinder - Now visualization specific things are in /viz
HDF5: Made petsc_gen_xdmf.py executable
Viewer+Options: Added Fortran interface for PetscObjectViewFromOptions()
IS: Stupid mistake. Damn you compiler
DMPlex: Now parallel HDF5 label output does not fail - However, it is also now clear that we will have to write the full interpolation connectivity in order for these to be meaningful
DMPlex: Added DMPlexCreatePointNumbering() - Made DMPlexCreateNumbering_Private() more flexible
IS: Added ISSortRemoveDups()
DMPlex: Added DMPlexInvertCell_Internal() - Stupid type matching
HDF5: Added petsc_gen_xdmf.py which processes PETSc HDF5 output and produces an Xdmf file
DMPlex: Force a serial load in DMLoad_HDF5(), after which we call DMPlexDistribute() - Eventually we would load into a naive partition
IS+HDF5: Corrected code for new block size interface where it is never negative
Viewer+HDF5: Added PetscViewerHDF5ReadSizes() - This allows me to check the size of a Vec or IS to be loaded

show more ...

DMPlex: Fix for DMPlexOrient()
- I was not symmetrizing the orientation graph on processes

DMPlex: Added DMPlexAddConeSize()

DMPlex: Fixed memory overwrite at depth 0 for viewing

DMPlex+Doc: Fixed function doc
Reported-by: Brad Aagaard <baagaard@usgs.ggov>

Merge branch 'knepley/feature-optcontrol-mg' into knepley/feature-fe-dg-space

* knepley/feature-optcontrol-mg:
TAO ex1: documentation
TAO ex1: Now we have a full optimal control problem - The da

Merge branch 'knepley/feature-optcontrol-mg' into knepley/feature-fe-dg-space

* knepley/feature-optcontrol-mg:
TAO ex1: documentation
TAO ex1: Now we have a full optimal control problem - The data for the state is perfect (Q is I) - The regularization is perfect (L is I and a_ref is the exact control)
TAO ex1: Constructed saddle point problem, but control is given exactly
PetscFE: Fix array initialization bug
TAO ex1: Starting an IPM example
DMPlex: When creating a section, we should only complete a label once
DMPlex: Now DMPlexProjectFieldLocal() correctly masks out fields with no function
DMPlex: Added DMPlexVecSetFieldClosure_Internal() - Allows fields to be masked
SNES ex62: Updates for printing changes
DMPlex: Fixes preallocation checks
SNES ex12: Updates for printing changes

show more ...

DMPlex: When creating a section, we should only complete a label once

DMPlex: Added DMPlexVecSetFieldClosure_Internal()
- Allows fields to be masked

DMPlex: Now Plex output is parallel
- We now write visualization specific topology in /viz/topology
- We output a point reordering, coneSizes, cones, and orientations
- Now longer need to interpolate

DMPlex: Now Plex output is parallel
- We now write visualization specific topology in /viz/topology
- We output a point reordering, coneSizes, cones, and orientations
- Now longer need to interpolate on load

show more ...

DMPlex: Use the presence of faceGeometry in the DM to signal we are using FVM
- This will become a PetscFVM object soon

Merge branch 'master' into knepley/feature-plex-hdf5-parallel-load

* master: (125 commits)
since PETSc cannot view factored matrices, generate an error rather than display the wrong stuff (this wa

Merge branch 'master' into knepley/feature-plex-hdf5-parallel-load

* master: (125 commits)
since PETSc cannot view factored matrices, generate an error rather than display the wrong stuff (this was really bad)
outdated example that doesn't match coding style removed and buggy
petscversion.h: workaround ifort -cpp lack of referential transparency
configure: ignore AR, MAKEFLAGS [if set in env]
incorrect first entry in array being zeroed in example
sizeof.c: remove dependence on petsc [and fix syntax error]
nightlybuilds: use consistant PETSC_ARCH, machine names
TS: Allow viewing of TS solutions
DMPlex: Fix comparison for complex
change MPIVersion streams to use MPI_Wtime() to be portable to Windows do not use scalar square root on real number; it upsets Microsoft
cannot use I in PETSc code because in complex it sometimes represents imaginary number cannot pass PetscScalar * to VecNorm()
in dashboard don't warn about SSL and BOI deprecated
TS ex11: Fixed output formatting
Doc+DMPlex: Added function documentation for some refinement functions
streams: translate code from 'bash' to 'sh' [fix freebsd]
barry is an idiot who doesn't have a clue as to what he is doing (and git's unfriendlyness doesn't help matters)
don't list examples under C that don't work with complex!
forgotten output file
removed valgrind warning about sending uninitialized data
nightly builds ignore unrecognized pragma and .c file as C++ code
...

Conflicts:
include/petsc-private/dmpleximpl.h

show more ...

Merge branch 'master' of bitbucket.org:petsc/petsc
Fuck if git tells me to git pull then I pull, if you want me to do something else than have git tell what to do

Plex+HDF5: Moved all HDF5 to a separate file, and mapped 2D periodic mesh to the cylinder
- Now visualization specific things are in /viz

Merge branch 'knepley/fix-plex-bc'

* knepley/fix-plex-bc:
TS ex11: Fixed output formatting
Doc+DMPlex: Added function documentation for some refinement functions
TS ex11: Must separate reconst

Merge branch 'knepley/fix-plex-bc'

* knepley/fix-plex-bc:
TS ex11: Fixed output formatting
Doc+DMPlex: Added function documentation for some refinement functions
TS ex11: Must separate reconstruction setup from geometry setup
DMPlex: DMPlexConstructGhostCells() should not be calling DMSetFromOptions()
DMPlex: FVM BC should only be applied to faces - Refinement adds in vertices/edges on refined faces
DMPlex: Must decide what kind of BC to apply for output

show more ...

Merge branch 'knepley/feature-plex-periodic'

* knepley/feature-plex-periodic:
TS: Allow viewing of TS solutions
DMPlex: Fix comparison for complex
Draw: Fixed bad check
SNES ex62: Enhance ex

Merge branch 'knepley/feature-plex-periodic'

* knepley/feature-plex-periodic:
TS: Allow viewing of TS solutions
DMPlex: Fix comparison for complex
Draw: Fixed bad check
SNES ex62: Enhance ex62 to use tensor-product cells
DMPlex: Now we are correctly creating the Mobius band - Coordinates for uniform refinement are still broken
DMPlex: If domain is periodic, uniform refinement of coordinates will respect this
DMPlex: Allow geometry routines to recognize periodic quads
DMPlex: Added DMPlexLocalizeCoordinates() - This pulls coordinates from vertices to the cell, which allows discontinuities - We use this for periodic problems
DM: Added preliminary support for periodic coordinates - Added DMGet/SetPeriodicity() which is for clipping coordinates to a periodic box
DMPlex: Needed a custom Fortran wrapper for DMPlexInterpolate() because now we require that the incoming mesh be NULL
DMPlex: Fix handling of Fortran NULLs
DMPlex: Updated test output - The DMView was changed
SNES ex12: Cleanup structure - Use PetscFECreateDefault()
PetscFE: Added a view for Basic
SNES ex12: Starting to cleanup test names
SNES ex62: Marked MF tests as broken

show more ...

DMPlex: Fix comparison for complex

DMPlex: Must decide what kind of BC to apply for output

Merge branch 'knepley/feature-plex-bc'

* knepley/feature-plex-bc:
DMPlex: Cells in the boundary marker should not be interpreted as constraints
DMPlex: Fix declaration
TS ex11: Replace ApplyBC

Merge branch 'knepley/feature-plex-bc'

* knepley/feature-plex-bc:
DMPlex: Cells in the boundary marker should not be interpreted as constraints
DMPlex: Fix declaration
TS ex11: Replace ApplyBC() with DMPlexInsertBoundaryValuesFVM()
DMPlex: Fixed DMPlexInsertBoundaryValuesFVM() - Remove redundant variable - Handle absence of id
TS ex11: Reorganization so that physics creation functions can take the DM - This required splitting the options block
DMPlex: DMPlexInsertBoundaryValues() --> DMPlexInsertBoundaryValuesFEM(), and added an FVM version - The FVM version depends on precalculated geometry which needs to be pulled into the library
TS ex11: Remove unused RHSFunction()
TS ex11: Remove Model from BC func - We do not want opaque objects outside of the context so we can pull this in the library
TS ex11: Now use TS infrastructure for local functions
DMPlex: Fix signature for IFunction - Note that the time is still not being passed down to the pointwise function
TS: Added local versions of assembly to DMTS - I was not sure whether everything was necessary
SNES ex62: Updated output
SNES ex12: Removed outdated BC handling - Removed unnecessary label completion - Updated output
DMPlex: Switched FEM assembly routines to use DMPlexInsertBoundaryValues()
DMPlex: Now DMPlexInsertBoundaryValue() only sets values for the given label - Added DMPlexProjectFunctionLabelLocal()
DMPlex: Now DMCreateSection() calls DMPlexLabelAddCells() - This is necessary for FEM boundary conditions, but we do not want it for FVM so we need some control divergence here soon
DMPlex: Added DMPlexLabelAddCells() - Needed this for FEM BC
DMPlex: Fix Gmsh test

show more ...

DMPlex: Added DMPlexLocalizeCoordinates()
- This pulls coordinates from vertices to the cell, which allows discontinuities
- We use this for periodic problems

Merge branch 'knepley/feature-plex-bc' into knepley/feature-plex-periodic

* knepley/feature-plex-bc:
DMPlex: Cells in the boundary marker should not be interpreted as constraints
DMPlex: Fix dec

Merge branch 'knepley/feature-plex-bc' into knepley/feature-plex-periodic

* knepley/feature-plex-bc:
DMPlex: Cells in the boundary marker should not be interpreted as constraints
DMPlex: Fix declaration
TS ex11: Replace ApplyBC() with DMPlexInsertBoundaryValuesFVM()
DMPlex: Fixed DMPlexInsertBoundaryValuesFVM() - Remove redundant variable - Handle absence of id
TS ex11: Reorganization so that physics creation functions can take the DM - This required splitting the options block
DMPlex: DMPlexInsertBoundaryValues() --> DMPlexInsertBoundaryValuesFEM(), and added an FVM version - The FVM version depends on precalculated geometry which needs to be pulled into the library
TS ex11: Remove unused RHSFunction()
TS ex11: Remove Model from BC func - We do not want opaque objects outside of the context so we can pull this in the library
TS ex11: Now use TS infrastructure for local functions
DMPlex: Fix signature for IFunction - Note that the time is still not being passed down to the pointwise function
TS: Added local versions of assembly to DMTS - I was not sure whether everything was necessary
SNES ex62: Updated output
SNES ex12: Removed outdated BC handling - Removed unnecessary label completion - Updated output
DMPlex: Switched FEM assembly routines to use DMPlexInsertBoundaryValues()
DMPlex: Now DMPlexInsertBoundaryValue() only sets values for the given label - Added DMPlexProjectFunctionLabelLocal()
DMPlex: Now DMCreateSection() calls DMPlexLabelAddCells() - This is necessary for FEM boundary conditions, but we do not want it for FVM so we need some control divergence here soon
DMPlex: Added DMPlexLabelAddCells() - Needed this for FEM BC
DMPlex: Fix Gmsh test

show more ...

DMPlex: Now parallel HDF5 label output does not fail
- However, it is also now clear that we will have to write the full interpolation connectivity in order for these to be meaningful

DMPlex: Added DMPlexCreatePointNumbering()
- Made DMPlexCreateNumbering_Private() more flexible

DMPlex: Added DMPlexInvertCell_Internal()
- Stupid type matching