Revert "Merge remote-tracking branch 'origin/jose/split-reduction' into maint"

should have checked namespace issues..

This reverts commit 52cbd52866ccab13bc62001b3276d0b542ea4117, reversing
changes

Revert "Merge remote-tracking branch 'origin/jose/split-reduction' into maint"

should have checked namespace issues..

This reverts commit 52cbd52866ccab13bc62001b3276d0b542ea4117, reversing
changes made to ccbd7ac4b2d5b3774d1161bff3961426efe80e9b.

show more ...

Merge branch 'maint'

Merge remote-tracking branch 'origin/jose/split-reduction' into maint

Make PetscSplitReduction public

remove incorrect compiler warning about possible unused variables

Reported-by: Jose Roman <jroman@dsic.upv.es>

Merge branch 'master' of bitbucket.org:petsc/petsc

fix: for --with-visibility=1

Merged upstream/master to feature-dmmoab

Merge remote-tracking branch 'upstream/master'

Merge branch 'master' into prbrune/snes-fascontinuation

* master: (855 commits)
fix for moab example had wrong blocksize of mapping
docs:remove src/docs/doxygen/manual [per Barry's instructions]

Merge branch 'master' into prbrune/snes-fascontinuation

* master: (855 commits)
fix for moab example had wrong blocksize of mapping
docs:remove src/docs/doxygen/manual [per Barry's instructions]
gfortran: remove 132 char limit from f90 free-form
SNES: Set up matrix-free *after* copying the user Jacobian to snes->pc
DMPlex-Gmsh: Use trueNumCells when building the coordinate section
Sys: Got guard logic reversed
Sys: The #define guard for PetscMalloc*() must match the one for PetscFree*()
PetscFV: I hate fucking complex
removed ltog scatter from DMDA since functionality can be obtained with gtol scatter and less memory
fix test of DMLocalToGlobalBegin/End and add actual test cases
improve object memory logging for DA, ISLocalToGlobal and ISBLOCK
lower high water mark of constructing parallel scatter by allocating only as much room in svalues as needed
owner in VecScatter construction is the rank, therefor can use PetscMPIInt to hold it do not PetscInt
another reuse of allocated array in 2d/3d DMDA create
improved high water mark of DMDASetups by implementing and using PETSC_USE_POINTER for ISCreateBlock()
PetscFV: Another complex fix
DMPlex+TS: Fixing types
ISIdentity() was wrong for block IS! Indices are NOT scaled by bs
PetscFV: Straighten out complex types
TS ex11: Fix casts and remove unused variable
...

show more ...

Merge branch 'master' into barry/saws-options

Merge branch 'jed/dmda-memusage'

Sys: Got guard logic reversed

Sys: The #define guard for PetscMalloc*() must match the one for PetscFree*()

Merge branch 'master' into stefano_zampini/pcbddc-multilevel-improvement

Conflicts:
src/docs/website/documentation/changes/dev.html
src/ksp/ksp/examples/tutorials/ex59.c
src/ksp/pc/impls/bddc/bdd

Merge branch 'master' into stefano_zampini/pcbddc-multilevel-improvement

Conflicts:
src/docs/website/documentation/changes/dev.html
src/ksp/ksp/examples/tutorials/ex59.c
src/ksp/pc/impls/bddc/bddcprivate.c

show more ...

Merge branch 'knepley/fix-pylith-parallel'

* knepley/fix-pylith-parallel: (22 commits)
DMPlex: In refinement, merge hybrid and non-hybrid SF creation (should be done for everything) - If missing,

Merge branch 'knepley/fix-pylith-parallel'

* knepley/fix-pylith-parallel: (22 commits)
DMPlex: In refinement, merge hybrid and non-hybrid SF creation (should be done for everything) - If missing, define *Max as *End - Fixed bound on hybrid 2D simplicial hybrid cell refinement - Fixes bug in which wrong remote point number was calculated if local process had no hybrid cells, but remote process did so refiners were mismatched
DMPlex: Fix for hybrid refinement - Must correct offsets when neighbor process has no hybrid points
DMPlex: Fix hex hybrid refinement SF - Fixed offset - Fixed number of cell faces - Fixed missing edge offset - Fixed number of cells faces - FIxed missing edge offset
PetscSection: Must initialize remoteOffsets
DMPlex: Fixed cohesive cell creation for clamped boundaries - Explicitly tag cross-edges (edges with two clamped endpoints) with 2 in boundary label - Add cross-edge to vertices when looking for things that need split points replaced - Remove extra check for boundary vertices
DMPlex: Fixes for marking cohesive submesh - Do not mark points as split which are in the boundary label - Mark boundary points from label as unsplit
DMPlex: Changing meaning of boundary label in DMPlexLabelCohesiveComplete() - Now it marks vertices already present in the original label - We mark these as unsplit, and ignore their stars - Remove old code
DMPlex: Fixed stupid typo
DMPlex: Fixed refinement to respect embedded manifolds
DMPlex: Fixed types in DMPlexOrient()
DMPlex: Fixed initialization in DMPlexLabelCohesiveComplete()
DMPlex: Fixed logic in hybrid cell tagging and added missing declaration
DMPlex Fixed huge error with last change set - Default cellHeight is 0
DMPlex: Allow cellHeight argument for DMPlexMarkBoundaryFaces()
DMPlex: Added blabel argument to DMPlexLabelCohesiveComplete() - Lets user designate the boundary of a dividing surface - Added GetSurfaceSize_Static()
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
PC+Doc: Fixed FieldSplit doc
Sys: Replace Malloc1() in multiple mallocs (fix size)
...

show more ...

added properly low memory usage support for VecScatter for blocksize > 12
added simple memory usage function

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

Merged petsc/petsc into master

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 ...

Sys: Replace Malloc1() in multiple mallocs (fix size)

Sys: Replace Malloc1() in multiple mallocs

Merge branch 'jed/malloc-xl-workaround'

* jed/malloc-xl-workaround:
malloc: PETSC_USE_MALLOC_COALESCED, work around IBM XL miscompile
config: fix isIBM compiler test

malloc: PETSC_USE_MALLOC_COALESCED, work around IBM XL miscompile

IBM XL 12.1 for BG/Q and POWER at optimization level -O and above
miscompiles code such as the following, as reduced from PetscMallo

malloc: PETSC_USE_MALLOC_COALESCED, work around IBM XL miscompile

IBM XL 12.1 for BG/Q and POWER at optimization level -O and above
miscompiles code such as the following, as reduced from PetscMalloc3.

struct {double *a,*b,*c;} *t;
t = malloc(sizeof *t);
(*(void**)&t->a = malloc(24)) && (*(void**)&t->b = t->a+1,*(void**)&t->c = t->b+1,0);
printf("%p %p %p\n",(void*)t->a,(void*)t->b,(void*)t->c);

The compiler incorrectly allows reordering of

*(void**)&t->b = t->a+1;

as though this statement did not modify t->b. To work around this, and
to be more flexible on malloc policy (including debugging malloc
issues), we add PETSC_USE_MALLOC_COALESCED which is normally enabled for
debugging=0 builds, but not when using broken compilers.

show more ...

Merge branch 'knepley/feature-plex-hdf5'

* knepley/feature-plex-hdf5: (63 commits)
DMPlex ex2: Added an example that checks a mesh
DMPlex: #ifdef support functions for HDF5
TS: Fix output sequ

Merge branch 'knepley/feature-plex-hdf5'

* knepley/feature-plex-hdf5: (63 commits)
DMPlex ex2: Added an example that checks a mesh
DMPlex: #ifdef support functions for HDF5
TS: Fix output sequence
DMPlex: Check for HDF5 attributes before writing them
DMPlex: Label with timestep number since we do not have times
Viewer+HDF5: Added PetscViewerHDF5HasAttribute()
TS: Set output sequence for timesteps
DMPlex: Handle output sequences - Added static to some functions
DM: Add idea of an output sequence - Added DMGet/SetOutputSequenceNumber() - Added doc for outputDM
REVERT TS: Timestep gets set into DMTS
DMPlex: Fix up HDF5 segregation
DMPlex: Replace MAXPATHLEN with PETSC_MAX_PATH_LEN
DMPlex: Fix HDF5 timestep handling
DMPlex: Set correct block size on HDF5 field vectors
TS: Timestep gets set into DMTS
DMTS: Added timestep
DMPlex: Fix output of fields - Corrected name - Fixed vector type
DMPlex: Return errors from HDF5 function
DMPlex+Viewer: Now output separate fields for viz - Added the PETSC_VIEWER_HDF5_VIZ format for creating viz fields - Added field type determination
Sys+HDF5: Added PetscDataType string and HDF5 support for it
...

Conflicts:
config/BuildSystem/config/packages/__init__.py
src/ts/interface/ts.c

show more ...