DMPlex: Fix check in refinement for tet face edges that intersect the cross-tet edge
- My re-interpolation test still does not work

DMPlex: Fix array restore in DMPlexGetFaces_Internal()

DMPlex: Now DMView_Plex() uses the DM name

DMPlex: Added code for 3D uniform refinement
- The dreaded monster checkin

DMPlex ex4: Added tests for mesh symmetry and point orientations

DMPlex ex4: Added a 3D refinement test
- Also removed the need to have a hybrid mesh

DMPlex: Made DMPlexGetFaces_Internal() non-static, and added DMPlexRestoreFaces_Internal() and DMPlexGetRawFaces_Internal()
- This is a nice way to get the canonical face orientations

DMPlex: Fix for DMPlexUninterpolate()

DMPlex: Added DMPlexGetSupportSection()

PetscFE: Changed PetscFEIntegrateResidual() to take auxiliary fields
- Also changed the PetscFEM struct to take auxiliary elements
- Added P_0 Lagrange element

DM: Added a tolerance to DMPrintLocalVec()
- Also added member to DMPlex to store tolerance, use -dm_plex_print_tol to set

Merge branch 'maint'

Documentation and examples from ATPESC 2013 tutorial

* maint:
KSPSetDM docs: fix KSPSetDMActive args, add cross-references
Added shallow water equation mesh (annulus-20.exo

Merge branch 'maint'

Documentation and examples from ATPESC 2013 tutorial

* maint:
KSPSetDM docs: fix KSPSetDMActive args, add cross-references
Added shallow water equation mesh (annulus-20.exo)
cleanup some docs error out if KSPSetDM() is used but not made inactive or DMKSPSetComputeOperators() is not called
docs on KSPSetNullSpace() that KSP auto uses any previously added to the Mat
shallow water examples for tutorial (mesh file annulus-20.exo not included)
tutorials/HandsOnExercise: run jobs with ./ex19; MPICH does not add '.' to path
TS ex11: I believe the default physics is simple advection. I asked Jed to put in shallow water on the annulus. He might be able to do it from Beijing airport.
third case of example 3 had wrong options
handson: update instructions [for ex42]
added <pre> to output files, added text to HandsOnTutorial.html added ex42 example run commands
HandsOn: rename all output files to .html - [to avoid popup on safari]
HandsOn: fix line numbers in URLs to ex*.c.html
added explanation of runs for HandsOnExercises
HandsOn: fix missing html format [from copy/paste]
HandsOn: update some runs
HandsOn: update the run commands
HandsOn: Force Jed to come up with more exciting examples than these
TS ex11: Added shallow water equations
TS ex11: Fixed docs
TS ex11: Added some explanatory text. Need more.
Makefile: Added TS ex11 to EXAMPLESC
Main page: Changed links to petsc-dev website
vis
selected 4 tutorial examples began to cleanup tutorial page introduction on them
added start of Hands on tutorial that works through a few examples bug fix for valgrind broadcasting uninitialized values

show more ...

PetscFE: Now cell blocks are settable from options
- Fixed bad merge in 8residual integration

Merge branch 'karlrupp/feature-snes-ex52-OpenCL' into knepley/feature-dt-fem

* karlrupp/feature-snes-ex52-OpenCL:
SNES ex52: OpenCL integration now natively supports double precision as well.
SN

Merge branch 'karlrupp/feature-snes-ex52-OpenCL' into knepley/feature-dt-fem

* karlrupp/feature-snes-ex52-OpenCL:
SNES ex52: OpenCL integration now natively supports double precision as well.
SNES ex52: Added option to switch between OpenCL and CUDA backends.
SNES ex52: Added test for multiple regular parallel refinements
SNES ex52: Fixed output flag for residual printing
SNES ex52: Added -refinement_rounds to allow multiple regular parallel refinements
SNES ex52: Added uniform parallel refinement - Added test
DMPlex: Uniform refinement does not work on uninterpolated meshes
SNES ex52: Changed output for 3D tests
SNES ex52: Fixed Makefile to support compilation of both the OpenCL and the CUDA version
SNES ex52: OpenCL version can now deal with multiple parallel blocks.
SNES ex52: Fixed OpenCL vector-type initialization for Laplacian as well.
SNES ex52: Fixed static initialization issue in OpenCL kernel for the AMD SDK.
SNES ex52: Fixed OpenCL test suite, added reference output
SNES ex52: Now builder runs for OpenCL are separate
SNES ex52: Unify dispatch function for CUDA and OpenCL - Now only one function works at a time (I think this is alright in the example) - Remove -opencl flag
SNES ex52: Added OpenCL profiling information.
SNES ex52: Now supporting elasticity in 2d via OpenCL.
SNES ex52: Updated OpenCL version such that spatial dimension is now a runtime parameter.
Added OpenCL version of SNES ex52 for the Laplacian.
Builder: Fixed problems for snes/ex52: "Cuda" -> "cuda", stripping "\r" from output for comparison with reference.

Conflicts:
src/snes/examples/tutorials/ex52.c

show more ...

added start of Hands on tutorial that works through a few examples
bug fix for valgrind broadcasting uninitialized values

DMPlex ex3: Update to latest interface
- Now runs without Python generation

DMPlex: Fix DMPlexComputeL2Diff() to take PetscFE

DMPlex: Fix memory leak with refinement

DMPlex+PetscFE: Moved integration functions from DMPlex to PetscFE
- Now PetscFE type indicates the integration method
- Added all machinery for PetscFE types

DMPlex: Small cleanup fo refinement code before we add tetrahedral stuff

DMPlex: Fixed some function names to cleanup the refinement file move

DMPlex: Added DMPlexEqual() and DMPlexUninterpolate()
- These will be needed for testing refined meshes

DMPlex: remove unused variables

Merge branch 'knepley/opt-plex-depth' into knepley/plex-depth-label

Fixes build error due to missing depthLabel field in DM_Plex.

* knepley/opt-plex-depth:
DMPlex: Use BFS to calculate depth ra

Merge branch 'knepley/opt-plex-depth' into knepley/plex-depth-label

Fixes build error due to missing depthLabel field in DM_Plex.

* knepley/opt-plex-depth:
DMPlex: Use BFS to calculate depth rather than crazy recursive scheme
DMPlex: Optimize depth retrieval by storing a pointer to the label - This is a little unsafe since someone could go in and create a new "depth" label out from under you

Conflicts:
src/dm/impls/plex/plex.c

show more ...

DMPlex: Put in METIS support