Plex ex8: Make functions static

Plex: Added support for hybrid (prism) cell geometry
- There are 2 differences from normal cells:
1) The bottom face is inverted, since the normal points toward the top face, instead of outward
2

Plex: Added support for hybrid (prism) cell geometry
- There are 2 differences from normal cells:
1) The bottom face is inverted, since the normal points toward the top face, instead of outward
2) The side faces are ordered as tensor products of intervals, not as quadrilaterals

show more ...

Merged in knepley/feature-bdfunc-normal (pull request #1500)

Plex: Feed the facet normal to a type of projection function

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

DMPlexCheck{Skeleton,Faces}: remove simplex arg; fix for undistributed meshes

Plex: Feed the facet normal to a type of projection function
- Added DM_BC_ESSENTIAL_BD_FIELD
- Added DMProjectPoint_BdField_Private

Merged in florianwechsung/petsc/fw/vertexrebalancing (pull request #1421)

DMPLEX/PCPATCH: Rebalancing shared points on parallel interfaces

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

Merged in florianwechsung/petsc/fw/vertexrebalancing (pull request #1421)

DMPLEX/PCPATCH: Rebalancing shared points on parallel interfaces

Approved-by: BarryFSmith <bsmith@mcs.anl.gov>
Approved-by: Lawrence Mitchell <wence@gmx.li>

show more ...

conditional definition of functions when petsc is build with parmetis

squash dead assignments

Plex+SNES: Moved some functions back from SNES to Plex
- Get/RestoreCellFields()
- Get/RestoreFaceFields()
- Get/RestoreFaceGeometry()
- Run the subdomain hook to copy the DMSolver

don't use dm_rebalance_partition_view to check the performance of the rebalancing as we get non-deterministic distributions, we just check in the test that we have actually improved things now

fixed incompatible-pointer-types error when building --with-64-bit-indices=1

Merge remote-tracking branch 'origin/haplav/fix-matpartitioningview-hierarch' [pull request #1470]

Merged in knepley/fix-plex-interpolate-sf-new1 (pull request #1473)

fix DMPlexInterpolatePointSF()

Approved-by: Matthew Knepley <knepley@gmail.com>

ex18: 7_hdf5 and 7_hdf5_ppu require !complex

ex18: 7_hdf5_hierarch requires !complex

DMPlexGetFullJoin: fix inarg validation

ex18: 6_tet made TODO for now

updated output file for dm_impls_plex_tests-ex18_2b

ex18: TODO examples with parallel partitioning of uninterpolated meshes

Plex: Send potential overlap faces to all roots

Plex ex18: Added another test that exposed a bug in parallel interpolation

Plex: Expose DMPlexInterpolatePointSF()

Plex: Rewrote function that puts interpolated points into the point SF in parallel
- It only worked for edges originally
- The orientation of the SF points still does not match when computed in paral

Plex: Rewrote function that puts interpolated points into the point SF in parallel
- It only worked for edges originally
- The orientation of the SF points still does not match when computed in parallel

show more ...

Revert "ex18: hotfix SF for test 4_tet%"

This reverts commit faed488c3615e97c021c2674cfb2b465114efbdc.

Plex test ex18: removed all TODOs from tests - must be fixed in this branch