1. petsc
  2. PETSc
  3. petsc

Branches

next

View source
next
  • Contributors
    1. Loading...
Author Commit Message Date Builds
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr403/pascgeopc/psanan/ksp-pipe-flex/master' into next
Satish Balay
Merge branch 'master' into next
Matt Knepley
Merge branch 'knepley/fix-configure-relative-installdir' into next * knepley/fix-configure-relative-installdir: Configure: Should use installdir rather than --prefix Configure: The --prefix argument should always be converted to an absolute path
Matt Knepley
Configure: Should use installdir rather than --prefix
Matt Knepley
Configure: The --prefix argument should always be converted to an absolute path
Patrick Sanan
KSPPIPEGCR: initialize sometimes-used vector to NULL This should address this warning from the nightlies: /sandbox/petsc/petsc.clone/src/ksp/ksp/impls/gcr/pipegcr/pipegcr.c: In function ‘KSPSolve_PIPEGCR_cycle’: /sandbox/petsc/petsc.clone/include/petsc/private/kspimpl.h:222:36: warning: ‘t’ may be used uninitialized in this function [-Wmaybe-uninitialized] else {ierr = MatMultTranspose(A,x,y);CHKERRQ(ierr);} …
BarryFSmith
Merge branch 'pr403/pascgeopc/psanan/ksp-pipe-flex/master' into next
Patrick Sanan
KSP tutorial tests: move examples involving KSPPIPEFGMRES, KSPPIPEGCR, and KSPPIPEFCG to TESTEXAMPLES_C_NOCOMPLEX
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'next' of bitbucket.org:petsc/petsc into next
BarryFSmith
Merge branch 'pr403/pascgeopc/psanan/ksp-pipe-flex/master' into next
Patrick Sanan
KSPPIPEFGMRES: add missing iteration count initialization Add a missing initialization for the (optional) iteration count parameter in the PIPEFGMRES cycle function. This should hopefully satisfy these warning from the nightlies: warnings attributed to commit https://bitbucket.org/petsc/petsc/commits/483d696 KSP: Add KSPPIPEFGMRES src/ksp/ksp/impls/gmres/pipefgmres/pipefgmres.c:374 [http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2016/02/…
Patrick Sanan
KSPPIPEFCG: remove extra semicolon Hopefully, this will remove the following warnings from one of the MS test machines: warnings attributed to commit https://bitbucket.org/petsc/petsc/commits/390d8e4 KSP: Add KSPPIPEFCG src/ksp/ksp/impls/fcg/pipefcg/pipefcg.c:609 [http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2016/02/11/build_next_arch-mswin_ps3.log] C:\cygwin\home\sbalay\PETSC~1.CLO\src\ksp\ksp\impls\fcg\pipef…
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr403/pascgeopc/psanan/ksp-pipe-flex/master' into next
BarryFSmith
Merge branch 'master' into next
Patrick Sanan
KSPPIPEFCG, KSPPIPEGCR: remove redundant explicit support for KSP_NORM_NONE See aee7fab88aac, which makes the corresponding change in other KSP implementations. Based on a bug which was Reported-by: Jakub Kruzik <jakub.kruzik@vsb.cz>
BarryFSmith
Merge branch 'master' into next
Jed Brown
Merge branch 'jed/fix-ksp-norm-none' into next * jed/fix-ksp-norm-none: Fix KSP_NORM_NONE priority when greater than 1 DMDARestoreElements: clarify man page and reference from DMDAGetElements
Jed Brown
Fix KSP_NORM_NONE priority when greater than 1 KSP_NORM_NONE is automatically supported with priority 1, but we had code set its priority as high as actual norms, causing it to be selected automatically for those methods. This commit removes those declarations of preferred support, making them in-line with other implementations. Reported-by: Jakub Kruzik <jakub.kruzik@vsb.cz>
Patrick Sanan
KSPPIPEFCG, KSPPIPEGCR, KSPPIPEGMRES: fixes for (not yet supported) complex compilation
sarich
Merge branch 'sarich/feature-tao-pounders-add-correlations' into next
sarich
Update output to reflect new residual formula
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'master' into next
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr401/tferma/virs-pcfieldsplit-fix/master' into next
BarryFSmith
Merge branch 'hongzh/fix-bug-for-cost-integral' into next
BarryFSmith
Merge branch 'hongzh/fix-bug-for-cost-integral' into next
BarryFSmith
Merge branch 'next' of bitbucket.org:petsc/petsc into next
Patrick Sanan
KSP Tutorial examples: loosen KSP convergence tolerance for some pipelined examples
Patrick Sanan
KSPFCG, KSPPIPEFCG, KSPPIPEGCR: fix for FCD truncation strategies enum used to set from options
Patrick Sanan
KSPPIPEFCG, KSPPIPEGCR, KSPFCG: remove macro This removes the macro to compute the number of previous search directions to use. This is code duplication, and perhaps an inline function added to a header (perhaps include/petsc/private/kspimpl.h) might be an alternative.
Patrick Sanan
KSPPIPEFCG: source formatting cleanup
Patrick Sanan
KSPPIPEGCR: source formatting cleanup
Patrick Sanan
KSPPIPEFGMRES: source formatting cleanup
Matt Knepley
Merge branch 'knepley/fix-plex-ex1f90' into next * knepley/fix-plex-ex1f90: Builder: Fix dependencies in some runs Plex: Fix API changes in ex1f90
BarryFSmith
Merge branch 'pr401/tferma/virs-pcfieldsplit-fix/master' into next
BarryFSmith
Merge branch 'pr404/tferma/virs-divergedlinearsolve-fix-rebase/master' into next
BarryFSmith
Merge branch 'hongzh/fix-bug-for-cost-integral' into next
BarryFSmith
Merge branch 'master' into next
Jed Brown
Merge branch 'jed/dmdarestoreelements-strict' into next * jed/dmdarestoreelements-strict: DMDARestoreElements: zero parameters to prevent accidental misuse after call
Jed Brown
Merge branch 'master' into next * master: warning filter for warning C4003: not enough actual parameters for macro 'PETSC_PASTE3_ was accidently put in the .F rule instead of the .c rule install: fix RDict.db error duing 'make install'
sarich
Merge branch 'sarich/feature-tao-pounders-add-correlations' into next
sarich
Merge branch 'sarich/feature-tao-pounders-add-correlations' of bitbucket.org:petsc/petsc into sarich/feature-tao-pounders-add-correlations
sarich
add forgotten files
sarich
add correlations to subproblem
Patrick Sanan
KSPFCG, KSPPIPEFCG, KSPPIPEGCR: fixes to respect changes in naming conventions for truncation strategies, and removal of counter of search space resets
Patrick Sanan
KSP: Add KSPPIPEFGMRES KSPPIPEFGMRES is modification of the pipelined GMRES algorithm described in Ghysels, P., Ashby, T. J., Meerbergen, K., & Vanroose, W. (2013). Hiding Global Communication Latency in the GMRES Algorithm on Massively Parallel Machines. SIAM J. Sci. Comput., 35(1), 48–71 to include variable preconditioning, using the same approach as employed in extending GMRES to FGMRES. Note that it includes a ``…
Patrick Sanan
KSP: Add KSPPIPEGCR Implements the PIPEGCR method, as described in "Pipelined, Flexible Krylov Subspace Methods" P. Sanan, S. M. Schnepp, D. A. May
Patrick Sanan
KSP: Add KSPPIPEFCG Implements the PIPEFCG algorithm as described in "Pipelined, Flexible Krylov Subspace Methods" P. Sanan, S. M. Schnepp, D. A. May
Patrick Sanan
KSPFCG - Define truncation strategies to be used with related solvers
Tobin Isaac
Merge branch 'tisaac/dmforest' into next This branch has progressed to where PetscFV can be used with DMForest in ts ex11. * tisaac/dmforest: DMPlex: remove old DMPlexGet* calls in f90 example DMP4est: fix label/ghost interaction with internal plex representation DM: make DMDestroyLabelLinkList() for internal usage DMP4est: add Vec viewers and loaders via plex DMP4est: cleanup DMP4est: also share value map in DMShareDiscreti…
Tobin Isaac
DMPlex: remove old DMPlexGet* calls in f90 example
Tobin Isaac
DMP4est: fix label/ghost interaction with internal plex representation We have to create labels for the internal plex, construct the ghost cells, and then copy the labels back from the ghosted plex.
Tobin Isaac
DM: make DMDestroyLabelLinkList() for internal usage
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'barry/downloads' into next
BarryFSmith
Merge branch 'barry/downloads' into next
Tobin Isaac
DMP4est: add Vec viewers and loaders via plex These can be replaced with p4est native vtk routines if need be, but that would require my digging into the layout of dofs to vertices, cells, etc., which I don't want to do right now.
Tobin Isaac
DMP4est: cleanup
Tobin Isaac
DMP4est: also share value map in DMShareDiscretization()
Tobin Isaac
DMP4est: stop using DMPlex to get global vectors. It creates confusion in VecGetDM(), and it's only advantage right now is the ability to piggyback off of the VecView*() routines that plex implements. We can put those back in later if needed.
Tobin Isaac
typo
Hong Zhang
Merge branch 'hzhang/superlu_dist-4.3-ex17' into next
Tobin Isaac
DMP4est: share point sf of ghosted plex back to p4est Even though the ghosted cells aren't really present for the p4est, right now all numbering has to coordinate between the two.
BarryFSmith
Merge branch 'barry/add-pcreset-hypre-for-master' into next
BarryFSmith
Merge branch 'pr394/stevengcj/blassuffix/master' into next
BarryFSmith
Merge branch 'barry/downloads' into next
BarryFSmith
Merge branch 'barry/downloads' into next
Tobin Isaac
dmplexts: try to convert if not plex
Tobin Isaac
ts ex11: use DMProjectFunction() for initial condition. It's overkill, but makes it more interoperable
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'barry/downloads' into next
BarryFSmith
Merge branch 'master' into next
Tobin Isaac
DMP4est: share application context with ghosted plex In ts ex11, I now have to address the fact that there is plex specific stuff in SetInitialCondition(). Work on that tomorrow.
Tobin Isaac
DMP4est: use DMLabelGetValue, since this doesn't need to build the index
Tobin Isaac
DMP4est: do not loop over ghost support cells
Tobin Isaac
DMP4est: first pass at accounting for ghost cells
Tobin Isaac
ts ex11: start mesh conversion code Doesn't work yet. All plex<->p4est code has assumed no ghost cells thus far. I still conjecture the FV can be implemented just as efficiently without ghost cells, but my options right now are (a) try to replace ghost cells, then address plex<->p4est fv concerns, or (b) make plex<->p4est fv work for ghost cells. Given the time constraints, I'm going to go for the latter and hope that I can convi…
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'hongzh/checkpointing-update' into next
BarryFSmith
Merge branch 'hongzh/add-checkpointing-support-for-adaptive-ts' into next
Satish Balay
Merge branch 'master' into next
sarich
Merge branch 'sarich/xsdk-add-test-arch' into next
Matt Knepley
Merge branch 'knepley/fix-configure-HOME' into next * knepley/fix-configure-HOME: Configure: Give proper default if HOME is missing from the environment Configure: I think RDict should use threads by default, even if the script does not
BarryFSmith
Merge branch 'hongzh/add-checkpointing-support-for-adaptive-ts' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
Tobin Isaac
Merge branch 'tisaac/buildCheck-ignore-vc-moaning' into next * tisaac/buildCheck-ignore-vc-moaning: checkBuilds.py: ignore Microsoft VC's false positives about bad macros. boost: requires bzlib.h to build - so add a check. Also update to latest version 1.60.0 updated Trilinos.py to check for a Zoltran routine since it no longer contains TPI
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
BarryFSmith
Merge branch 'barry/fix-log-summary-format' into next
Matt Knepley
Merge branch 'knepley/fix-configure-saving' into next * knepley/fix-configure-saving: Configure: Fix retrieval logging in package Configure: Fix logging for shell calls in retrieval
Matt Knepley
Merge branch 'knepley/fix-configure-saving' into next * knepley/fix-configure-saving: Configure: I think RDict should use threads by default, even if the script does not Configure: If not using threads for RDict, only save at the end Configure: If using --help, the RDict should be readonly Bib: More refs
Satish Balay
Merge branch 'master' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
Matt Knepley
Merge branch 'knepley/fix-plex-natural-mem' into next * knepley/fix-plex-natural-mem: Plex: Fix memory leak in natural-to-global In the manual page, http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/KSP/MatGetSchurComplement.html#MatGetSchurComplement towards the bottom, when discussing using PetscObjectCompose() to provide a user method to compute the Schur complement, the string which is queried should be docs: replace os.popen('ls'…
Matt Knepley
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: Forest: Fixed memory leak
Matt Knepley
Forest: Fixed memory leak
Matt Knepley
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: Plex: Fixed DMPlexProjectField()
Matt Knepley
Plex: Fixed DMPlexProjectField()
BarryFSmith
Merge branch 'pr395/dmeiser/fix-veccuspgetarraywrite/master' into next
Satish Balay
Merge branch 'hzhang/update-networkex' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
Satish Balay
Merge branch 'hzhang/update-networkex' into next
sarich
Merge branch 'sarich/check-fcflags-fflags' into next
BarryFSmith
Merge branch 'barry/Sylvain-gmres-symmetric' into next
BarryFSmith
Merge branch 'pr394/stevengcj/blassuffix/master' into next
Hong Zhang
Merge branch 'hzhang/inode-erroriffpe' into next
Satish Balay
Merge branch 'balay/to-maint-metis-parmetis-nocxx' into next
Matt Knepley
Merge branch 'knepley/fix-configure-dirs' into next * knepley/fix-configure-dirs: Configure: Need to check that CMake was found when installing a CMakePackage
Matt Knepley
Merge branch 'knepley/fix-configure-dirs' into next * knepley/fix-configure-dirs: Configure: Must create the lib/petsc/conf if it does not exist
Satish Balay
Merge branch 'balay/superlu_dist-4.3-p1' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'barry/fix-matseqaijgetarrayf90' into next
Matt Knepley
Merge branch 'master' into tisaac/dmforest * master: (143 commits) Revert "revolve: c++ compile fix" - as revolve now does not require 'bool' anymore revolve: c++ compile fix revolve: enable tests and add to nightlybuilds fix download url for revolve remove outdated unused script removed outdated use of the term petsc-dev for the development version of PETSc streams: precompute NPMAX - and use this value. And now user can use…
Satish Balay
Merge branch 'master' into next
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'barry/fix-matseqaijgetarrayf90' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Satish Balay
Merge branch 'balay/fix-streams-npmax' into next
Matt Knepley
Merge branch 'knepley/fix-dm-clone-coord-section' into next * knepley/fix-dm-clone-coord-section: DM: When cloning, need to propagate the coordinate DM section
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Satish Balay
Merge branch 'master' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
Satish Balay
Merge branch 'balay/superlu_dist-4.3' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
Stefano Zampini
Merge branch 'stefano_zampini/feature-pardiso-schur' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Matt Knepley
Merge branch 'master' into tisaac/dmforest * master: (177 commits) DT ex3: The test should check the number of digits requested DT ex3: One test fails in single precision Plex: Do not automatically create a default section if no fields are defined Plex: DMPlexCopyCoordinates() needs to create the coordinate section if it is missing superlu: no longer requires xerbla slamch dlamch ? superlu: update to v5.1 skip MatSolve() if m…
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
Hong Zhang
Merge branch 'hzhang/superlu-5.1' into next
BarryFSmith
Merge branch 'hzhang/update-networkex' into next
Matt Knepley
Merge branch 'knepley/fix-quadrature-test' into next * knepley/fix-quadrature-test: DT ex3: The test should check the number of digits requested DT ex3: One test fails in single precision
Matt Knepley
Merge branch 'knepley/feature-plex-jacp' into next * knepley/feature-plex-jacp: Plex: Do not automatically create a default section if no fields are defined Plex: DMPlexCopyCoordinates() needs to create the coordinate section if it is missing
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Matt Knepley
Merge branch 'knepley/feature-plex-jacp' into next * knepley/feature-plex-jacp: SNES ex12: Added parallel tests - These test parallel, unstructured interpolation/restriction using FAS Plex+SNES: Now in DMPlexComputeJacobian_Internal(), we also form the Jacobian preconditioner matrix FE: Now we can form either the Jacobian or the Jacobian preconditioned matrix using a flag DS: Now have point wise function slots for the Jacobian precondition…
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Matt Knepley
Merge branch 'knepley/fix-quadrature-test' into next * knepley/fix-quadrature-test: DT ex3: Change tolerance and eliminate tests which need MPFR from regular runs superlu: fix complex build fix conditional Generate an Error if the user calls for TS type Euler with IFunction defined
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hongzh/checkpointing' into next
Tobin Isaac
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: DA: fix C++ externing of newly added prototypes builder: added DMP8est regression tests DMP4est: fix symmetry operations DMP4est: fix uninitialized roots in PetscSFReduce DMP8est: fix indexing in DMPforestLabelsInitialize() DMP8est: fix bad indexing when creating connectivity DMP4est: added serial 2d regression tests to builder.py
Tobin Isaac
DA: fix C++ externing of newly added prototypes
Tobin Isaac
builder: added DMP8est regression tests
Tobin Isaac
DMP4est: fix symmetry operations - DMP8est face orientation table had a missing minus sign - I had the order of the DihedralCompose operands reversed. It's tough to keep track of covariant/contravariant sometimes.
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
BarryFSmith
Merge branch 'hongzh/checkpointing' into next
BarryFSmith
Merge branch 'pr380/semihozmen/semihozmen/ksp-eigen-comp-bugfix/master' into next
BarryFSmith
Merge branch 'master' into next
Tobin Isaac
DMP4est: fix uninitialized roots in PetscSFReduce
Tobin Isaac
DMP8est: fix indexing in DMPforestLabelsInitialize()
Tobin Isaac
DMP8est: fix bad indexing when creating connectivity
Tobin Isaac
DMP4est: added serial 2d regression tests to builder.py
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Matt Knepley
Merge branch 'knepley/fix-dm-clone-coordinates' into next * knepley/fix-dm-clone-coordinates: DM: Clone the coordinate DM instead of reusing in Clone()
Satish Balay
Merge branch 'master' into next
Tobin Isaac
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: PetscSF: compile even if MPI_Reduce_local() is not available
Tobin Isaac
PetscSF: compile even if MPI_Reduce_local() is not available
Matt Knepley
Merge branch 'knepley/fix-dm-clone-coordinates' into next * knepley/fix-dm-clone-coordinates: DM: Clone() should also copy the coordinate DM Tests: Run CUDA tests with datafiles only when using double precision. Tests: Reverted KSP ex10 with CuSparse to use ILU. Tests: Fixed wrong reference output for KSP tutorial 10 with cusparse.
Hong Zhang
Merge branch 'hzhang/update-networkex' into next
Tobin Isaac
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: Fix bad function rename and missed rename related to DMProject routines
Tobin Isaac
Fix bad function rename and missed rename related to DMProject routines
Tobin Isaac
Merge branch 'tisaac/dmforest' into next * tisaac/dmforest: snes ex62: fix bad function rename
Tobin Isaac
snes ex62: fix bad function rename
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Tobin Isaac
Merge branch 'tisaac/dmforest' into next This branch introduces the ability to convert a DMPlex mesh to a DMP4est mesh, and generalizes some DMPlex ideas to DM ideas to accomplish it (labels, boundaries, projection functions). * tisaac/dmforest: (83 commits) Fix bad rename builder2.py: handler when 'requires' isn't a list DM: added DMComputeL2GradientDiff() for Plex and DA DMDA: use DMProjectFunctionLocal() and DMComputeL2Diff()…
Tobin Isaac
Fix bad rename
Tobin Isaac
builder2.py: handler when 'requires' isn't a list
Tobin Isaac
DM: added DMComputeL2GradientDiff() for Plex and DA
Tobin Isaac
DMDA: use DMProjectFunctionLocal() and DMComputeL2Diff()
Tobin Isaac
Merge remote-tracking branch 'origin/knepley/feature-plex-add-time' into tisaac/dmforest Both branches changeds DM projection behavior, so I wanted to sort it out before merging * origin/knepley/feature-plex-add-time: (59 commits) SNES ex69: Updated to new function API, fixed test output SNES ex75: Fixed for new function API add fortran stub for TSMonitorLGSetTransform() and fix TSMonitorLG for multiple fields but no names provided for fields add -binary_write_double so __flo…
Tobin Isaac
DMPlex: do not complete cells when completing boundary labels This leads to a possible growth of the boundary into the interior if the same procedure is done more than once.
Tobin Isaac
DMP4est: Set internal plex from options
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr388/mapdes/fix-petscsf-create-section-sf/master' into next
Satish Balay
Merge branch 'master' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr391/zianekhodja/lilia/tsirm/master' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'master' into next
BarryFSmith
Merge branch 'pr391/zianekhodja/lilia/tsirm/master' into next
BarryFSmith
Merge branch 'barry/feature-linesearch-monitor' into next
BarryFSmith
Merge branch 'master' into next
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Tobin Isaac
DMP4est: fix DMPforestLabelsInitialize() For non-cells, I was always taking the "it's on the outside of the tree" path, instead of checking whether the points were in the interior of the tree first.
Tobin Isaac
DMBoundary: Give the link list a reference counter like DMLabel This allows behind the scenes sharing of boundaries in DMP4est with a DMPlex representation.
Tobin Isaac
DMPforestLabelsInitialize(): don't transfer depth label
Tobin Isaac
DMLabel: eliminate default value from regression value output files.
Tobin Isaac
DMForest: aliases for -dm_refine and -dm_refine_hierarchy to match DMPlex
Tobin Isaac
DMP4est: label transfer does not throw an error (in serial in 2D) remove code that was accidentally left commented out
Tobin Isaac
DMPlex: transfer default value when submeshing labels
Tobin Isaac
DMP4est: share constraint matrix with plex
Tobin Isaac
DMP4est: fix P4estFaceToPetscFace Was not the inverse permutation of PetscFaceToP4estFace before, now it is.
Tobin Isaac
PetscSF_Basic: implement PetscSFReduceEnd() for generic MPI_Op using MPI_Reduce_local() Loop over all of the unpack sites and call MPI_Reduce_local(): better than no implementation.
Tobin Isaac
DMLabel: enable explicit control over default label values. It seemed strange to me that DMLabelGetValue() had a default value that means not set, but that DMLabelSetValue() could set a label to have this value. So now, DMLabelSetValue() will ignore attempts to set a value to the default value, but the user can change what the default value is. DMLabel: fix DMLabelDistribute() The recently added default value was not transfered properly, and was accesse…
Tobin Isaac
DMP4est: implement hierarchical transfer SFs Each cell creates a list of the points in its closure that have dofs in the scheme, and a PetscSFNode is created for each point in that closure. Then, the coarse mesh is transfered to overlap the first, so that cell overlap from fine to coarse can be determined. Then, in looping over the closure lists of the fine cells, we can match fine points to their containing coarse points, which allows cons…
Tobin Isaac
DMP4est: initialize labels from baseDM This part is made easy by the fact the the base DM is redundant: it's just a question of finding the right coarse cell to look up. Inheriting labels from a coarse/fine DM will require setting up the intergrid transfer SF's.
Tobin Isaac
DMP4est: user DMCreateLocalVector_Section_Private May need to set View/Load routines at a later date.
Tobin Isaac
DMP4est: ex12 almost runs clean. - I had to fix some function namings in the DMP4est implementation. - Every p4est connectivity needs a geometry in our implementation. - Create DMShareDiscretization to indicate that a second DM is essentially a different representation of the same data as the first, and to allow them to have the same PetscDS related data. There is still a problem with reference counting related to {SNES,KSP,…
Tobin Isaac
DMP4est: do not raise an error for p4est/libsc memory mismatches
Tobin Isaac
PetscInfo_Private: fix Valid Pointer numbers
Tobin Isaac
PetscFinalize(): Destroy the log after finalizing packages. Finalizing a package could generate log/info data, in which case we need to have the log data structures around.
Tobin Isaac
snes tutorial ex12: remove DMPlexComputeL2DiffVec() I don't see any advantage of storing the cell L_2 errors over the residuals themselves.
Tobin Isaac
DM: add DMComputeL2Diff() Lifted from DMPlex DM: typo fix
Tobin Isaac
DMForest: use plex to create default section
Tobin Isaac
DM: move DMProjectFuncton() up from DMPlex
Tobin Isaac
DMP4est: for now, farm out vecs and mats to plex
Tobin Isaac
DMP4est: don't let libsc catch signals Petsc's segv handling is more robust, libsc was overriding it.
Tobin Isaac
DMP4est: copy label names when creating hierarchies I would like to get away with lazy label value copying if I can: determining a point's label value from a coarse ancestor's.
Tobin Isaac
DMP4est: enumerate facets The trickiest part of converting back and forth between a DMPlex and a p4est is that p4est does not keep track of facets explicitly, so when we want to track a label value from the original DMPlex to a facets of a quadrant/octant, we need some help.
Tobin Isaac
DM: DMBoundary is now in DM, not in DMPlex All DMBoundary routines depend only on DMLabel, not on the DAG, so there's nothing to prevent it being used in other DMs.
Tobin Isaac
DMForest: bring in p4est to plex convertino from p4est library I think we still need to set up labels on the p4est side and transfer them to the plex.
Tobin Isaac
snes tutorial ex12: replace DMPlexCopyCoordinates() with DMSetCoordinateDM() I think that coordinate copying is only necessary with the DAG has changed: in all of these cases, DMClone() has been used, so the DAG is the same.
Tobin Isaac
DMp4est: multilevel automatic hierarchy creation workin in snes ex12
Tobin Isaac
DMP4est: some debugging towards automatic hierarchy creation Once I was able to test hierarchy creation in ex12, I uncovered some bugs in DMP4est related to constructing a connectivity from a plex, naming of functions, etc.
Tobin Isaac
DMP4est: work on coarsening 1) We want the coarse mesh to inherit as many options as possible from the parent, so there is now a DMForestTemplate() method for this type of operation 2) Partitioning for coarsening is a major p4est option, so I added it to the DMP4est interface
Tobin Isaac
DM: add DMGet/SetFineDM This requires the introduction of some cycle-detection when destroying a DM. DM: refine NonCyclicReferenceCounting Only recurse in the event of a doubly-linked list. This is to let the swapping that is used to create a DMPlex hierarchy keep working unaffected.
Tobin Isaac
DMForest: introduce initial refinement level. The idea being that, when a minimum and an initial refinement level are specified, then a hierarchy is created that spans them, much like -dm_plex_refine_hierarchy.
Tobin Isaac
pforest: fix naming/registering of conversion methods
Tobin Isaac
SNES example tutorial 12: allow tensor elements
Tobin Isaac
Plex: fix order of setting Coarse DM's when refining When we move coarse DMs out of the plex structure, we have to modify DMPlexReplace_Static() and DMPlexSwap_Static() to move the hierarchy when the plexes that are switching places.
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
sarich
add forgotten files
sarich
add correlations to subproblem
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Satish Balay
Merge branch 'master' into next
Hong Zhang
Merge branch 'hzhang/matpackage-erroriffpe' into next
Satish Balay
Merge branch 'master' into next
Matt Knepley
Merge branch 'knepley/fix-plex-general-interp' into next * knepley/fix-plex-general-interp: Plex: Fixed complex interpolation problem snesvi should only use DM to compute bounds if it can compute bounds fixes for shashi example, converges to some degree with rs method declare everything in example nontrivial nonlinear problem arising from reaction Bib: Many more refs
Satish Balay
Merge branch 'hzhang/matpackage-erroriffpe' into next
935 commits not shown.