| f0bcadac | 06-Apr-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Merge branch 'pr255/mapdes/sf-multiple-in-flight/master'
* pr255/mapdes/sf-multiple-in-flight/master:
SF: fix multiple in flight comm rounds for basic SFType |
| f145bd14 | 06-Apr-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Fix compiler warnings about unreachable code |
| 056833d7 | 05-Apr-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
Merge branch 'barry/feature-hdf5-flexible-initial-dimension' |
| 8e84e3e8 | 02-Apr-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Merge branch 'master' into knepley/feature-viewer-partial-read
* master: (48 commits)
ViennaCL: Fixes memory leak when destructing an aijviennacl matrix.
MatLoad_MPIDense() always allocated arra
Merge branch 'master' into knepley/feature-viewer-partial-read
* master: (48 commits)
ViennaCL: Fixes memory leak when destructing an aijviennacl matrix.
MatLoad_MPIDense() always allocated array space even if it was previously provided
Clarified lumping options for PCFieldSplitSchurPreType SELFP.
fixed up PetscInfo in GAMG
TSErrorWeightedNorm[2|Infinity]: API changes
MatLoad_SeqDense() always allocated new space for the data even if the user already provided it
Fixed DMCompositeGetAccessArray() to support readonly vectors
Fix obviously-wrong aspects of documentation on SNESVINEWTONRSLS and SNESVINEWTONSSLS.
fixed example to compile - demonstrates a bug
var name 'len' was F90 key word
For user provided Jacobian functions they should always assembly into the second matrix, the one from which the preconditioner is built because the first may be used as matrix-free with -snes_mf_operator
Suggest --with-64-bit-indices if users mistakenly use --with-64bit-indices
Fix configure logging of integer size.
missing underscore in #define of vecgetarrayread translation of fortran stub for VECGETARRAYREAD
updated test output files to reflect differences introduced previously in PCView output
-ksp_type preonly does not support -ksp_monitor_short so remove from test line to prevent extranious warning output
missing CHKERRQ; and ugly double declaration of ierr
Plex: Reorganized PlexPoint interface - Added DMPlexGetPointLocalField() and DMPlexGetPointGlobalField() - Now the DMPlexGetLocalOffset*_Private() functions return start and end - Uniformly use private functions
Plex: Put in explicit cast
MatSetValuesMPISBAIJ will now print correct global indices when error generated due to new nonzero location
...
# Please enter a commit message to explain why this merge is necessary,
# especially if it merges an updated upstream into a topic branch.
#
# Lines starting with '#' will be ignored, and an empty message aborts
# the commit.
show more ...
|
| 45038af3 | 07-Feb-2015 |
Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> |
SF: fix multiple in flight comm rounds for basic SFType
The check to find the correct pack to complete a communication round
didn't check the key at all, so do that.
Adds simple test that two comms
SF: fix multiple in flight comm rounds for basic SFType
The check to find the correct pack to complete a communication round
didn't check the key at all, so do that.
Adds simple test that two comms rounds over the same SF produce the
correct answer.
show more ...
|
| ef605946 | 01-Apr-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
turned on optimizations to remove unneeded copies to the same memory in VecScatter and hence DMLocalToLocalBegin/End() |
| 82ea9b62 | 29-Mar-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
PETSc Vecs with a bs of 1, coming from a generic VecCreate() or DMCreateGlobalVector() with a DMDA
will NOT have an extra dimension attached to them in HDF5 files created with VecView() with HDF5 vie
PETSc Vecs with a bs of 1, coming from a generic VecCreate() or DMCreateGlobalVector() with a DMDA
will NOT have an extra dimension attached to them in HDF5 files created with VecView() with HDF5 viewers,
to have that extra dimension even if the bs is on use either PetscViewerHDF5SetBaseDimension2() or
-viewer_hdf5_base_dimension2. Similarly to handle VecLoad() you need to set this value in the viewer if the
file has the extra dimension
show more ...
|
| c24d300c | 29-Mar-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
add VecView/Load example for HDF5
upgrade VecView example for HDF5 to also load and check it the same |
| b821176e | 27-Mar-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
missing underscore in #define of vecgetarrayread translation of fortran stub for VECGETARRAYREAD |
| f8e4bde8 | 27-Mar-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Viewer: PetscViewerRead() should return the number of items actually read
- This also applies to all its implementations |
| 91c1f63a | 25-Mar-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
Fixed manual pages for VecStrideSet()
also updated a few other related manual pages
Reported-by: Italo Tasso <italo@tasso.com.br> |
| ce6fdca1 | 24-Mar-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
Microsoft C++ compiler cannot perform complex operation on double
Microsoft C++ compiler cannot initialize complex structure
Missing fortran uppercase name replace for VecGet/RestoreArrayRead() |
| d9e4b6ef | 23-Mar-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Merge branch 'knepley/feature-harmonic-advection'
* knepley/feature-harmonic-advection: (86 commits)
SNES: Some docs for NASM
TS: Added view for MIMEX
Plex: Silence compiler warnings
Builder
Merge branch 'knepley/feature-harmonic-advection'
* knepley/feature-harmonic-advection: (86 commits)
SNES: Some docs for NASM
TS: Added view for MIMEX
Plex: Silence compiler warnings
Builder: Check for missing numProcs was wrong
PetscFE: Turn off Fortran binding for PetscFECompositeGetMapping()
Plex: Fix big bug in DMPlexGetGlobalFieldOffset_Internal() - If offset is negative (dof is on another proc), then we must have end < start
Builder: Allow override of numProcs
Plex: Fix Fortran wrappers for DMLabelView() and DMPlexDistribute()
Plex+Label: Added Fortran wrapper for DMPlexCreateLabel()
Plex: Dammit, how did this not get checked in?
Plex: Refiner check for hybrid cells must use hybrid faces as well - Ghost cells in FV method are hybrid
TS: MIMEX now uses work vectors for correct grid
Plex: Adding support for FV in automatic interpolator construction - Notice that the injector cannot do the right thing until we scrap VecScatter
Plex+FEM: Using PetscFVGetDualSpace()
SNES+Doc: Added Options Database keys
PetscFV: Added PetscFVRefine() - This allows interpolation to finer grids needed for FAS
PetscFV: Added a dual space for calculating projections
DM: Better error checking in DMCoarsen()
DMDT+PetscFE: Added PetscQuadratureExpandComposite() - Replaces PetscFECompositeExpandQuadrature() - Added PetscFECompositeGetMapping()
Plex: Fixed complex type error
...
Conflicts:
src/dm/impls/plex/ftn-custom/makefile
src/ts/interface/ts.c
show more ...
|
| 0ef36bf8 | 23-Mar-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Vec: Silence compiler warning |
| 22d96a0c | 06-Mar-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Section+Vec: Added PetscSectionGet/RestoreField_Internal() and PetscSectionVecNorm()
- This allows a user to get the hypernorm for a vector with several fields in it |
| aca31432 | 05-Mar-2015 |
Jed Brown <jed@jedbrown.org> |
Vec: fix typo in comment |
| 0ef1316f | 01-Mar-2015 |
Jed Brown <jed@jedbrown.org> |
VecStash BTS: fix block stash InsertMode accounting
Triggered by ksp/ksp/examples/tutorials/ex56
Reported-by: Barry Smith <bsmith@mcs.anl.gov>
Reported-by: Mark Adams <mfadams@lbl.gov> |
| 4153a436 | 01-Mar-2015 |
Jed Brown <jed@jedbrown.org> |
VecStash BTS: fix indexing bug counting sends to rank 0
Triggered by src/vec/vec/examples/tutorials/ex8
Reported-by: Barry Smith <bsmith@mcs.anl.gov>
Reported-by: Hong Zhang <hzhang@mcs.anl.gov> |
| 794a961b | 28-Feb-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
mistakenly used hsize_t for return type from an HDF5 function that returns int
fix code formatting |
| a1dbdba5 | 27-Feb-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
Merge branch 'maint'
Conflicts:
src/dm/impls/da/gr2.c
src/vec/vec/utils/vecio.c |
| 7bcbaff4 | 27-Feb-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
Removed direct use of HDF5_NATIVE_DOUBLE to support single precision as well |
| c5af9b60 | 25-Feb-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Merge branch 'knepley/fix-plex-submesh-section'
* knepley/fix-plex-submesh-section:
DM: Fix FUNCT
DM: Fixed naming of utility function
Plex: Fixed array reuse bug in PetscSectionCreateGlobalSe
Merge branch 'knepley/fix-plex-submesh-section'
* knepley/fix-plex-submesh-section:
DM: Fix FUNCT
DM: Fixed naming of utility function
Plex: Fixed array reuse bug in PetscSectionCreateGlobalSectionCensored()
DM+Plex: Added sanity checking for global sections received from user
Plex: Unified interpolated submesh creation - Now both versions create an SF - Moved DMPlexFilterPoint_Internal() - DMPlexMarkCohesiveSubmesh_Interpolated() now takes a DMLabel instead of a string
Plex: Allow NULL Label in DMPlexMarkSubmesh_Interpolated()
show more ...
|
| 4ce5b3ad | 25-Feb-2015 |
Matthew G. Knepley <knepley@gmail.com> |
Vec: Do not set block size until we are forced to do it for subvector |
| c98fd787 | 23-Feb-2015 |
Barry Smith <bsmith@mcs.anl.gov> |
MatLoad_ and VecLoad_ routines that use options database information from the binary file .info files need to PetscViewerSetUp() before accessing the options database because the "lazy" setup of bina
MatLoad_ and VecLoad_ routines that use options database information from the binary file .info files need to PetscViewerSetUp() before accessing the options database because the "lazy" setup of binary viewers defered the opening of the files until binary reading was first done which was after the calls to determine matrix and vector block sizes.
show more ...
|
| 14a2fe63 | 21-Feb-2015 |
Jed Brown <jed@jedbrown.org> |
VecAssemblyEnd_MPI_BTS: fix donotstash code path
VecAssemblyEnd needs to set insertmode = NOT_SET_VALUES even if no
values have been set.
Reported-by: Hong Zhang <hzhang@mcs.anl.gov>
Suggested-by:
VecAssemblyEnd_MPI_BTS: fix donotstash code path
VecAssemblyEnd needs to set insertmode = NOT_SET_VALUES even if no
values have been set.
Reported-by: Hong Zhang <hzhang@mcs.anl.gov>
Suggested-by: Barry Smith <bsmith@mcs.anl.gov>
show more ...
|