Merge remote-tracking branch 'origin/barry/remove-big-g'

Merge branch 'karlrupp/viennacl-getarray'

* karlrupp/viennacl-getarray: Provides VecViennaCLGetArray*() routines
in order to directly call custom OpenCL kernels for
viennacl::vector<PetscScalar>

Merge branch 'karlrupp/viennacl-getarray'

* karlrupp/viennacl-getarray: Provides VecViennaCLGetArray*() routines
in order to directly call custom OpenCL kernels for
viennacl::vector<PetscScalar> objects. This branch also fixes a missing
registration of the vector type and uses the 1.4.2 release instead of 1.4.1.

show more ...

Merge branch 'master' into barry/remove-big-g

Merge branch 'master' into barry/wirth-fusion-materials

ViennaCL: VecViennaCLGetArray*() routines now public.

Allows the user to get pointers to the viennacl::vector<PetscScalar> objects
inside a Vec and e.g. directly launch OpenCL kernels with them.

Re

ViennaCL: VecViennaCLGetArray*() routines now public.

Allows the user to get pointers to the viennacl::vector<PetscScalar> objects
inside a Vec and e.g. directly launch OpenCL kernels with them.

Reported-by: Mani Chandra <mc0710@gmail.com>

show more ...

ViennaCL: Reenabled -vec_type viennacl

This was available already, but got lost at some point.

insure that all casts to (double) are enclosed in ()

completed removal of %G from PETSc, we now generate an error if used

DMPlex: Small fix

PetscSection: Added PetscSectionGetField()

Merge branch 'master' into barry/wirth-fusion-materials
bad manual merge needs to be fixed
this is what happens when branches hang around too long

Conflicts:
src/ts/examples/tutorials/advection-dif

Merge branch 'master' into barry/wirth-fusion-materials
bad manual merge needs to be fixed
this is what happens when branches hang around too long

Conflicts:
src/ts/examples/tutorials/advection-diffusion-reaction/ex10.c

show more ...

VecValidValues() with CUSP: Skipping check if data is only valid on the GPU.

The validation was run also if only the data on the GPU was valid,
which is useless and may even lead to segfaults.
This

VecValidValues() with CUSP: Skipping check if data is only valid on the GPU.

The validation was run also if only the data on the GPU was valid,
which is useless and may even lead to segfaults.
This commit fixes the problem for CUSP vectors.
A more general fix for other accelerator libraries is postponed for now.

show more ...

Doc: Added blocksize argument to VecCreateSeqWithArray()
Reported-by: Torquil Macdonald Sørensen <torquil@gmail.com>

PetscSection: Change clIndices member to clPoints
- Added some docs

fix %d that should be %D

simplify awk script in test example since failed on http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2013/12/30/examples_master_arch-opensolaris-misc_n-gage.log

Merge branch 'jose/mat-vec-fixes'

* jose/mat-vec-fixes:
Correctly check input type of Mat argument in MatMPIAIJGetLocalMat
Fix Matlab viewer for parallel complex vectors

Conflicts:
src/vec/vec

Merge branch 'jose/mat-vec-fixes'

* jose/mat-vec-fixes:
Correctly check input type of Mat argument in MatMPIAIJGetLocalMat
Fix Matlab viewer for parallel complex vectors

Conflicts:
src/vec/vec/impls/mpi/pdvec.c

show more ...

Merge branch 'master' into barry/fix-quad-examples

Merge branch 'barry/fix-math-functions' into barry/fix-quad-examples

Conflicts:
src/snes/examples/tutorials/ex22.c
src/sys/classes/random/examples/tutorials/ex2.c

removed direct use of sin(), sqrt(), isnormal() etc from C source code and C examples

needed for __float128 precision

fix a place with Jed's transformation to PetscMalloc1() failed and convert some cases to PetscCalloc()

also fix a couple examples that fail for 64 bit integers

fixes for quad precision

for some reason %G sometimes works and sometimes doesn't needs debugging

VecView_MPI_ASCII: avoid PetscObjectGetName unless really needed

Historically, VecView output has not included names, but the previous
change to eagerly call PetscObjectGetName() changed output like

VecView_MPI_ASCII: avoid PetscObjectGetName unless really needed

Historically, VecView output has not included names, but the previous
change to eagerly call PetscObjectGetName() changed output like

-Vec Object: 2 MPI processes
+Vec Object:Vec_0x84000005_0 2 MPI processes

This commit reverts to only name vectors when necessary. It fixes the
previously-inconsistent behavior where ranks other than 0 neglected to
call PetscObjectGetName for PETSC_VIEWER_ASCII_VTK and
PETSC_VIEWER_ASCII_VTK_CELL.

Reported-by: Barry Smith <bsmith@mcs.anl.gov>

show more ...

fixes for single precision and complex mostly in Fortran code which did not handle single precision properly at all

VecView_*_ASCII: close leak for INFO and INFO_DETAIL

Add unit test for -vec_view ascii::ascii_info

Reported-by: Barry Smith <bsmith@mcs.anl.gov>