update master so that --with-visibility and split libraries work with mpiuni

mark function declaration and implementation which takes no arguments with void
to prevent warnings from some compilers

Merge branch 'knepley/feature-fe-derivatives'

* knepley/feature-fe-derivatives:
Doc+Plex: Small fix for distribution
PetscFE: Correct type for reference space derivatives
SNES ex62: Updated to

Merge branch 'knepley/feature-fe-derivatives'

* knepley/feature-fe-derivatives:
Doc+Plex: Small fix for distribution
PetscFE: Correct type for reference space derivatives
SNES ex62: Updated to new derivative mapping
SNES ex12: Updated to new derivative mapping
PetscFE: Changed mapping of derivatives - Now they are transformed only pointwise, instead of transforming the whole tabulation
SNES ex62: Works with new derivative mapping
SNES ex12: Works with new derivative mapping
PetscFE: Changed PetscFEIntegrateResidual_Basic() derivative mapping - Instead of mapping the basis derivatives, we only map the pointwise derivatives gradU and f_1 Suggested-by: Jed Brown <jedbrown@mcs.anl.gov>
PetscFE: Removed PetscGenerateFEMQuadrature*.py

Conflicts:
config/builder.py

show more ...

sizeof.c: remove dependence on petsc [and fix syntax error]

change MPIVersion streams to use MPI_Wtime() to be portable to Windows
do not use scalar square root on real number; it upsets Microsoft

Merge branch 'barry/streams'

streams: translate code from 'bash' to 'sh' [fix freebsd]

removed valgrind warning about sending uninitialized data

PetscFE: Removed PetscGenerateFEMQuadrature*.py

added MPI_MAX_PROCESSOR_NAME for MPIUni
fixed streams/process.py script to work cleanly with NPMAX=1

do not require -fopenmp to compile streams
webclient code requires PETSC_USE_SOCKET_VIEWER

removed "data points" from perfect scaling plot of streams

Reported-by: Karl Rupp <rupp@iue.tuwien.ac.at>

remove traceback from python error

make sure that streams benchmark actual speedup takes at least one third of the y axis

use ${PYTHON} process.py instead of ./process.py in makefiles

Reported-by: Jed Brown jedbrown@jedbrown.org

Barry is an idiot, forgot to put back try except in plotting streams

when plotting streams speedup also plot actual bandwidth numbers on right axis

run streams test will alltests

update FAQ to discuss running the new streams processors
final cleanup of running streams

pretty plots of results from running streams

updates for handling streams output

Work on simplifying and automating running of streams to test machine

TAO: s/PETSC_NULL/NULL/ in manual

Touches a few non-TAO files.

Sys: add PetscMalloc1 macro, array allocation without redundant types

The type is inferred from the pointer return type. This patch is
automated via the following script:

git grep -l 'PetscMalloc(

Sys: add PetscMalloc1 macro, array allocation without redundant types

The type is inferred from the pointer return type. This patch is
automated via the following script:

git grep -l 'PetscMalloc(.*sizeof' src | xargs perl -pi -e 's@PetscMalloc\(([^,;]*[^,; ]) *\* *sizeof\([^,;()]+\),@PetscMalloc1($1,@'

This commit contains an additional bug-fix in csrperm.c, fixing pointer
arity. The code was introduced in 2006, but the allocation could not
have been correct at any time. This probably means that
MatDuplicate_SeqAIJPERM has never been tested.

a54129beb540034ba105796c682d589e7e1111f2
Richard Tran Mills <rmills@ornl.gov>

Added MATSEQCSRPERM support for MatDuplicate() and conversion to/from
MATSEQAIJ. Note that these changes are not quite debugged.

show more ...