Begin renaming the PetscOptions objects for the future where PetscOptions will not be a singleton
Split PetscViewerASCIISpinLock into three locks because it was used in nested functions leading to deadlockReported-by: Michele De Stefano <MStefano@slb.com>
Merge remote-tracking branch 'origin/balay/ftn-command_argument'
fix unused variable warning
fix for wrongly named FORTRANNULLOBJECTDEREFERENCE
allow second object argument to be null
F2003: attempt to use command_argument_count() and get_command_argument() for fortran command line argumentsAlso wrap them in fortran routines [instead of attempting to call directly from C]
Merge branch 'master' into barry/fix-petscviewer-attempt-2
Generate a very useful error message if PetscLogView(), PetscLogEventGetPerfInfo() or PetscLogEventGetFlops()is called with a preceeding call to PetscLogBegin() and friends (also -log_summary which
Generate a very useful error message if PetscLogView(), PetscLogEventGetPerfInfo() or PetscLogEventGetFlops()is called with a preceeding call to PetscLogBegin() and friends (also -log_summary which triggers a PetscLogBegin())Reported-by: Theler German Guillermo <gtheler@cites-gss.com>
show more ...
Rework the threadsafety model in PETSc to use PetscSpinLock...() routines for lockingLocks around opening ASCII viewers and around PetscCommDuplicate() and PetscCommDestroy()Support for downloading
Rework the threadsafety model in PETSc to use PetscSpinLock...() routines for lockingLocks around opening ASCII viewers and around PetscCommDuplicate() and PetscCommDestroy()Support for downloading concurrencykit and using concurrencykit spin locks
changed second argument of PetscObjectViewFromOptions() to PetscObject and all routines that use this routine
stripped out all PETSc threadcomm code
updated PETSc directory layout to match standard packaging strategiesinclude/petsc finclude,private,mpiunilib/petsc confbin/petsc*
PetscFPTCreate() was not called in petscinitialize_()
remove some dead code for AD and old AMS java and Matlab utilities
manual merge of conflict
remove PETSC_DESIRE_COMPLEX and instead include complex with petscmath.h UNLESS PETSC_SKIP_COMPLEX is set (for SSL includes currently)
win32: set printf output format to print 2-digit exponent [instead of the default 3-digit]
doc: fix missing manpage for PetscInitializeFortran
Viewer+Options: Added Fortran interface for PetscObjectViewFromOptions()
Merge branch 'barry/add-fortran-num-consts'
win32: fix fortran upcase for petscsetcommonblocknumeric_()libpetsc.lib(zstartf.o) : error LNK2001: unresolved external symbol _petscsetcommonblocknumeric_@32
cuda: fix mismatched include for cublasInit()/home/balay/petsc.clone-3/src/sys/objects/ftn-custom/zstart.c: In function ‘petscinitialize_’:/home/balay/petsc.clone-3/src/sys/objects/ftn-custom/zsta
cuda: fix mismatched include for cublasInit()/home/balay/petsc.clone-3/src/sys/objects/ftn-custom/zstart.c: In function ‘petscinitialize_’:/home/balay/petsc.clone-3/src/sys/objects/ftn-custom/zstart.c:446:3: warning: implicit declaration of function ‘cublasInit’ [-Wimplicit-function-declaration]
numerical constants such as PETSC_PI, PETSC_REAL_MAX are automatically placed in Fortran common blocknow numerical constants are always equal in both languages
support for MPIU operations for quad precision in Fortran now works
12345678910