Section+Doc: Small doc fix
Move hash.h to include/petsc/private/.
fix numerous documentation errors due to cut and paste etcReported-by: Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk>
Add initial proper support for multiple implementations of ISGlobalToLocalMappingApply()Provides two subtypes basic and hash. Selects hash automatically if the memory needed for basic(end - start)
Add initial proper support for multiple implementations of ISGlobalToLocalMappingApply()Provides two subtypes basic and hash. Selects hash automatically if the memory needed for basic(end - start) is greater than 4 times the size of the local to global mapping (n) and 1,000,000Utilizes the hash support provided by Lawrence MitchellThanks-to: Lawrence Mitchell
show more ...
Fix bug introduced (write to idxout)
Fix comments for sowing
ISLtoG: Call setfromoptions in create?
ISLtoG: Allow controlling GtoL behaviour from optionsUse -ltogm_use_hash_table to use a memory scalable approach forapplying the global to local mapping.
ISLtoG: Add option to use hash table to GtoLThis removes non-scalable O(Nglobal) memory usage forISGlobalToLocalMappingApply.
f90impl.h needs to be available for users with --prefix installationReported-by: Satish Balay <balay@mcs.anl.gov>
Remove the use and definition of __FUNCT__ throughout the codeSince all modern C/C++ compilers provide this functionality we no longer need to provide it manually in PETScTime: 1.5 hoursThanks-t
Remove the use and definition of __FUNCT__ throughout the codeSince all modern C/C++ compilers provide this functionality we no longer need to provide it manually in PETScTime: 1.5 hoursThanks-to: Andreas Mang <andreas@ices.utexas.edu>
Merge branch 'stefano_zampini/allow-late-matsetblocksizes'
Updated and simplified Fortran interfaceReduced from 4 distinct approaches to using PETSc from Fortran to 1No more .h90 files (Fortran 90 functionality is now always available)Parameters and func
Updated and simplified Fortran interfaceReduced from 4 distinct approaches to using PETSc from Fortran to 1No more .h90 files (Fortran 90 functionality is now always available)Parameters and function prototypes are now provide through Fortran modules, not includes;only #define are provided in the petscXXXdef.h filesPETSc objects can be declared as either type(tXXX) or XXX. For example type(tVec) or VecMoved now private Fortran include files out of public include directory since uses don't use them;they are only used to generate the Fortran modulesEasier to support and test; we can now easily add more function prototypesAlso had to rework generation of manual pages to consistently use MANSEC and SUBMANSECsince these variables also handle how the Fortran modules are generated.Time: 57 hours
ISLocalToGlobalMapping: fix ISLocalToGlobalMappingSetBlockSizecode is simpler now
Added ISLocalToGlobalMappingDuplicateif refcnt > 0 we cannot change the block size of the l2gmap in-place
PetscLayout: change the block size of the l2gmap only if it is blocked
ISLocalToGlobalMapping: fix SetBlockSize method.added other cases to ex8.c and verbose output
ISLocalToGlobalMapping: free globals and manage correctly the negative indices
First attempt to allow changing the block size of a matrix (and any associated l2gmap)
Merge remote-tracking branch 'origin/maint'
PetscSection: Fix View() to use viewer with right communicator
cast memzeros to void to avoid compiler complaints about const qualifiers
PetscSection: add PetscSectionSymThis object's is essentially a Petsc-idiom way of letting the userspecify implementations ofPetscSectionGetPointSyms()PetscSectionRestorePointSyms()PetscSectio
PetscSection: add PetscSectionSymThis object's is essentially a Petsc-idiom way of letting the userspecify implementations ofPetscSectionGetPointSyms()PetscSectionRestorePointSyms()PetscSectionGetFieldPointSyms()PetscSectionRestoreFieldPointSyms()Like PetscDualSpaceGetSymmetries(), the symmetries are descibed by R *P, where R is a flip (or rotation with complex scalars). I had thoughtabout a more managed approach, like PetscSectionGet/SetValues, whichentirely hid the existence of symmetries to begin with, but:- then you have to differentiate between different PetscDataTypes (e.g., for indices, just perform the permutation, not the flips/rotations), and- what about inserting an element matrix? When inserting an element matrix, it makes the most sense to apply flips/rotations to the element matrix, and the permutations to the indices where the matrix is to be inserted with MatSetValues().so in the end I decided to just provide a simple interface to pass thesymmetries to the users and let them figure out how they want to usethem.
fix typo in isltog comment
MatIS: raise errors in MatSetValues if any of the indices is not mapped locallyRemoved buggy ISG2LMapApply (just used by MATIS). Use ISGlobalToLocalMappingApply insteadAdded MatSetValuesBlocked
1...<<11121314151617181920>>...35