use MatSetValues_MPIAIJ_Symbolic instead of MatSetValues for parallel MatMatMult
added function 'MatSetValues_MPIAIJ_Symbolic'
Merge branch 'pr1042/a_selinger/feature-matmatmult/master' [PR #1042]* pr1042/a_selinger/feature-matmatmult/master: - added another sequential matmatmult algorithm (combined) - added another para
Merge branch 'pr1042/a_selinger/feature-matmatmult/master' [PR #1042]* pr1042/a_selinger/feature-matmatmult/master: - added another sequential matmatmult algorithm (combined) - added another parallel matmatmult algorithm (seqMPI) - added test for new seq and parallel matrix multiplication algorithms
show more ...
Merge branch 'maint'
MatMatMult: Replaced qsort() by PetscSortInt().This should fix issues with 64bit-integers.
Fix error with -mat_aijmkl_no_spmv2 with MKL versions >= 18 update 2.Commit-type: bug-fixReported-by: Richard Tran Mills <rmills@rmills.org>
fixed error which arose for non-square matrices
updated files according to pull request comments
added another parallel matmatmult algorithm (seqMPI)
added another sequential matmatmult algorithm (combined)
Merge branch 'barry/rm-fshift-error-hack-sor/maint'
Remove hack of using special fshift value of -1 to indicate error handling, was broken for MatSOR_SeqAIJand did nothing constructive in the other casesCommit-type: bug-fix,style-fixReported-by: V
Remove hack of using special fshift value of -1 to indicate error handling, was broken for MatSOR_SeqAIJand did nothing constructive in the other casesCommit-type: bug-fix,style-fixReported-by: VEYS Stéphane <Stephane.VEYS@cea.fr>
Fix a typo in a debug test
MatMPIAIJGetCondensed: compose IS with global columns with the return matrix
MATIS: added support for naive MPIXAIJ -> MATIS conversion
Mat: added API to inquire congruency of LayoutsFix a bug in MatDiagonalSet_MPIAIJ
Merge remote-tracking branch 'origin/maint'
Merge branch 'hzhang/mumps-spRHS'
MatMumps: log flops for factorization and elimination
MatMumps: small fix to MatView
MatMumps: no need to call the Schur support for zero sized Schur complements
fix warning reported by the nightly tests
Fix a wrong macroFix a code claiming it is wrong. The macro was wrong in multiple places besides changing nnz
1...<<61626364656667686970>>...265