Merge branch 'tisaac/feature-mat-permute-via-createsubmatrix' into 'main'Make MatPermute call MatCreateSubMatrix as a fallbackSee merge request petsc/petsc!3636
MatConvert_SeqDense_SeqAIJ(): check for square A
Make MatCreateSubMatrix_SeqDense match current interface of MatCreateSubMatrix and test permutation
MATSEQDENSECUDA: add MatCopy support
MatDenseGetColumnVec: prevent from allocating needless GPU memory
MATAIJ: some fixes for device builds
CUDA: add a couple of missing Lazy initializations
Convert MPI error type to PETSc error with string message for all MPI callsNow PETSc examples will ONLY return PETSc error codes and never MPI error codes directly so we can understand and post-pro
Convert MPI error type to PETSc error with string message for all MPI callsNow PETSc examples will ONLY return PETSc error codes and never MPI error codes directly so we can understand and post-process their errors better.The test harness will now automatically retry tests that fail with MPI, this may help with Intel MPI that produces seemingly random failures.Commit-type: error-checking/spend 30m
show more ...
Merge remote-tracking branch 'origin/release'
checkbadSource: apply rules to *.cu *.cpp sources, and expand CHKERRQ check to CHKERR(Q|MPI|CUDA|CUBLAS|CUSPARSE)
Add MatOption MAT_FORCE_DIAGONAL_ENTRIES; rm MAT_NEW_DIAGONALSReported-by: Nourgaliev, Robert Nr <nourgaliev1@llnl.gov>implemented for C = At*B with aij for all the algorithmsimplemented for C = A
Add MatOption MAT_FORCE_DIAGONAL_ENTRIES; rm MAT_NEW_DIAGONALSReported-by: Nourgaliev, Robert Nr <nourgaliev1@llnl.gov>implemented for C = At*B with aij for all the algorithmsimplemented for C = A*B with aij matrices
Add missing PetscStackCallBLAS()Commit-type: i.e. error-checking/spend 25m
Need initialize to prevent test systems from complaining about possibly uninitialized variables.Commit-type: portability-fix/spend 5mReported-by: GitLab CI
SeqDense treats ILU as LU allowing PC defaults to work when dense matrixis not yet knownDMSHELL uses dense format if no format provided.Fix manual pages for MatFactorCommit-type: feature, test
SeqDense treats ILU as LU allowing PC defaults to work when dense matrixis not yet knownDMSHELL uses dense format if no format provided.Fix manual pages for MatFactorCommit-type: feature, testing-fix, documentation/spend 1h
Mat: Replace 0 -> NULL for pointers
Add support of CUDA/HIP initialization at the same time
Merge remote-tracking branch 'origin/maint'
cudavecimpl.h move to include/petsc/private as its used by private/sfimpl.hReported-by: Nidish <nb25@rice.edu>
Do lazy petsc cuda initialization and validationSo that we can avoid overburdening the GPU in CI.See more at https://gitlab.com/petsc/petsc/-/issues/360#note_373723821
MATSEQDENSE: fix multadd and multtransposeadd operations
Limit chance of integer overflow in calls to PetscLogFlops()Don't view inner KSP when it has not yet been created.Commit-type: bug-fixFunded-by: ECP/spend 20mReported-by: Karl Yang <y.juntao@
Limit chance of integer overflow in calls to PetscLogFlops()Don't view inner KSP when it has not yet been created.Commit-type: bug-fixFunded-by: ECP/spend 20mReported-by: Karl Yang <y.juntao@hotmail.com>
MatTranspose_SeqDense: reuse matrix memory for rectangular matrices and MAT_INPLACE_MATRIX
MatDenseGetArrayWrite: Add SEQDENSECUDA implementationwe can create a matrix with only GPU memory, and if the user asks for CPU memory, this needs to be allocatedReported-by: Jose Roman jroman@ds
MatDenseGetArrayWrite: Add SEQDENSECUDA implementationwe can create a matrix with only GPU memory, and if the user asks for CPU memory, this needs to be allocatedReported-by: Jose Roman jroman@dsic.upv.es
12345678910>>...38