deal.II Newsletter #256
Rene Gassmoeller
Hello everyone!
This is deal.II newsletter #256.
It automatically reports recently merged features and discussions about the deal.II finite element library.
Below you find a list of recently proposed or merged features:
#15295: Unify multiple copies of the same function. (proposed by bangerth)
#15294: Document that compute_aspect_ratio_of_cells() requires LAPACK. (proposed by drwells; merged)
#15293: Wrap some PETSc documentation lines. (proposed by drwells; merged)
#15292: Add flag to NM::MappingInfo reinit functions (proposed by jh66637)
#15291: Add changelog entry. (proposed by bangerth; merged)
#15290: Add CITATIONS.bib file to suggest citation in GitHub (proposed by masterleinad; merged)
#15289: TensorProductMatrixSymmetricSum: small modifications (proposed by peterrum; merged)
#15288: MGTwoLevelTransfer: Clean old vector content in prolongate_add (proposed by kronbichler; merged)
#15287: MappingFEField: Remove update_JxW_values for Piola transform (proposed by bergbauer; merged)
#15286: Deprecate our own CUDA backend (proposed by Rombur)
#15285: Fix GitHub Linux CI (proposed by masterleinad; merged)
#15284: Use distinct names for PETSC call_and_possibly_capture_exception functions (proposed by masterleinad; merged)
#15283: Use Kokkos CMake targets (proposed by masterleinad)
#15282: Fix a couple clang-tidy warnings (proposed by masterleinad; merged)
#15281: FEPointEvaluation: Integrate should use JxW (proposed by bergbauer; merged)
#15278: Use dealii-dependencies Docker for GitHub workflow (proposed by pcafrica; merged)
#15277: Fix data format in some changelogs. (proposed by marcfehling; merged)
#15276: recoverable SNES and TS (proposed by stefanozampini)
#15275: Fix some matrix_free tests without vectorization (proposed by masterleinad; merged)
#15274: FEPointEvaluation: skip renumbering when possible (proposed by bergbauer; merged)
#15273: Tensor product kernels: Evaluate integrate values only (proposed by bergbauer; merged)
#15271: Convert PETSc SNES interfaces to use callback error conventions. (proposed by bangerth; merged)
#15270: Fix warnings when compiling with nvcc_wrapper (proposed by Rombur; merged)
#15269: Make the PETSc TS interfaces conform to callback conventions. (proposed by bangerth; merged)
#15267: Add 'requires' clauses for matrix template arguments to PETSc TS and SNES solvers. (proposed by bangerth; merged)
#15264: Suppress 'requires' clauses for doxygen if doxygen is too old. (proposed by bangerth; merged)
#15263: Convert IDA to use the callback standard. (proposed by bangerth; merged)
#15262: Test CUDAWrappers::MatrixFree not only for DEAL_II_WITH_CUDA (proposed by masterleinad; merged)
#15261: remove inexact CGAL kernels and add test (proposed by jh66637; merged)
#15217: Make MatrixFree evaluation kernels more flexible w.r.t. number type (proposed by kronbichler; merged)
#15209: Integrate DG into non-nested MG (proposed by jh66637; merged)
#15165: Introduce DistributedComputeIntersectionLocationsInternal (proposed by jh66637; merged)
#14569: Tensor product matrix collection fix vectorization (proposed by peterrum; merged)
#14536: Choose to initialize ghost elements with reinit(partitioner). (proposed by marcfehling; merged)
#14346: Add process_dofs_vectorized_transpose() variants with constant_offset (proposed by peterrum; merged)
#14327: Precompute inverse diagonal (FDM) (proposed by peterrum; merged)
And this is a list of recently opened or closed discussions:
#15280 Redesign SolutionTransfer (opened)
#15279 Document backward incompatibilities in IDA/KINSOL, etc. (opened and closed)
#15272 New release with missing headers (opened and closed)
#14840 Add C++20-style concepts to deal.II classes and functions (closed)
A list of all major changes since the last release can be found at this website.
Thanks for being part of the community!
Let us know about questions, problems, bugs or just share your experience by writing to this mailing list, or by opening issues or pull requests on Github.
Additional information can be found at our official website.