Changes: 3.17

General:

• PETSc now requires a C99 compliant C compiler in all cases. Previously C99 was only required when building PETSc, but this now extends to public interfaces and header-files

• PETSc now requires a C++11 compliant C++ compiler. Note this requirement is only enforced if C++ is used; it is acceptable to have a compiler that does not support C++11 if you only ever build C source

• PETSc now requires at least Microsoft Visual Studio 2015 when using the Microsoft Visual C/C++ Compiler

• Change MPIU_Allreduce() to always returns a PETSc error code that should be checked with CHKERRQ(ierr) or PetscCall()

Configure/Build:

• Change minimum value of --with-cxx-dialect argument from “03” to “11”

• C++ dialect will now also be inferred from compiler flags, although users will be warned that they should let PETSc auto-detect the flag when setting the dialect this way

• Change C++ dialect flag option to be consistent with compiler flags; --with-cxx-dialect=gnu++14 means you want -std=gnu++14, no more, no less

• Fix for requesting no C++ dialect flag via --with-cxx-dialect=0. Previously configure would bail out immediately without running the tests and therefore wouldn’t set any of the capability defines. configure now runs all tests, just doesn’t add the flag in the end

• Fix a number of corner-cases when handling C++ dialect detection

• Remove deprecated PETSC_VERSION_PATCH so as to not have confusion with patch releases where the subminor version changes

• Change PETSC_HAVE_MKL to PETSC_HAVE_MKL_LIBS

• Add PETSC_HAVE_MKL_INCLUDES

• Enable HYPRE GPU for 64bit indices build (using HYPRE’s mixed-int configuration)

• Reuse PETSc-installed hwloc when installing OpenMPI

Sys:

• Add MPI_Comm_get_name() and MPI_Comm_set_name() to MPIUNI

• Remove petsccublas.h and petschipblas.h

• Remove -petsc_use_default_null_stream and -[cuda|hip]_synchronize options

• Remove PetscCUDASynchronize and PetscHIPSynchronize. Their operation is now managed by PetscDeviceContext via its PetscStreamType attribute

• Remove PetscCUDAInitialize(), PetscCUDAInitializeCheck(), PetscHIPInitialize(), and PetscHIPInitializeCheck(). Their function is now handled by PetscDeviceInitialize() and PetscDeviceInitialized()

• Remove PetscCUBLASInitializeHandle(), PetscCUSOLVERDnInitializeHandle(), PetscHIPBLASInitializeHandle(), and PetscHIPSOLVERInitializeHandle(). Their function is now handled implicitly by PetscDeviceContext

• Remove petsc_gputimer_begin and petsc_gputimer_begin

• Add -device_enable, -device_select and -device_view startup-options to control coarse-grained device initialization strategy

• Replace -[cuda|hip]_device with split options -device_enable_[cuda|hip] and -device_select_[cuda|hip] to enable fine-grained control of device selection and initialization strategy

• Replace -[cuda|hip]_view with -device_view_[cuda|hip]

• Add PetscDeviceInitType to enumerate PETSc device initialization strategies

• Add PetscDeviceInitialize() to eagerly initialize a PetscDeviceType, and PetscDeviceInitialized() to query the corresponding initialization state

• Change PetscDeviceCreate() to also accept a PetscInt devid, to create a PetscDevice for a specific device

• Add PetscDeviceView()

• Move PetscInt64_FMT and MPIU_INT64 definitions to petscsystypes.h

• Add PetscBLASInt_FMT, PETSC_MPI_COMM_FMT, and PETSC_MPI_WIN_FMT format specifiers

• Add petscmacros.h header to house common PETSc preprocessor macros

• Add PetscUnreachable() to indicate unreachable code section to compiler

• Add PetscHasAttribute() macro to query for existence of an __attribute__ specifier

• Add PetscCommGetComm() and PetscCommRestoreComm() to allow reuse of MPI communicator with external packages, as some MPI implementations have broken MPI_Comm_free()

• Add PetscExpand(), PetscConcat(), PetscCompl(), and PetscExpandToNothing()

• Add PETSC_CONSTEXPR_14, PETSC_NULLPTR, and PETSC_NODISCARD

• Add PetscSizeT as a language-agnostic equivalent of size_t from <stddef.h>

• Add PetscCount as a signed datatype for counts, equivalent to ptrdiff_t from <stddef.h>.

• Add PetscCountCast, PetscSortIntWithCountArray(), and PetscSortIntWithIntCountArrayPair()

• Deprecate SETERRQ1() - SETERRQ9() in favor of SETERRQ() which is now variadic

• Deprecate PetscInfo1() - PetscInfo9() in favor of PetscInfo() which is now variadic

• Deprecate PETSC_INLINE, inline is a standard keyword since C99 and C++11

• Deprecate PETSC_STATIC_INLINE, as both static and inline are standard keywords since C99 and C++11

• Remove PETSC_C_RESTRICT, restrict is a standard keyword since C99

• Change SETERRMPI() to be variadic

• Change SETERRABORT() to be variadic

• Add PetscCheck() and PetscAssert() for checking a boolean condition is true. The former is always enabled, while the latter is enabled only in debug builds.

• PetscDevice initialization for CUDA and HIP will now respect CUDA_VISIBILE_DEVICES and HIP_VISIBLE_DEVICES environment variables respectively

• Add PETSC_ATTRIBUTE_COLD to inform compilers that a function is unlikely to be called

• Add PetscCall(), PetscCallVoid(), PetscCallMPI(), PetscCallAbort(), PetscCallContinue(), PetscCallThrow(), and PetscCallCXX(). These supersede CHKERRQ(), CHKERRV(), CHKERRMPI(), CHKERRABORT(), CHKERRCONTINUE(), CHKERRXX(), and CHKERRCXX() respectively

• Add PetscCallCUDA(), PetscCallCUBLAS(), PetscCallCUSPARSE(), PetscCallCUSOLVER(), PetscCallCUFFT(), and PetscCallCURAND(). These supersede CHKERRCUDA(), CHKERRCUBLAS(), CHKERRCUSPARSE(), CHKERRCUSOLVER(), CHKERRCUFFT(), and CHKERRCURAND() respectively

• Add PetscCallHIP(), PetscCallHIPBLAS(), and PetscCallHIPSOLVER(). These supersede CHKERRHIP(), CHKERRHIPBLAS(), and CHKERRHIPSOLVER() respectively

• Add PetscCallCEED() which supersedes CHKERRQ_CEED()

• Soft-deprecate CHKERR variants listed above in C/C++ sources. New code should prefer the PetscCall variants though no compiler diagnostics will be emitted if the old versions are used. One may use ${PETSC_DIR}/share/petsc/chkerrconvert.py to perform the conversions automatically, though users should note that this script is considered “beta” software. It was used to facilitate the CHKERR changes during development and is being exposed to users in order to smooth the transition – while every reasonable effort was made to ensure correctness, we provide no guarantee of it.

PetscViewer:

• Add PetscViewerHDF5SetDefaultTimestepping() and PetscViewerHDF5SetDefaultTimestepping() to deal with HDF5 files missing the timestepping attribute

PetscDraw:

• Add PetscDrawSPAddPointColorized() to change scatter point color based on third input value

AO:

IS:

• ISLocalToGlobalMappingCreateSF(): allow passing start = PETSC_DECIDE

• Add ISGeneralSetIndicesFromMask()

• Add ISSetLayout()

• Add PetscSectionSymDistribute()

• Add ISLocalToGlobalMappingGetType()

VecScatter / PetscSF:

• Add MPI-4.0 large count support. With an MPI-4.0 compliant MPI implementation and 64-bit indices, one can now pass over 2 billion elements in a single message in either VecScatter or PetscSF

• Add PetscSFFetchAndOpWithMemTypeBegin(), which is similar to PetscSFFetchAndOpBegin(), but with explicit memory types

• Change PetscSFSetGraph() and PetscSFSetGraphLayout() to sort leaves, remove const from ilocal and iremote arguments

• Add PetscSFConcatenate()

PF:

Vec:

• Change VecTaggerComputeBoxes() and VecTaggerComputeIS() to return a boolean whose value is true if the list was created

• Add -vec_bind_below option for specifying size threshold below which GPU is not used for Vec operations

• Add VecSetBindingPropagates()

• Add VecGetBindingPropagates()

• For CUDA and ViennaCL and HIP GPU vectors, VecCreate() no longer allocates the array on CPU eagerly, it is only allocated if it is needed

• VecGetArrayAndMemType() and VecGetArrayReadAndMemType() now always return a device pointer (copying the data to the device if needed) for the standard CUDA, HIP, and CUDA/HIP Kokkos vectors. Previously, they did so only when the device had the latest data

• Add VecGetArrayWriteAndMemType() and VecRestoreArrayWriteAndMemType(), which are similar to the VecGetArrayReadAndMemType() family, but only write to the vector on device

PetscPartitioner:

Mat:

• Add -mat_bind_below option for specifying size threshold below which GPU is not used for Mat operations

• Add MatSetBindingPropagates()

• Add MatGetBindingPropagates()

• Add MatSeqAIJGetArrayWrite() and MatSeqAIJRestoreArrayWrite() to get write-access to the value array of MatSeqAIJ on CPU

• Add MatCUSPARSESetUseCPUSolve() Use CPU solve with cuSparse for LU factorization that are on the CPU.

• Change MatCreateIS() behavior when NULL is passed for the mappings. Now a NULL map implies matching local and global spaces

• Add support of MatSetValuesCOO() and MatSetPreallocationCOO() for matrix type AIJKOKKOS. Additionally, for AIJKOKKOS, they support negative indices and remote entries

• Add MatMultHermitianTransposeEqual() and MatMultHermitianTransposeAddEqual()

• Add MatSetPreallocationCOOLocal() to set preallocation for matrices using a coordinate format of the entries with local indices

• Change MatStructures enumeration to avoid spaces and match capitalization of other enumerations

• Change size argument of MatSetPreallocationCOO() to PetscCount

• Add MATORDERINGMETISND use METIS for nested dissection ordering of MatSeqAIJ, with options nseps, niter, ufactor and pfactor under the common prefix -mat_ordering_metisnd_

• Change options -matproduct_<product_type>_via to -mat_product_algorithm

• Add -mat_superlu_dist_3d and -mat_superlu_dist_d <n> to support using SuperLU_DIST’s version 7.2 3d decomposition algorithms

• MATIS now supports negative and repeated indices in the local-to-global map and the COO assembly routines. This allows for better integration with libceed

• Add MatISGetLocalToGlobalMapping() to retrieve the logical map for assembled subdomain problem

PC:

• Add MG option -pc_mg_galerkin_mat_product_algorithm [cusparse|hypre] and PCMGGalerkinSetMatProductAlgorithm() to use cuSparse or hypre’s SpGEMM for Galerkin products in hypre

• Add PC type PCBJKOKKOS a new, experimental batch Kokkos solver -pc_type bjkokkos -pc_bjkokkos_ksp_type [tfqmr|bicg] -pc_bjkokkos_pc_type jacobi -ksp_type preonly

• Add -pc_svd_monitor ::all option to print all the singular values instead of a maximum of ten

PCMG:

• Add PCMGGetGridComplexity() to get operator and grid complexity of MG hierarchy

• Change PCGAMG default to use PCJACOBI smoothing instead of PCSOR. This also allows the default configuration to use GPUs effectively, and to deliver equivalent convergence. For the old default, use -mg_levels_pc_type sor.

• Change PCGAMG eigenvalue estimation to use KSPCG when MAT_SPD has been set (see MatSetOption()) and KSPCR when MAT_SYMMETRIC or MAT_HERMITIAN has been set. These are usually somewhat more accurate and reliable than the previous default of KSPGMRES, and in tune with KSPCHEBYSHEV. Note that Chebyshev will generally not be a suitable smoother for indefinite matrices.

• Change PCGAMG to use PCGAMGSetUseSAEstEig() by default when the smoother uses Jacobi preconditioning.

KSP:

• Outer most KSPSolve() will error if KSP_DIVERGED_ITS and `KSPSetErrorIfNotConverged() is used

• Add KSPQMRCGS to support qmrcgstab with right preconditioning

• Add KSPGuessSetTolerance()

• Add a new model option to KSPGuessFischerSetModel()

SNES:

• Add SNESNewtonTRDCGetRhoFlag(), SNESNewtonTRDCSetPreCheck(), SNESNewtonTRDCGetPreCheck(), SNESNewtonTRDCSetPostCheck(), SNESNewtonTRDCGetPostCheck()

SNESLineSearch:

TS:

• Add TSSundialsSetUseDense() and options database option -ts_sundials_use_dense to use a dense linear solver (serial only) within CVODE, instead of the default iterative solve

• Change timestepper type TSDISCGRAD to include additional conservation terms based on formulation from [Gonzalez 1996] for Hamiltonian systems: - Add TSDiscGradIsGonzalez() to check flag for whether to use additional conservative terms in discrete gradient formulation - Add TSDiscGradUseGonzalez() to set discrete gradient formulation with or without additional conservative terms. Without flag, the discrete gradients timestepper is just backwards euler

• Add TSRemoveTrajectory to destroy and remove the internal TSTrajectory object from TS

• Change TSMonitorSPSwarmSolution() to have correct axes labels and bounds

Tao:

• Add TaoGetGradient(), TaoGetObjectiveAndGradient() and TaoGetHessian()

• Deprecate TaoSetInitialVector() in favor of TaoSetSolution()

• Deprecate TaoGetSolutionVector() in favor of TaoGetSolution()

• Deprecate TaoGetGradientVector() in favor of TaoGetGradient()

• Deprecate TaoSetObjectiveRoutine() in favor of TaoSetObjective()

• Deprecate TaoSetGradientRoutine() in favor of TaoSetGradient()

• Deprecate TaoSetObjectiveAndGradientRoutine() in favor of TaoSetObjectiveAndGradient()

• Deprecate TaoSetHessianRoutine() in favor of TaoSetHessian()

• Change TaoGetObjective(). Use TaoGetSolutionStatus(tao,NULL,&fct,NULL,NULL,NULL,NULL) instead

DM/DA:

• Add DMLabelGetNonEmptyStratumValuesIS(), similar to DMLabelGetValueIS() but counts only nonempty strata

• Add DMLabelCompare() for DMLabel comparison

• Add DMCompareLabels() comparing DMLabels of two DMs

• DMCopyLabels() now takes DMCopyLabelsMode argument determining duplicity handling

• Add -dm_bind_below option for specifying size threshold below which GPU is not used for Vec and Mat objects associated with a DM

• Add DMCreateMassMatrixLumped() to support explicit timestepping, also add DMTSCreateRHSMassMatrix(), DMTSCreateRHSMassMatrixLumped(), and DMTSDestroyRHSMassMatrix()

• Promote DMGetFirstLabelEntry() to public API and rename

• Add bias vector argument to DMGetDefaultConstraints() and DMSetDefaultConstraints(). Passing bias=NULL recovers prior behavior.

• Change DMGetAuxiliaryVec(), DMSetAuxiliaryVec(), and DMGetAuxiliaryLabels() to take an equation part number

DMSwarm:

• Add DMSwarmGetNumSpecies() and DMSwarmSetNumSpecies() to support PIC

• Add DMSwarmComputeLocalSize(), DMSwarmComputeLocalSizeFromOptions(), DMSwarmInitializeCoordinates(), DMSwarmInitializeVelocities(), DMSwarmInitializeVelocitiesFromOptions() to assist initialization of PIC methods

DMPlex:

• Add DMExtrude() which now the default extrusion

• Change DMPlexExtrude() to use DMPlexTransform underneath

• Add DMPlexTransformExtrudeSetNormalFunction() to allow computed normals

• Add DMGetNaturalSF() and DMSetNaturalSF()

• Change -dm_plex_csr_via_mat to -dm_plex_csr_alg which takes a DMPlexCSRAlgorithm name

• Add public API for metric-based mesh adaptation:

  • Move DMPlexMetricCtx from public to private and give it to DMPlex

  • Add DMPlexMetricSetFromOptions() to assign values to DMPlexMetricCtx

  • Add DMPlexMetricSetIsotropic() for declaring whether a metric is isotropic

  • Add DMPlexMetricIsIsotropic() for determining whether a metric is isotropic

  • Add DMPlexMetricSetUniform() for declaring whether a metric is uniform

  • Add DMPlexMetricIsUniform() for determining whether a metric is uniform

  • Add DMPlexMetricSetRestrictAnisotropyFirst() for declaring whether anisotropy should be restricted before normalization

  • Add DMPlexMetricRestrictAnisotropyFirst() for determining whether anisotropy should be restricted before normalization

  • Add DMPlexMetricSetMinimumMagnitude() for specifying the minimum tolerated metric magnitude

  • Add DMPlexMetricGetMinimumMagnitude() for retrieving the minimum tolerated metric magnitude

  • Add DMPlexMetricSetMaximumMagnitude() for specifying the maximum tolerated metric magnitude

  • Add DMPlexMetricGetMaximumMagnitude() for retrieving the maximum tolerated metric magnitude

  • Add DMPlexMetricSetMaximumAnisotropy() for specifying the maximum tolerated metric anisostropy

  • Add DMPlexMetricGetMaximumAnisotropy() for retrieving the maximum tolerated metric anisotropy

  • Add DMPlexMetricSetTargetComplexity() for specifying the target metric complexity

  • Add DMPlexMetricGetTargetComplexity() for retrieving the target metric complexity

  • Add DMPlexMetricSetNormalizationOrder() for specifying the order of L-p normalization

  • Add DMPlexMetricGetNormalizationOrder() for retrieving the order of L-p normalization

  • Change DMPlexMetricCtx so that it is only instantiated when one of the above routines are called

  • Change DMPlexMetricEnforceSPD() to have more arguments:

    • control whether anisotropy is restricted

    • output the modified metric, rather than modifying the input

    • output the determinant

  • Change DMPlexMetricNormalize() to have another argument, for controlling whether anisotropy is restricted

• Change DMAdaptor so that its -adaptor_refinement_h_min/h_max/a_max/p command line arguments become -dm_plex_metric_h_min/h_max/a_max/p

• Add 2D and 3D mesh adaptation interface to Mmg and 3D mesh adaptation interface to ParMmg. Mmg/ParMmg specific changes:

  • Change DMPlexBuildFromCellListParallel() to have another argument, for the connectivity

  • Change DMPlexCreateFromCellListParallelPetsc() to have another argument, for the connectivity

  • Add DMPlexMetricSetVerbosity() for setting the verbosity of the metric-based mesh adaptation package

  • Add DMPlexMetricGetVerbosity() for getting the verbosity of the metric-based mesh adaptation package

  • Add DMPlexMetricSetNoInsertion() to turn off node insertion and deletion for (Par)Mmg

  • Add DMPlexMetricNoInsertion() to determine whether node insertion and deletion are turned off for (Par)Mmg

  • Add DMPlexMetricSetNoSwapping() to turn off facet swapping for (Par)Mmg

  • Add DMPlexMetricNoSwapping() to determine whether facet swapping is turned off for (Par)Mmg

  • Add DMPlexMetricSetNoMovement() to turn off node movement for (Par)Mmg

  • Add DMPlexMetricNoMovement() to determine whether node movement is turned off for (Par)Mmg

  • Add DMPlexMetricSetNoSurf() to turn off surface modification for (Par)Mmg

  • Add DMPlexMetricNoSurf() to determine whether surface modification is turned off for (Par)Mmg

  • Add DMPlexMetricSetGradationFactor() to set the metric gradation factor

  • Add DMPlexMetricGetGradationFactor() to get the metric gradation factor

  • Add DMPlexMetricSetHausdorffNumber() to set the metric Hausdorff number

  • Add DMPlexMetricGetHausdorffNumber() to get the metric Hausdorff number

  • Add DMPlexMetricSetNumIterations() to set the number of ParMmg adaptation iterations

  • Add DMPlexMetricGetNumIterations() to get the number of ParMmg adaptation iterations

• Change DMPlexCoordinatesLoad() to take a PetscSF as argument

• Change DMPlexLabelsLoad() to take the PetscSF argument and load in parallel

• Change DMPlexCreateFromFile() to take the mesh name as argument

• Change DMAdaptMetric so that it takes an additional argument for cell tags

• Change DMTransformAdaptLabel so that it takes an additional argument for cell tags

• Change DMGenerateRegister so that it registers routines that take an additional argument for cell tags

• Change DMPlexFindVertices() to take Vec and IS arguments instead of arrays

• Add DMPlexTSComputeRHSFunctionFEM() to support explicit timestepping

• Newly created DMPlex will be distributed by default; this previously required -dm_distribute or explicit calls to DMPlexDistribute()

• Add DMPlexDistributeGetDefault() and DMPlexDistributeSetDefault() to determine and set the default for DMPlex distribution

• Add meshing of the Schwarz-P and Gyroid triply periodic minimal surface (see DMPlexCreateTPSMesh()). These meshes can be automatically generated using -dm_plex_shape schwarz_p or -dm_plex_shape gyroid, with optional levels of refinement and extrusion to 3D solids with prescribed thickness.

• Add DMCreateFEDefault() as a convenience method for creating the right element on a mesh

• Add DMPlexCreateReferenceCell()

• Remove deprecated DMPlexCreateFromCellList() and DMPlexCreateFromCellListParallel()

• Add DMSetMatrixPreallocateSkip() to save initialization time when MatSetPreallocationCOO() will be used.

• Improve loading performance related to coordinate projection in common cases.

• Add DMPlexGetOrdering1D() for 1D Plex problems

• Add DMPlexComputeClementInterpolant() averaging operator

DMPlexLandau:

• Add DMPlexLandauCreateVelocitySpace() Create DMComposite of DMPlex for Landau collision operator

• Add DMPlexLandauDestroyVelocitySpace() Destroy DMComposite of DMPlex for Landau collision operator

• Add DMPlexLandauIFunction() Landau collision operator

• Add DMPlexLandauIJacobian() Landau collision operator

FE/FV:

• Deprecate PetscSpacePolynomialGetSymmetric() and PetscSpacePolynomialSetSymmetric(): symmetric polynomials were never supported and support is no longer planned

• Remove PetscSpacePolynomialType enum and associated array of strings PetscSpacePolynomialTypes: other polynomial spaces are now handled by other implementations of PetscSpace

• Add PETSCSPACEPTRIMMED that implements trimmed polynomial spaces (also known as the spaces in Nedelec face / edge elements of the first kind)

• Replace PetscDSGet/SetHybrid() with PetscDSGet/SetCohesive()

• Add PetscDSIsCohesive(), PetscDSGetNumCohesive(), and PetscDSGetFieldOffsetCohesive()

• Add argument to PetscFEIntegrateHybridJacobian() to indicate the face for the integration

• Add PetscFECreateByCell() and PetscFECreateLagrangeByCell() to create FE spaces on specific cell types

• Replace PetscDualSpaceCreateReferenceCell() with DMPlexCreateReferenceCell()

• Add PetscDualSpaceEqual() and PetscQuadratureEqual()

DMNetwork:

• DMNetworkAddComponent() now requires a valid component key for each call

• Add DMNetworkSharedVertexGetInfo()

DMStag:

DT:

• Add PetscDTPTrimmedEvalJet() to evaluate a stable basis for trimmed polynomials, and PetscDTPTrimmedSize() for the size of that space

• Add PetscDSGetRHSResidual() and PetscDSSetRHSResidual() to support explicit timestepping

• Add PetscDTTensorQuadratureCreate() to combine different quadratures, such as on a prism

• Add PetscProbComputeKSStatistic() to apply the Kolmogorov-Smirnov test

• Add probability distributions PetscPDFMaxwellBoltzmann1D(), PetscCDFMaxwellBoltzmann1D(), PetscPDFMaxwellBoltzmann2D(), PetscCDFMaxwellBoltzmann2D(), PetscPDFMaxwellBoltzmann3D(), PetscCDFMaxwellBoltzmann3D(), PetscPDFGaussian1D(), PetscCDFGaussian1D(), PetscPDFSampleGaussian1D(), PetscPDFGaussian2D(), PetscPDFSampleGaussian2D(), PetscPDFConstant1D(), PetscCDFConstant1D()

Fortran:

Logging:

• Add PetscLogIsActive() to determine if logging is in progress or not