Releases: parthenon-hpc-lab/parthenon
Releases · parthenon-hpc-lab/parthenon
Regression test gold standard v25
Updates the output_hdf5
test to include a Metadata::None
variable.
v24.08
Release 24.08
Date: 2024-08-30
Added (new features/APIs/variables/...)
- [PR 1151] Add time offset
c
to LowStorageIntegrator - [PR 1147] Add
par_reduce_inner
functions - [PR 1159] Add additional timestep controllers in parthenon/time.
- [PR 1148] Add
GetPackDimension
toStateDescriptor
for calculating pack sizes beforeMesh
initialization - [PR 1143] Add tensor indices to VariableState, add radiation constant to constants, add TypeLists, allow for arbitrary containers for solvers
- [PR 1140] Allow for relative convergence tolerance in BiCGSTAB solver.
- [PR 1047] General three- and four-valent 2D forests w/ arbitrary orientations.
- [PR 1130] Enable
parthenon::par_reduce
for MD loops with Kokkos 1D Range - [PR 1119] Formalize MeshData partitioning.
- [PR 1128] Add cycle and nbtotal to hst
- [PR 1099] Functionality for outputting task graphs in GraphViz format.
- [PR 1091] Add vector wave equation example.
- [PR 991] Add fine fields.
- [PR 1106] Add CMake options for turning on ASAN and HWASAN
- [PR 1100] Custom refinement ops propagated to fluxes
- [PR 1090] SMR with swarms
- [PR 1079] Address XDMF/Visit Issues
- [PR 1084] Properly free swarm boundary MPI requests
- [PR 1020] Add bi- and trilinear interpolation routines
- [PR 1081] Add GetSize and GetIndex to sparse pack
- [PR 1026] Particle BCs without relocatable device code
- [PR 1037] Add SwarmPacks
- [PR 1068] Add ability to dump sparse pack contents as a string
- [PR 1062] UserWorkBeforeRestartOutput
- [PR 1043] Unify flux correction with boundary communication, make fluxes variables, allow fluxes for non-cell fields
- [PR 1060] Add the ability to request new MeshData/MeshBlockData objects by selecting variables by UID.
- [PR 1039] Add ability to output custom coordinate positions for Visit/Paraview
- [PR 1019] Enable output for non-cell-centered variables
Changed (changing behavior/API/variables/...)
- [PR 1153] Allow base grid with fewer blocks than ranks before initial AMR
- [PR 1105] Refactor parameter input for linear solvers
- [PR 1078] Add reduction fallback in 1D. Add IndexRange overload for 1D par loops
- [PR 1024] Add .outN. to history output filenames
- [PR 1004] Allow parameter modification from an input file for restarts
Fixed (not changing behavior/API/variables/...)
- [PR 1145] Fix remaining swarm D->H->D copies
- [PR 1150] Reduce memory consumption for buffer pool
- [PR 1146] Fix an issue outputting >4GB single variables per rank
- [PR 1152] Fix memory leak in task graph outputs related to
abi::__cxa_demangle
- [PR 1146] Fix an issue outputting >4GB single variables per rank
- [PR 1144] Fix some restarts w/non-CC fields
- [PR 1132] Fix regional dependencies for iterative task lists and make solvers work for arbirtrary MeshData partitioning
- [PR 1139] only add --expt-relaxed-constexpr for COMPILE_LANGUAGE:CXX
- [PR 1131] Make deallocation of fine and sparse fields work
- [PR 1127] Add WithFluxes to IsRefined check
- [PR 1111] Fix undefined behavior due to bitshift of negative number in LogicalLocation
- [PR 1092] Updates to DataCollection and MeshData to remove requirement of predefining MeshBlockData
- [PR 1113] Prevent division by zero
- [PR 1112] Remove shared_ptr cycle in forest::Tree
- [PR 1104] Fix reading restarts due to hidden ghost var
- [PR 1098] Move to symmetrized logical coordinates and fix SMR bug
- [PR 1095] Add missing include guards in hdf5 restart
- [PR 1093] Fix forest size for symmetry dimensions
- [PR 1089] Fix loading restart files without derefinement counter
- [PR 1079] Address XDMF/Visit Issues
- [PR 1088] Correctly fill fluxes for non-cell variables in SparsePacks
- [PR 1083] Correctly fill VariableFluxPack for edge fluxes in 2D
- [PR 1087] Make sure InnerLoopPatternTVR is resolved on device properly when it is the default loop pattern
- [PR 1071] Fix bug in static mesh refinement related to redefinition of Mesh::root_level
- [PR 1073] Fix bug in AMR and sparse restarts
- [PR 1070] Correctly exclude flux vars from searches by default
- [PR 1049] Catch task failures from threads
- [PR 1058] Vector history not being output if no scalar history present
- [PR 1057] Fix history output after restarts
- [PR 1053] Set the correct root level on restart
- [PR 1024] Add features to history output
- [PR 1031] Fix bug in non-cell centered AMR
Infrastructure (changes irrelevant to downstream codes)
- [PR 1117] Enable CI pipelines on AMD GPUs with ROCM/HIP
- [PR 1114] Enable sanitizers for extended CI host build
- [PR 1123] Default initialize ProResInfo.dir
- [PR 1121] Default initialize BndInfo.dir
- [PR 1116] Fix NumPy 2.0 test script breakage
- [PR 1055] Refactor mesh constructors
- [PR 1066] Re-introduce default loop patterns and exec spaces
- [PR 1064] Forbid erroneous edge case when adding MeshData on a partition
- [PR 1035] Fix multigrid infrastructure to work with forest
- [PR 1048] Tiny fixes to custom coords logic
- [PR 1028] Internal reorganization of LogicalLocation files
- [PR 1009] Move from a single octree to a forest of octrees
Removed (removing behavior/API/varaibles/...)
- [PR 1108] Remove NaN payload tags infrastructure
Incompatibilities (i.e. breaking changes)
- [[PR 1135]](http...
Regression test gold standard v24
Update the boundary exchange gold standard file.
Regression test gold standard v23
Regression test gold standard v22
Update gold files for sparse_advection to account for the fixed shape of the sparse scalar fields.
Release 24.03
Release 24.03
Date: 2024-03-21
Added (new features/APIs/variables/...)
- [PR 852] Add Mesh version of UserWorkBeforeOutput
- [PR 998] tensor indices added to sparse pack
- [PR 999] Add a post-initialization hook
- [PR 987] New tasking infrastructure and capabilities
- [PR 969] New macro-based auto-naming of profiling regions and kernels
- [PR 981] Add IndexSplit
- [PR 983] Add Contains to SparsePack
- [PR 968] Add per package registration of boundary conditions
- [PR 948] Add solver interface and update Poisson geometric multi-grid example
- [PR 996] Remove dynamic allocations from swarm particle creation
Changed (changing behavior/API/variables/...)
- [PR 973] Multigrid performance upgrades
Fixed (not changing behavior/API/variables/...)
- [PR1023] Fix broken param of a scalar bool
- [PR1012] Remove accidentally duplicated code
- [PR992] Allow custom PR ops with sparse pools
- [PR988] Fix bug in neighbor finding routine for small, periodic, refined meshes
- [PR986] Fix bug in sparse boundary communication BndInfo cacheing
- [PR978] remove erroneous sparse check
Infrastructure (changes irrelevant to downstream codes)
- [PR 1017] Make regression tests more verbose on failure
- [PR 1007] Split template instantiations for HDF5 Read/Write attributes to speed up compile times
- [PR 990] Partial refactor of HDF5 I/O code for readability/extendability
- [PR 982] add some gut check testing for parthenon-VIBE
Incompatibilities (i.e. breaking changes)
- [PR 987] Change the API for what was IterativeTasks
- [PR 974] Change GetParentPointer to always return T*
- [PR 996] Remove dynamic allocations from swarm particle creation
New Contributors
Full Changelog: v23.11...v24.03
Regression test gold standard v21
Update gold files for advection_outflow to include derived var produced by UserMeshWorkBeforeOutput
Release 23.11
Release 23.11
Date: 2023-11-16
Added (new features/APIs/variables/...)
- [PR 962] Add support for in-situ histograms/profiles
- [PR 911] Add infrastructure for geometric multi-grid
- [PR 971] Add UserWorkBeforeLoop
- [PR 907] PEP1: Allow subclassing StateDescriptor
- [PR 932] Add GetOrAddFlag to metadata
- [PR 931] Allow SparsePacks with subsets of blocks
- [PR 921] Add more flexible ways of adding and using MeshData/MeshBlockData objects to DataCollections
- [PR 900] Add Morton numbers and expand functionality of LogicalLocation
- [PR 902] Add ability to output NaNs for de-allocated sparse fields
- [PR 887] Add ability to dump more types of params and read them from restarts
- [PR 884] Add constant derivative BC and expose GenericBC
- [PR 892] Cost-based load balancing and memory diagnostics
- [PR 889] Add PreCommFillDerived
- [PR 872] Boundary communication for non-cell centered fields
- [PR 877] Add flat sparse packs
- [PR 868] Add block-local face, edge, and nodal fields and allow for packing
Changed (changing behavior/API/variables/...)
- [PR 977] Change to CalVer from SemVer
- [PR 975] Construct staged integrators via arbitrary name
- [PR 976] Move UserWorkBeforeLoop to be after first output
- [PR 965] Allow leading whitespace in input parameters
- [PR 926] Internal refinement op registration
- [PR 897] Deflate compression filter is not called any more if compression is soft disabled
- [PR 896] Update Kokkos integration to support installed version. Use
serial
(flat MPI) host parallelization by default (instead of OpenMP) - [PR 894] Demand that sparse pool order sparse ids
- [PR 888] Bump Kokkos submodule to 4.0.1
- [PR 885] Expose PackDescriptor and use uids in SparsePacks
Fixed (not changing behavior/API/variables/...)
- [PR 955] Only permit rank0 to mkdir when -d flag specified
- [PR 952] Fix format string in sparse advection example
- [PR 947] Add missing ForceRemeshComm dependencies
- [PR 928] Fix boundary comms during refinement next to refined blocks
- [PR 937] Fix multiple line continuations
- [PR 933] Remove extraneous debug check
- [PR 917] Update Iterative Tasking Infrastructure
- [PR 890] Fix bugs in sparse communication and prolongation
Infrastructure (changes irrelevant to downstream codes)
- [PR 967] Change INLINE to FORCEINLINE on par_for_inner overloads
- [PR 938] Restructure buffer packing/unpacking kernel hierarchical parallelism
- [PR 944] Move sparse pack identifier creation to descriptor
- [PR 904] Move to prolongation/restriction in one for AMR and communicate non-cell centered fields
- [PR 918] Refactor RegionSize
- [PR 901] Implement shared element ownership model
Removed (removing behavior/API/varaibles/...)
- [PR 930 Remove ParthenonManager::ParthenonInit as it is error-prone and the split functions are the recommended usage.
Regression test gold standard v20
Upgolded to account for fixed behavior when outputting string parameters.
Regression test gold standard v19
Revert true sparse gold file to version from v16 after fixing bugs that were introduced into the gold file in v17.