1# Changes/Release Notes 2 3On this page we provide a summary of the main API changes, new features and examples for each release of libCEED. 4 5(main)= 6 7## Current `main` branch 8 9### Interface changes 10 11- Add `bool` field type for `CeedQFunctionContext` and related interfaces to use `bool` fields. 12- `CEED_BASIS_COLLOCATED` removed; users should only use `CEED_BASIS_NONE`. 13- Remove unneeded pointer for `CeedElemRestrictionGetELayout`. 14 15### New features 16 17- Add `CeedOperatorCreateAtPoints` which evaluates the `CeedQFunction` at arbitrary locations in each element, for use in Particle in Cell, Material Point Method, and similar methods. 18- Add `CeedElemRestrictionGetLLayout` to provide L-vector layout for strided `CeedElemRestriction` created with `CEED_BACKEND_STRIDES`. 19- Add `CeedVectorReturnCeed` and similar when parent `Ceed` context for a libCEED object is only needed once in a calling scope. 20- Enable `#pragma once` for JiT source; remove duplicate includes in JiT source string before compilation. 21 22### Examples 23 24- Add deal.II example with CEED BP suite. 25 26(v0-12)= 27 28## v0.12 (Oct 31, 2023) 29 30### Interface changes 31 32- Update `CeedOperatorContext*` functions to `CeedOperator*Context*` functions for consistency. 33For example, `CeedOperatorContextGetFieldLabel` was renamed to `CeedOperatorGetContextFieldLabel`. 34- Removed `CeedBasisSetNumQuadraturePoints` as redundant and bug-prone interface. 35 36### New features 37 38- Added {c:func}`CeedOperatorGetFieldByName` to access a specific `CeedOperatorField` by its name. 39- Update `/cpu/self/memcheck/*` backends to help verify `CeedVector` array access assumptions and `CeedQFunction` user output assumptions. 40- Update {c:func}`CeedOperatorLinearAssembleDiagonal` to provide default implementation that supports `CeedOperator` with multiple active bases. 41- Added Sycl backends `/gpu/sycl/ref`, `/gpu/sycl/shared`, and `/gpu/sycl/gen`. 42- Added {c:func}`CeedBasisApplyAtPoints` for evaluation of values and derivatives at arbitrary points inside elements. 43- Added support for non-tensor $H(\text{curl})$ finite element spaces with {c:func}`CeedBasisCreateHcurl`. 44- Added {c:func}`CeedElemRestrictionCreateCurlOriented`, similar to {c:func}`CeedElemRestrictionCreateOriented`, for element restrictions requiring more general element transformations such as those for high-order $H(\text{curl})$ spaces on tetrahedra (see [https://dl.acm.org/doi/pdf/10.1145/3524456](https://dl.acm.org/doi/pdf/10.1145/3524456)). 45- Added {c:func}`CeedOperatorLinearAssemblePointBlockDiagonalSymbolic` to create COO mapping for mapping out of {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`. 46- Added support for application codes which manage multiple {ref}`Ceed` objects, parallelized across OpenMP threads. 47 48### Examples 49 50- Add `DMSwarm` example demonstrating interpolation from background mesh to swarm points and projection from swarm points to background mesh. 51 52#### {ref}`example-petsc-bps` 53 54- Requires PETSc version 3.19 or later. 55 56#### {ref}`example-petsc-navier-stokes` 57 58- Updated restart and checkpointing interface. 59- Add data-driven subgrid-stress model. 60- Add differential filtering of solution. 61- Add turbulence statistics collection over spanwise-symmetric geometries. 62- Add Taylor-Green vortex initial condition. 63- Add Riemann-based outflow boundary conditions. 64- Added vortex shedding and flow past cylinder example, including calculations for lift, drag, and heat transfer. 65- Add Internal Damping Layer (IDL) for helping turbulent simulation stability. 66- Derive `CeedBasis` from `PetscFE`, and various other internal maintainability updates. 67 68(v0-11)= 69 70## v0.11 (Dec 24, 2022) 71 72### Interface changes 73 74- Added {c:func}`CeedOperatorSetName` for more readable {c:func}`CeedOperatorView` output. 75- Added {c:func}`CeedBasisCreateProjection` to facilitate interpolation between nodes for separate `CeedBases`. 76- Rename and move {c:func}`CeedCompositeOperatorGetNumSub` and {c:func}`CeedCompositeOperatorGetSubList` to public interface. 77- Renamed `CEED_BASIS_COLLOCATED` to `CEED_BASIS_NONE` for clarity. 78Some users previously misinterpreted a `CeedOperator` field using `CEED_BASIS_COLLOCATED` as meaning that the entire `CeedOperator` used a quadrature space that is collocated with the nodal space of the active bases. 79 80### New features 81 82- Update `/cpu/self/memcheck/*` backends to help verify `CeedQFunctionContext` data sizes provided by user. 83- Improved support for $H(\text{div})$ bases. 84- Added `CeedInt_FMT` to support potential future use of larger integer sizes. 85- Added `CEED_QFUNCTION_ATTR` for setting compiler attributes/pragmas to `CEED_QFUNCTION_HELPER` and `CEED_QFUNCTION`. 86- OCCA backend updated to latest OCCA release; DPC++ and OMP OCCA modes enabled. 87Due to a limitation of the OCCA parser, typedefs are required to use pointers to arrays in QFunctions with the OCCA backend. 88This issue will be fixed in a future OCCA release. 89 90### Bugfix 91 92- Fix bug in setting device id for GPU backends. 93- Fix storing of indices for `CeedElemRestriction` on the host with GPU backends. 94- Fix `CeedElemRestriction` sizing for {c:func}`CeedOperatorAssemblePointBlockDiagonal`. 95- Fix bugs in CPU implementation of {c:func}`CeedOperatorLinearAssemble` when there are different number of active input modes and active output modes. 96 97### Examples 98 99#### {ref}`example-petsc-navier-stokes` 100 101- Various performance enhancements, analytic matrix-free and assembled Jacobian, and PETSc solver configurations for GPUs. 102- Refactored to improve code reuse and modularity. 103- Support for primitive variables for more accurate boundary layers and all-speed flow. 104- Added $YZ\beta$ shock capturing scheme and Shock Tube example. 105- Added Channel example, with comparison to analytic solutions. 106- Added Flat Plate with boundary layer mesh and compressible Blasius inflow condition based on Chebyshev collocation solution of the Blasius equations. 107- Added strong and weak synthetic turbulence generation (STG) inflow boundary conditions. 108- Added "freestream" boundary conditions based on HLLC Riemann solver. 109- Automated stabilization coefficients for different basis degree. 110 111#### {ref}`example-petsc-bps` 112 113- Support for convergence studies. 114 115### Maintainability 116 117- Refactored `/gpu/cuda/shared` and `/gpu/cuda/gen` as well as `/gpu/hip/shared` and `/gpu/hip/gen` backend to improve maintainablity and reduce duplicated code. 118- Enabled support for `p > 8` for `/gpu/*/shared` backends. 119- Switch to `clang-format` over `astyle` for automatic formatting; Makefile command changed to `make format` from `make style`. 120- Improved test harness. 121 122(v0-10-1)= 123 124## v0.10.1 (Apr 11, 2022) 125 126### Interface changes 127 128- Added {c:func}`CeedQFunctionSetUserFlopsEstimate` and {c:func}`CeedOperatorGetFlopsEstimate` to facilitate estimating FLOPs in operator application. 129 130### New features 131 132- Switched MAGMA backends to use runtime compilation for tensor basis kernels (and element restriction kernels, in non-deterministic `/gpu/*/magma` backends). 133This reduces time to compile the library and increases the range of parameters for which the MAGMA tensor basis kernels will work. 134 135### Bugfix 136 137- Install JiT source files in install directory to fix GPU functionality for installed libCEED. 138 139(v0-10)= 140 141## v0.10 (Mar 21, 2022) 142 143### Interface changes 144 145- Update {c:func}`CeedQFunctionGetFields` and {c:func}`CeedOperatorGetFields` to include number of fields. 146- Promote to the public API: QFunction and Operator field objects, `CeedQFunctionField` and `CeedOperatorField`, and associated getters, {c:func}`CeedQFunctionGetFields`; {c:func}`CeedQFunctionFieldGetName`; {c:func}`CeedQFunctionFieldGetSize`; {c:func}`CeedQFunctionFieldGetEvalMode`; {c:func}`CeedOperatorGetFields`; {c:func}`CeedOperatorFieldGetElemRestriction`; {c:func}`CeedOperatorFieldGetBasis`; and {c:func}`CeedOperatorFieldGetVector`. 147- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added. 148- Add {c:func}`CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients. 149- Promote {c:func}`CeedOperatorCheckReady`to the public API to facilitate interactive interfaces. 150- Warning added when compiling OCCA backend to alert users that this backend is experimental. 151- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`. 152- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name. 153- Clarify documentation for {c:func}`CeedVectorTakeArray`; this function will error if {c:func}`CeedVectorSetArray` with `copy_mode == CEED_USE_POINTER` was not previously called for the corresponding `CeedMemType`. 154- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitialized arrays; require initialized data for {c:func}`CeedVectorGetArray`. 155- Added {c:func}`CeedQFunctionContextRegisterDouble` and {c:func}`CeedQFunctionContextRegisterInt32` with {c:func}`CeedQFunctionContextSetDouble` and {c:func}`CeedQFunctionContextSetInt32` to facilitate easy updating of {c:struct}`CeedQFunctionContext` data by user defined field names. 156- Added {c:func}`CeedQFunctionContextGetFieldDescriptions` to retrieve user defined descriptions of fields that are registered with `CeedQFunctionContextRegister*`. 157- Renamed `CeedElemTopology` entries for clearer namespacing between libCEED enums. 158- Added type `CeedSize` equivalent to `ptrdiff_t` for array sizes in {c:func}`CeedVectorCreate`, {c:func}`CeedVectorGetLength`, `CeedElemRestrictionCreate*`, {c:func}`CeedElemRestrictionGetLVectorSize`, and {c:func}`CeedOperatorLinearAssembleSymbolic`. This is a breaking change. 159- Added {c:func}`CeedOperatorSetQFunctionUpdated` to facilitate QFunction data re-use between operators sharing the same quadrature space, such as in a multigrid hierarchy. 160- Added {c:func}`CeedOperatorGetActiveVectorLengths` to get shape of CeedOperator. 161 162### New features 163 164- `CeedScalar` can now be set as `float` or `double` at compile time. 165- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends. 166- Added support for JiT of QFunctions with `#include "relative/path/local-file.h"` statements for additional local files. Note that files included with `""` are searched relative to the current file first, then by compiler paths (as with `<>` includes). To use this feature, one should adhere to relative paths only, not compiler flags like `-I`, which the JiT will not be aware of. 167- Remove need to guard library headers in QFunction source for code generation backends. 168- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present. 169- Added {c:func}`CeedStringAllocCopy` to reduce repeated code for copying strings internally. 170- Added {c:func}`CeedPathConcatenate` to facilitate loading kernel source files with a path relative to the current file. 171- Added support for non-tensor $H(\text{div})$ elements, to include CPU backend implementations and {c:func}`CeedBasisCreateHdiv` convenience constructor. 172- Added {c:func}`CeedQFunctionSetContextWritable` and read-only access to `CeedQFunctionContext` data as an optional feature to improve GPU performance. By default, calling the `CeedQFunctionUser` during {c:func}`CeedQFunctionApply` is assumed to write into the `CeedQFunctionContext` data, consistent with the previous behavior. Note that if a user asserts that their `CeedQFunctionUser` does not write into the `CeedQFunctionContext` data, they are responsible for the validity of this assertion. 173- Added support for element matrix assembly in GPU backends. 174 175### Maintainability 176 177- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage. 178- `Include-what-you-use` makefile target added as `make iwyu`. 179- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase. 180- Put GPU JiTed kernel source code into separate files. 181- Dropped legacy version support in PETSc based examples to better utilize PETSc DMPlex and Mat updates to support libCEED; current minimum PETSc version for the examples is v3.17. 182 183(v0-9)= 184 185## v0.9 (Jul 6, 2021) 186 187### Interface changes 188 189- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact. 190 191### New features 192 193- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use. 194- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends. 195- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations. 196- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed). 197 198### Performance improvements 199 200### Examples 201 202- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations. 203- Fluid mechanics example adds GPU support and improves modularity. 204 205### Deprecated backends 206 207- The `/cpu/self/tmpl` and `/cpu/self/tmpl/sub` backends have been removed. These backends were intially added to test the backend inheritance mechanism, but this mechanism is now widely used and tested in multiple backends. 208 209(v0-8)= 210 211## v0.8 (Mar 31, 2021) 212 213### Interface changes 214 215- Error handling improved to include enumerated error codes for C interface return values. 216- Installed headers that will follow semantic versioning were moved to {code}`include/ceed` directory. These headers have been renamed from {code}`ceed-*.h` to {code}`ceed/*.h`. Placeholder headers with the old naming schema are currently provided, but these headers will be removed in the libCEED v0.9 release. 217 218### New features 219 220- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features. 221- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly. 222- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators. 223 224### Performance improvements 225 226- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`. 227- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`. 228 229### Examples 230 231- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions. 232- {ref}`example-petsc-elasticity` example updated with Neo-Hookean hyperelasticity in current configuration as well as improved Neo-Hookean hyperelasticity exploring storage vs computation tradeoffs. 233- {ref}`example-petsc-navier-stokes` example updated with isentropic traveling vortex test case, an analytical solution to the Euler equations that is useful for testing boundary conditions, discretization stability, and order of accuracy. 234- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree. 235 236(v0-7)= 237 238## v0.7 (Sep 29, 2020) 239 240### Interface changes 241 242- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors. 243 As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`. 244 These changes improve support for mixed finite element methods. 245- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature. 246- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity. 247- Linear Operators can be assembled as point-block diagonal matrices with {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`, provided in row-major form in a {code}`ncomp` by {code}`ncomp` block per node. 248- Diagonal assemble interface changed to accept a {ref}`CeedVector` instead of a pointer to a {ref}`CeedVector` to reduce memory movement when interfacing with calling code. 249- Added {c:func}`CeedOperatorLinearAssembleAddDiagonal` and {c:func}`CeedOperatorLinearAssembleAddPointBlockDiagonal` for improved future integration with codes such as MFEM that compose the action of {ref}`CeedOperator`s external to libCEED. 250- Added {c:func}`CeedVectorTakeAray` to sync and remove libCEED read/write access to an allocated array and pass ownership of the array to the caller. 251 This function is recommended over {c:func}`CeedVectorSyncArray` when the {code}`CeedVector` has an array owned by the caller that was set by {c:func}`CeedVectorSetArray`. 252- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory. 253- Added {c:func}`CeedOperatorMultigridLevelCreate`, {c:func}`CeedOperatorMultigridLevelCreateTensorH1`, and {c:func}`CeedOperatorMultigridLevelCreateH1` to facilitate creation of multigrid prolongation, restriction, and coarse grid operators using a common quadrature space. 254 255### New features 256 257- New HIP backend: `/gpu/hip/ref`. 258- CeedQFunction support for user `CUfunction`s in some backends 259 260### Performance improvements 261 262- OCCA backend rebuilt to facilitate future performance enhancements. 263- PETSc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`. 264 265### Examples 266 267- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions. 268 269### Deprecated backends 270 271- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`. 272 273(v0-6)= 274 275## v0.6 (Mar 29, 2020) 276 277libCEED v0.6 contains numerous new features and examples, as well as expanded 278documentation in [this new website](https://libceed.org). 279 280### New features 281 282- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly 283 1-1 correspondence with the C interface, plus some convenience features. For instance, 284 data stored in the {cpp:type}`CeedVector` structure are available without copy as 285 {py:class}`numpy.ndarray`. Short tutorials are provided in 286 [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/). 287- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point, 288 {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal 289 ({c:func}`CeedOperatorAssembleLinearDiagonal`). These operations are useful for 290 preconditioning ingredients and are used in the libCEED's multigrid examples. 291- The inverse of separable operators can be obtained using 292 {c:func}`CeedOperatorCreateFDMElementInverse` and applied with 293 {c:func}`CeedOperatorApply`. This is a useful preconditioning ingredient, 294 especially for Laplacians and related operators. 295- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`, 296 {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`. 297- Make public accessors for various attributes to facilitate writing composable code. 298- New backend: `/cpu/self/memcheck/serial`. 299- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA 300 backends. (Single source is coming soon for OCCA backends.) 301- Fix some missing edge cases in CUDA backend. 302 303### Performance Improvements 304 305- MAGMA backend performance optimization and non-tensor bases. 306- No-copy optimization in {c:func}`CeedOperatorApply`. 307 308### Interface changes 309 310- Replace {code}`CeedElemRestrictionCreateIdentity` and 311 {code}`CeedElemRestrictionCreateBlocked` with more flexible 312 {c:func}`CeedElemRestrictionCreateStrided` and 313 {c:func}`CeedElemRestrictionCreateBlockedStrided`. 314- Add arguments to {c:func}`CeedQFunctionCreateIdentity`. 315- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification 316 with {cpp:enum}`CeedInterlaceMode`. This is now an attribute of the 317 {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no 318 longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and 319 {c:func}`CeedElemRestrictionApply`. 320 321### Examples 322 323libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`. 324Notable additions include: 325 326- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of 327 a domain in 1, 2, and 3 dimensions by applying a Laplacian. 328 329- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area 330 of domains (like the cube and sphere) by direct integration on a surface mesh; 331 demonstrates geometric dimension different from topological dimension. 332 333- PETSc {ref}`example-petsc-bps`: 334 335 - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support. 336 - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements 337 and transparent CUDA support. 338 - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`): 339 generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric 340 dimension different from topological dimension. 341 342- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid 343 solver with algebraic multigrid coarse solve. 344 345- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly 346 `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`), 347 implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and 348 quasi-2D computational domain support. 349 350- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for 351 linear elasticity, small-strain hyperelasticity, and globalized finite-strain 352 hyperelasticity using p-multigrid with algebraic multigrid coarse solve. 353 354(v0-5)= 355 356## v0.5 (Sep 18, 2019) 357 358For this release, several improvements were made. Two new CUDA backends were added to 359the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art 360performance using single-source {ref}`CeedQFunction`. From this release, users 361can define Q-Functions in a single source code independently of the targeted backend 362with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU 363compilation of the user provided {ref}`CeedQFunction` code. To allow a unified 364declaration, the {ref}`CeedQFunction` API has undergone a slight change: 365the `QFunctionField` parameter `ncomp` has been changed to `size`. This change 366requires setting the previous value of `ncomp` to `ncomp*dim` when adding a 367`QFunctionField` with eval mode `CEED EVAL GRAD`. 368 369Additionally, new CPU backends 370were included in this release, such as the `/cpu/self/opt/*` backends (which are 371written in pure C and use partial **E-vectors** to improve performance) and the 372`/cpu/self/ref/memcheck` backend (which relies upon the 373[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user 374{ref}`CeedQFunction` have no undefined values). 375This release also included various performance improvements, bug fixes, new examples, 376and improved tests. Among these improvements, vectorized instructions for 377{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD` 378instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and 379identity Q-Functions were introduced, and the PETSc benchmark problems were expanded 380to include unstructured meshes handling were. For this expansion, the prior version of 381the PETSc BPs, which only included data associated with structured geometries, were 382renamed `bpsraw`, and the new version of the BPs, which can handle data associated 383with any unstructured geometry, were called `bps`. Additionally, other benchmark 384problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively), 385and BP5 and BP6 (the collocated versions---for which the quadrature points are the same 386as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc 387examples. Furthermoew, another standalone libCEED example, called `ex2`, which 388computes the surface area of a given mesh was added to this release. 389 390Backends available in this release: 391 392| CEED resource (`-ceed`) | Backend | 393|--------------------------|-----------------------------------------------------| 394| `/cpu/self/ref/serial` | Serial reference implementation | 395| `/cpu/self/ref/blocked` | Blocked reference implementation | 396| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks | 397| `/cpu/self/opt/serial` | Serial optimized C implementation | 398| `/cpu/self/opt/blocked` | Blocked optimized C implementation | 399| `/cpu/self/avx/serial` | Serial AVX implementation | 400| `/cpu/self/avx/blocked` | Blocked AVX implementation | 401| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 402| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 403| `/cpu/occa` | Serial OCCA kernels | 404| `/gpu/occa` | CUDA OCCA kernels | 405| `/omp/occa` | OpenMP OCCA kernels | 406| `/ocl/occa` | OpenCL OCCA kernels | 407| `/gpu/cuda/ref` | Reference pure CUDA kernels | 408| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 409| `/gpu/cuda/shared` | Optimized pure CUDA kernels using shared memory | 410| `/gpu/cuda/gen` | Optimized pure CUDA kernels using code generation | 411| `/gpu/magma` | CUDA MAGMA kernels | 412 413Examples available in this release: 414 415:::{list-table} 416:header-rows: 1 417:widths: auto 418* - User code 419 - Example 420* - `ceed` 421 - * ex1 (volume) 422 * ex2 (surface) 423* - `mfem` 424 - * BP1 (scalar mass operator) 425 * BP3 (scalar Laplace operator) 426* - `petsc` 427 - * BP1 (scalar mass operator) 428 * BP2 (vector mass operator) 429 * BP3 (scalar Laplace operator) 430 * BP4 (vector Laplace operator) 431 * BP5 (collocated scalar Laplace operator) 432 * BP6 (collocated vector Laplace operator) 433 * Navier-Stokes 434* - `nek5000` 435 - * BP1 (scalar mass operator) 436 * BP3 (scalar Laplace operator) 437::: 438 439(v0-4)= 440 441## v0.4 (Apr 1, 2019) 442 443libCEED v0.4 was made again publicly available in the second full CEED software 444distribution, release CEED 2.0. This release contained notable features, such as 445four new CPU backends, two new GPU backends, CPU backend optimizations, initial 446support for operator composition, performance benchmarking, and a Navier-Stokes demo. 447The new CPU backends in this release came in two families. The `/cpu/self/*/serial` 448backends process one element at a time and are intended for meshes with a smaller number 449of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of 450eight interlaced elements and are intended for meshes with higher numbers of elements. 451The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU 452performance. The `/cpu/self/xsmm/*` backends rely upon the 453[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU 454performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA. 455The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable 456performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple 457parallelization approach, where each thread treats a finite element. Using just in time 458compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this 459backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend 460achieve good peak performance for 1D, 2D, and low order 3D problems, but performance 461deteriorates very quickly when threads run out of registers. 462 463A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED 464examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`). 465This example solves the time-dependent Navier-Stokes equations of compressible gas 466dynamics in a static Eulerian three-dimensional frame, using structured high-order 467finite/spectral element spatial discretizations and explicit high-order time-stepping 468(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc, 469so that the pointwise physics (defined at quadrature points) is separated from the 470parallelization and meshing concerns. 471 472Backends available in this release: 473 474| CEED resource (`-ceed`) | Backend | 475|--------------------------|-----------------------------------------------------| 476| `/cpu/self/ref/serial` | Serial reference implementation | 477| `/cpu/self/ref/blocked` | Blocked reference implementation | 478| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 479| `/cpu/self/avx/serial` | Serial AVX implementation | 480| `/cpu/self/avx/blocked` | Blocked AVX implementation | 481| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 482| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 483| `/cpu/occa` | Serial OCCA kernels | 484| `/gpu/occa` | CUDA OCCA kernels | 485| `/omp/occa` | OpenMP OCCA kernels | 486| `/ocl/occa` | OpenCL OCCA kernels | 487| `/gpu/cuda/ref` | Reference pure CUDA kernels | 488| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 489| `/gpu/magma` | CUDA MAGMA kernels | 490 491Examples available in this release: 492 493:::{list-table} 494:header-rows: 1 495:widths: auto 496* - User code 497 - Example 498* - `ceed` 499 - * ex1 (volume) 500* - `mfem` 501 - * BP1 (scalar mass operator) 502 * BP3 (scalar Laplace operator) 503* - `petsc` 504 - * BP1 (scalar mass operator) 505 * BP3 (scalar Laplace operator) 506 * Navier-Stokes 507* - `nek5000` 508 - * BP1 (scalar mass operator) 509 * BP3 (scalar Laplace operator) 510::: 511 512(v0-3)= 513 514## v0.3 (Sep 30, 2018) 515 516Notable features in this release include active/passive field interface, support for 517non-tensor bases, backend optimization, and improved Fortran interface. This release 518also focused on providing improved continuous integration, and many new tests with code 519coverage reports of about 90%. This release also provided a significant change to the 520public interface: a {ref}`CeedQFunction` can take any number of named input and output 521arguments while {ref}`CeedOperator` connects them to the actual data, which may be 522supplied explicitly to `CeedOperatorApply()` (active) or separately via 523`CeedOperatorSetField()` (passive). This interface change enables reusable libraries 524of CeedQFunctions and composition of block solvers constructed using 525{ref}`CeedOperator`. A concept of blocked restriction was added to this release and 526used in an optimized CPU backend. Although this is typically not visible to the user, 527it enables effective use of arbitrary-length SIMD while maintaining cache locality. 528This CPU backend also implements an algebraic factorization of tensor product gradients 529to perform fewer operations than standard application of interpolation and 530differentiation from nodes to quadrature points. This algebraic formulation 531automatically supports non-polynomial and non-interpolatory bases, thus is more general 532than the more common derivation in terms of Lagrange polynomials on the quadrature points. 533 534Backends available in this release: 535 536| CEED resource (`-ceed`) | Backend | 537|-------------------------|-----------------------------------------------------| 538| `/cpu/self/blocked` | Blocked reference implementation | 539| `/cpu/self/ref` | Serial reference implementation | 540| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 541| `/cpu/occa` | Serial OCCA kernels | 542| `/gpu/occa` | CUDA OCCA kernels | 543| `/omp/occa` | OpenMP OCCA kernels | 544| `/ocl/occa` | OpenCL OCCA kernels | 545| `/gpu/magma` | CUDA MAGMA kernels | 546 547Examples available in this release: 548 549:::{list-table} 550:header-rows: 1 551:widths: auto 552* - User code 553 - Example 554* - `ceed` 555 - * ex1 (volume) 556* - `mfem` 557 - * BP1 (scalar mass operator) 558 * BP3 (scalar Laplace operator) 559* - `petsc` 560 - * BP1 (scalar mass operator) 561 * BP3 (scalar Laplace operator) 562* - `nek5000` 563 - * BP1 (scalar mass operator) 564 * BP3 (scalar Laplace operator) 565::: 566 567(v0-21)= 568 569## v0.21 (Sep 30, 2018) 570 571A MAGMA backend (which relies upon the 572[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this 573release. This initial integration set up the framework of using MAGMA and provided the 574libCEED functionality through MAGMA kernels as one of libCEED’s computational backends. 575As any other backend, the MAGMA backend provides extended basic data structures for 576{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements 577the fundamental CEED building blocks to work with the new data structures. 578In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data 579but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled 580internally, and thus seamlessly to the user, through the functions/methods that are 581provided to support them. 582 583Backends available in this release: 584 585| CEED resource (`-ceed`) | Backend | 586|-------------------------|---------------------------------| 587| `/cpu/self` | Serial reference implementation | 588| `/cpu/occa` | Serial OCCA kernels | 589| `/gpu/occa` | CUDA OCCA kernels | 590| `/omp/occa` | OpenMP OCCA kernels | 591| `/ocl/occa` | OpenCL OCCA kernels | 592| `/gpu/magma` | CUDA MAGMA kernels | 593 594Examples available in this release: 595 596:::{list-table} 597:header-rows: 1 598:widths: auto 599* - User code 600 - Example 601* - `ceed` 602 - * ex1 (volume) 603* - `mfem` 604 - * BP1 (scalar mass operator) 605 * BP3 (scalar Laplace operator) 606* - `petsc` 607 - * BP1 (scalar mass operator) 608* - `nek5000` 609 - * BP1 (scalar mass operator) 610::: 611 612(v0-2)= 613 614## v0.2 (Mar 30, 2018) 615 616libCEED was made publicly available the first full CEED software distribution, release 617CEED 1.0. The distribution was made available using the Spack package manager to provide 618a common, easy-to-use build environment, where the user can build the CEED distribution 619with all dependencies. This release included a new Fortran interface for the library. 620This release also contained major improvements in the OCCA backend (including a new 621`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to 622compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an 623`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application 624of the Laplace operator) was also added to this release. 625 626Backends available in this release: 627 628| CEED resource (`-ceed`) | Backend | 629|-------------------------|---------------------------------| 630| `/cpu/self` | Serial reference implementation | 631| `/cpu/occa` | Serial OCCA kernels | 632| `/gpu/occa` | CUDA OCCA kernels | 633| `/omp/occa` | OpenMP OCCA kernels | 634| `/ocl/occa` | OpenCL OCCA kernels | 635 636Examples available in this release: 637 638:::{list-table} 639:header-rows: 1 640:widths: auto 641* - User code 642 - Example 643* - `ceed` 644 - * ex1 (volume) 645* - `mfem` 646 - * BP1 (scalar mass operator) 647 * BP3 (scalar Laplace operator) 648* - `petsc` 649 - * BP1 (scalar mass operator) 650* - `nek5000` 651 - * BP1 (scalar mass operator) 652::: 653 654(v0-1)= 655 656## v0.1 (Jan 3, 2018) 657 658Initial low-level API of the CEED project. The low-level API provides a set of Finite 659Elements kernels and components for writing new low-level kernels. Examples include: 660vector and sparse linear algebra, element matrix assembly over a batch of elements, 661partial assembly and action for efficient high-order operators like mass, diffusion, 662advection, etc. The main goal of the low-level API is to establish the basis for the 663high-level API. Also, identifying such low-level kernels and providing a reference 664implementation for them serves as the basis for specialized backend implementations. 665This release contained several backends: `/cpu/self`, and backends which rely upon the 666[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`, 667`/gpu/occa`, and `/omp/occa`. 668It also included several examples, in the `examples` folder: 669A standalone code that shows the usage of libCEED (with no external 670dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1 671(with the application of the mass operator); and a `petsc` example to perform BP1 672(with the application of the mass operator). 673 674Backends available in this release: 675 676| CEED resource (`-ceed`) | Backend | 677|-------------------------|---------------------------------| 678| `/cpu/self` | Serial reference implementation | 679| `/cpu/occa` | Serial OCCA kernels | 680| `/gpu/occa` | CUDA OCCA kernels | 681| `/omp/occa` | OpenMP OCCA kernels | 682 683Examples available in this release: 684 685| User code | Example | 686|-----------------------|-----------------------------------| 687| `ceed` | ex1 (scalar Laplace operator) | 688| `mfem` | BP1 (scalar mass operator) | 689| `petsc` | BP1 (scalar mass operator) | 690``` 691