1# Changes/Release Notes 2 3On this page we provide a summary of the main API changes, new features and examples 4for each release of libCEED. 5 6(main)= 7 8## Current `main` branch 9 10### Interface changes 11 12- Update {c:func} `CeedQFunctionGetFields` and {c:func} `CeedOperatorGetFields` to include number of fields. 13- 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`. 14- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added. 15- Add {c:func} `CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients. 16- Promote {c:func} `CeedOperatorCheckReady`to the public API to facilitate interactive interfaces. 17- Warning added when compiling OCCA backend to alert users that this backend is experimental. 18- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`. 19 20### New features 21 22- `CeedScalar` can now be set as `float` or `double` at compile time. 23 24### Maintainability 25 26- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage. 27 28(v0-9)= 29 30## v0.9 (Jul 6, 2021) 31 32### Interface changes 33 34- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact. 35 36### New features 37 38- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use. 39- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends. 40- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations. 41- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed). 42 43### Performance improvements 44 45### Examples 46 47- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations. 48- Fluid mechanics example adds GPU support and improves modularity. 49 50### Deprecated backends 51 52- 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. 53 54(v0-8)= 55 56## v0.8 (Mar 31, 2021) 57 58### Interface changes 59 60- Error handling improved to include enumerated error codes for C interface return values. 61- 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. 62 63### New features 64 65- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features. 66- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly. 67- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators. 68 69### Performance improvements 70 71- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`. 72- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`. 73 74### Examples 75 76- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions. 77- {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. 78- {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. 79- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree. 80 81(v0-7)= 82 83## v0.7 (Sep 29, 2020) 84 85### Interface changes 86 87- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors. 88 As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`. 89 These changes improve support for mixed finite element methods. 90- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature. 91- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity. 92- 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. 93- 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. 94- 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. 95- 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. 96 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`. 97- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory. 98- 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. 99 100### New features 101 102- New HIP backend: `/gpu/hip/ref`. 103- CeedQFunction support for user `CUfunction`s in some backends 104 105### Performance improvements 106 107- OCCA backend rebuilt to facilitate future performance enhancements. 108- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`. 109 110### Examples 111 112- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions. 113 114### Deprecated backends 115 116- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`. 117 118(v0-6)= 119 120## v0.6 (Mar 29, 2020) 121 122libCEED v0.6 contains numerous new features and examples, as well as expanded 123documentation in [this new website](https://libceed.readthedocs.io). 124 125### New features 126 127- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly 128 1-1 correspondence with the C interface, plus some convenience features. For instance, 129 data stored in the {cpp:type}`CeedVector` structure are available without copy as 130 {py:class}`numpy.ndarray`. Short tutorials are provided in 131 [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/). 132- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point, 133 {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal 134 ({c:func}`CeedOperatorAssembleLinearDiagonal`). These operations are useful for 135 preconditioning ingredients and are used in the libCEED's multigrid examples. 136- The inverse of separable operators can be obtained using 137 {c:func}`CeedOperatorCreateFDMElementInverse` and applied with 138 {c:func}`CeedOperatorApply`. This is a useful preconditioning ingredient, 139 especially for Laplacians and related operators. 140- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`, 141 {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`. 142- Make public accessors for various attributes to facilitate writing composable code. 143- New backend: `/cpu/self/memcheck/serial`. 144- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA 145 backends. (Single source is coming soon for OCCA backends.) 146- Fix some missing edge cases in CUDA backend. 147 148### Performance Improvements 149 150- MAGMA backend performance optimization and non-tensor bases. 151- No-copy optimization in {c:func}`CeedOperatorApply`. 152 153### Interface changes 154 155- Replace {code}`CeedElemRestrictionCreateIdentity` and 156 {code}`CeedElemRestrictionCreateBlocked` with more flexible 157 {c:func}`CeedElemRestrictionCreateStrided` and 158 {c:func}`CeedElemRestrictionCreateBlockedStrided`. 159- Add arguments to {c:func}`CeedQFunctionCreateIdentity`. 160- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification 161 with {cpp:enum}`CeedInterlaceMode`. This is now an attribute of the 162 {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no 163 longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and 164 {c:func}`CeedElemRestrictionApply`. 165 166### Examples 167 168libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`. 169Notable additions include: 170 171- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of 172 a domain in 1, 2, and 3 dimensions by applying a Laplacian. 173 174- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area 175 of domains (like the cube and sphere) by direct integration on a surface mesh; 176 demonstrates geometric dimension different from topological dimension. 177 178- PETSc {ref}`example-petsc-bps`: 179 180 - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support. 181 - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements 182 and transparent CUDA support. 183 - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`): 184 generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric 185 dimension different from topological dimension. 186 187- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid 188 solver with algebraic multigrid coarse solve. 189 190- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly 191 `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`), 192 implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and 193 quasi-2D computational domain support. 194 195- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for 196 linear elasticity, small-strain hyperelasticity, and globalized finite-strain 197 hyperelasticity using p-multigrid with algebraic multigrid coarse solve. 198 199(v0-5)= 200 201## v0.5 (Sep 18, 2019) 202 203For this release, several improvements were made. Two new CUDA backends were added to 204the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art 205performance using single-source {ref}`CeedQFunction`. From this release, users 206can define Q-Functions in a single source code independently of the targeted backend 207with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU 208compilation of the user provided {ref}`CeedQFunction` code. To allow a unified 209declaration, the {ref}`CeedQFunction` API has undergone a slight change: 210the `QFunctionField` parameter `ncomp` has been changed to `size`. This change 211requires setting the previous value of `ncomp` to `ncomp*dim` when adding a 212`QFunctionField` with eval mode `CEED EVAL GRAD`. 213 214Additionally, new CPU backends 215were included in this release, such as the `/cpu/self/opt/*` backends (which are 216written in pure C and use partial **E-vectors** to improve performance) and the 217`/cpu/self/ref/memcheck` backend (which relies upon the 218[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user 219{ref}`CeedQFunction` have no undefined values). 220This release also included various performance improvements, bug fixes, new examples, 221and improved tests. Among these improvements, vectorized instructions for 222{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD` 223instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and 224identity Q-Functions were introduced, and the PETSc benchmark problems were expanded 225to include unstructured meshes handling were. For this expansion, the prior version of 226the PETSc BPs, which only included data associated with structured geometries, were 227renamed `bpsraw`, and the new version of the BPs, which can handle data associated 228with any unstructured geometry, were called `bps`. Additionally, other benchmark 229problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively), 230and BP5 and BP6 (the collocated versions---for which the quadrature points are the same 231as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc 232examples. Furthermoew, another standalone libCEED example, called `ex2`, which 233computes the surface area of a given mesh was added to this release. 234 235Backends available in this release: 236 237| CEED resource (`-ceed`) | Backend | 238|--------------------------|-----------------------------------------------------| 239| `/cpu/self/ref/serial` | Serial reference implementation | 240| `/cpu/self/ref/blocked` | Blocked reference implementation | 241| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks | 242| `/cpu/self/opt/serial` | Serial optimized C implementation | 243| `/cpu/self/opt/blocked` | Blocked optimized C implementation | 244| `/cpu/self/avx/serial` | Serial AVX implementation | 245| `/cpu/self/avx/blocked` | Blocked AVX implementation | 246| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 247| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 248| `/cpu/occa` | Serial OCCA kernels | 249| `/gpu/occa` | CUDA OCCA kernels | 250| `/omp/occa` | OpenMP OCCA kernels | 251| `/ocl/occa` | OpenCL OCCA kernels | 252| `/gpu/cuda/ref` | Reference pure CUDA kernels | 253| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 254| `/gpu/cuda/shared` | Optimized pure CUDA kernels using shared memory | 255| `/gpu/cuda/gen` | Optimized pure CUDA kernels using code generation | 256| `/gpu/magma` | CUDA MAGMA kernels | 257 258Examples available in this release: 259 260:::{list-table} 261:header-rows: 1 262:widths: auto 263* - User code 264 - Example 265* - `ceed` 266 - * ex1 (volume) 267 * ex2 (surface) 268* - `mfem` 269 - * BP1 (scalar mass operator) 270 * BP3 (scalar Laplace operator) 271* - `petsc` 272 - * BP1 (scalar mass operator) 273 * BP2 (vector mass operator) 274 * BP3 (scalar Laplace operator) 275 * BP4 (vector Laplace operator) 276 * BP5 (collocated scalar Laplace operator) 277 * BP6 (collocated vector Laplace operator) 278 * Navier-Stokes 279* - `nek5000` 280 - * BP1 (scalar mass operator) 281 * BP3 (scalar Laplace operator) 282::: 283 284(v0-4)= 285 286## v0.4 (Apr 1, 2019) 287 288libCEED v0.4 was made again publicly available in the second full CEED software 289distribution, release CEED 2.0. This release contained notable features, such as 290four new CPU backends, two new GPU backends, CPU backend optimizations, initial 291support for operator composition, performance benchmarking, and a Navier-Stokes demo. 292The new CPU backends in this release came in two families. The `/cpu/self/*/serial` 293backends process one element at a time and are intended for meshes with a smaller number 294of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of 295eight interlaced elements and are intended for meshes with higher numbers of elements. 296The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU 297performance. The `/cpu/self/xsmm/*` backends rely upon the 298[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU 299performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA. 300The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable 301performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple 302parallelization approach, where each thread treats a finite element. Using just in time 303compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this 304backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend 305achieve good peak performance for 1D, 2D, and low order 3D problems, but performance 306deteriorates very quickly when threads run out of registers. 307 308A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED 309examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`). 310This example solves the time-dependent Navier-Stokes equations of compressible gas 311dynamics in a static Eulerian three-dimensional frame, using structured high-order 312finite/spectral element spatial discretizations and explicit high-order time-stepping 313(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc, 314so that the pointwise physics (defined at quadrature points) is separated from the 315parallelization and meshing concerns. 316 317Backends available in this release: 318 319| CEED resource (`-ceed`) | Backend | 320|--------------------------|-----------------------------------------------------| 321| `/cpu/self/ref/serial` | Serial reference implementation | 322| `/cpu/self/ref/blocked` | Blocked reference implementation | 323| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 324| `/cpu/self/avx/serial` | Serial AVX implementation | 325| `/cpu/self/avx/blocked` | Blocked AVX implementation | 326| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 327| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 328| `/cpu/occa` | Serial OCCA kernels | 329| `/gpu/occa` | CUDA OCCA kernels | 330| `/omp/occa` | OpenMP OCCA kernels | 331| `/ocl/occa` | OpenCL OCCA kernels | 332| `/gpu/cuda/ref` | Reference pure CUDA kernels | 333| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 334| `/gpu/magma` | CUDA MAGMA kernels | 335 336Examples available in this release: 337 338:::{list-table} 339:header-rows: 1 340:widths: auto 341* - User code 342 - Example 343* - `ceed` 344 - * ex1 (volume) 345* - `mfem` 346 - * BP1 (scalar mass operator) 347 * BP3 (scalar Laplace operator) 348* - `petsc` 349 - * BP1 (scalar mass operator) 350 * BP3 (scalar Laplace operator) 351 * Navier-Stokes 352* - `nek5000` 353 - * BP1 (scalar mass operator) 354 * BP3 (scalar Laplace operator) 355::: 356 357(v0-3)= 358 359## v0.3 (Sep 30, 2018) 360 361Notable features in this release include active/passive field interface, support for 362non-tensor bases, backend optimization, and improved Fortran interface. This release 363also focused on providing improved continuous integration, and many new tests with code 364coverage reports of about 90%. This release also provided a significant change to the 365public interface: a {ref}`CeedQFunction` can take any number of named input and output 366arguments while {ref}`CeedOperator` connects them to the actual data, which may be 367supplied explicitly to `CeedOperatorApply()` (active) or separately via 368`CeedOperatorSetField()` (passive). This interface change enables reusable libraries 369of CeedQFunctions and composition of block solvers constructed using 370{ref}`CeedOperator`. A concept of blocked restriction was added to this release and 371used in an optimized CPU backend. Although this is typically not visible to the user, 372it enables effective use of arbitrary-length SIMD while maintaining cache locality. 373This CPU backend also implements an algebraic factorization of tensor product gradients 374to perform fewer operations than standard application of interpolation and 375differentiation from nodes to quadrature points. This algebraic formulation 376automatically supports non-polynomial and non-interpolatory bases, thus is more general 377than the more common derivation in terms of Lagrange polynomials on the quadrature points. 378 379Backends available in this release: 380 381| CEED resource (`-ceed`) | Backend | 382|-------------------------|-----------------------------------------------------| 383| `/cpu/self/blocked` | Blocked reference implementation | 384| `/cpu/self/ref` | Serial reference implementation | 385| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 386| `/cpu/occa` | Serial OCCA kernels | 387| `/gpu/occa` | CUDA OCCA kernels | 388| `/omp/occa` | OpenMP OCCA kernels | 389| `/ocl/occa` | OpenCL OCCA kernels | 390| `/gpu/magma` | CUDA MAGMA kernels | 391 392Examples available in this release: 393 394:::{list-table} 395:header-rows: 1 396:widths: auto 397* - User code 398 - Example 399* - `ceed` 400 - * ex1 (volume) 401* - `mfem` 402 - * BP1 (scalar mass operator) 403 * BP3 (scalar Laplace operator) 404* - `petsc` 405 - * BP1 (scalar mass operator) 406 * BP3 (scalar Laplace operator) 407* - `nek5000` 408 - * BP1 (scalar mass operator) 409 * BP3 (scalar Laplace operator) 410::: 411 412(v0-21)= 413 414## v0.21 (Sep 30, 2018) 415 416A MAGMA backend (which relies upon the 417[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this 418release. This initial integration set up the framework of using MAGMA and provided the 419libCEED functionality through MAGMA kernels as one of libCEED’s computational backends. 420As any other backend, the MAGMA backend provides extended basic data structures for 421{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements 422the fundamental CEED building blocks to work with the new data structures. 423In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data 424but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled 425internally, and thus seamlessly to the user, through the functions/methods that are 426provided to support them. 427 428Backends available in this release: 429 430| CEED resource (`-ceed`) | Backend | 431|-------------------------|---------------------------------| 432| `/cpu/self` | Serial reference implementation | 433| `/cpu/occa` | Serial OCCA kernels | 434| `/gpu/occa` | CUDA OCCA kernels | 435| `/omp/occa` | OpenMP OCCA kernels | 436| `/ocl/occa` | OpenCL OCCA kernels | 437| `/gpu/magma` | CUDA MAGMA kernels | 438 439Examples available in this release: 440 441:::{list-table} 442:header-rows: 1 443:widths: auto 444* - User code 445 - Example 446* - `ceed` 447 - * ex1 (volume) 448* - `mfem` 449 - * BP1 (scalar mass operator) 450 * BP3 (scalar Laplace operator) 451* - `petsc` 452 - * BP1 (scalar mass operator) 453* - `nek5000` 454 - * BP1 (scalar mass operator) 455::: 456 457(v0-2)= 458 459## v0.2 (Mar 30, 2018) 460 461libCEED was made publicly available the first full CEED software distribution, release 462CEED 1.0. The distribution was made available using the Spack package manager to provide 463a common, easy-to-use build environment, where the user can build the CEED distribution 464with all dependencies. This release included a new Fortran interface for the library. 465This release also contained major improvements in the OCCA backend (including a new 466`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to 467compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an 468`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application 469of the Laplace operator) was also added to this release. 470 471Backends available in this release: 472 473| CEED resource (`-ceed`) | Backend | 474|-------------------------|---------------------------------| 475| `/cpu/self` | Serial reference implementation | 476| `/cpu/occa` | Serial OCCA kernels | 477| `/gpu/occa` | CUDA OCCA kernels | 478| `/omp/occa` | OpenMP OCCA kernels | 479| `/ocl/occa` | OpenCL OCCA kernels | 480 481Examples available in this release: 482 483:::{list-table} 484:header-rows: 1 485:widths: auto 486* - User code 487 - Example 488* - `ceed` 489 - * ex1 (volume) 490* - `mfem` 491 - * BP1 (scalar mass operator) 492 * BP3 (scalar Laplace operator) 493* - `petsc` 494 - * BP1 (scalar mass operator) 495* - `nek5000` 496 - * BP1 (scalar mass operator) 497::: 498 499(v0-1)= 500 501## v0.1 (Jan 3, 2018) 502 503Initial low-level API of the CEED project. The low-level API provides a set of Finite 504Elements kernels and components for writing new low-level kernels. Examples include: 505vector and sparse linear algebra, element matrix assembly over a batch of elements, 506partial assembly and action for efficient high-order operators like mass, diffusion, 507advection, etc. The main goal of the low-level API is to establish the basis for the 508high-level API. Also, identifying such low-level kernels and providing a reference 509implementation for them serves as the basis for specialized backend implementations. 510This release contained several backends: `/cpu/self`, and backends which rely upon the 511[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`, 512`/gpu/occa`, and `/omp/occa`. 513It also included several examples, in the `examples` folder: 514A standalone code that shows the usage of libCEED (with no external 515dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1 516(with the application of the mass operator); and a `petsc` example to perform BP1 517(with the application of the mass operator). 518 519Backends available in this release: 520 521| CEED resource (`-ceed`) | Backend | 522|-------------------------|---------------------------------| 523| `/cpu/self` | Serial reference implementation | 524| `/cpu/occa` | Serial OCCA kernels | 525| `/gpu/occa` | CUDA OCCA kernels | 526| `/omp/occa` | OpenMP OCCA kernels | 527 528Examples available in this release: 529 530| User code | Example | 531|-----------------------|-----------------------------------| 532| `ceed` | ex1 (scalar Laplace operator) | 533| `mfem` | BP1 (scalar mass operator) | 534| `petsc` | BP1 (scalar mass operator) | 535``` 536