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