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