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- `Include-what-you-use` makefile target added as `make iwyu`. 28 29(v0-9)= 30 31## v0.9 (Jul 6, 2021) 32 33### Interface changes 34 35- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact. 36 37### New features 38 39- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use. 40- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends. 41- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations. 42- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed). 43 44### Performance improvements 45 46### Examples 47 48- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations. 49- Fluid mechanics example adds GPU support and improves modularity. 50 51### Deprecated backends 52 53- 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. 54 55(v0-8)= 56 57## v0.8 (Mar 31, 2021) 58 59### Interface changes 60 61- Error handling improved to include enumerated error codes for C interface return values. 62- 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. 63 64### New features 65 66- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features. 67- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly. 68- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators. 69 70### Performance improvements 71 72- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`. 73- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`. 74 75### Examples 76 77- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions. 78- {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. 79- {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. 80- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree. 81 82(v0-7)= 83 84## v0.7 (Sep 29, 2020) 85 86### Interface changes 87 88- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors. 89 As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`. 90 These changes improve support for mixed finite element methods. 91- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature. 92- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity. 93- 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. 94- 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. 95- 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. 96- 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. 97 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`. 98- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory. 99- 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. 100 101### New features 102 103- New HIP backend: `/gpu/hip/ref`. 104- CeedQFunction support for user `CUfunction`s in some backends 105 106### Performance improvements 107 108- OCCA backend rebuilt to facilitate future performance enhancements. 109- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`. 110 111### Examples 112 113- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions. 114 115### Deprecated backends 116 117- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`. 118 119(v0-6)= 120 121## v0.6 (Mar 29, 2020) 122 123libCEED v0.6 contains numerous new features and examples, as well as expanded 124documentation in [this new website](https://libceed.readthedocs.io). 125 126### New features 127 128- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly 129 1-1 correspondence with the C interface, plus some convenience features. For instance, 130 data stored in the {cpp:type}`CeedVector` structure are available without copy as 131 {py:class}`numpy.ndarray`. Short tutorials are provided in 132 [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/). 133- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point, 134 {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal 135 ({c:func}`CeedOperatorAssembleLinearDiagonal`). These operations are useful for 136 preconditioning ingredients and are used in the libCEED's multigrid examples. 137- The inverse of separable operators can be obtained using 138 {c:func}`CeedOperatorCreateFDMElementInverse` and applied with 139 {c:func}`CeedOperatorApply`. This is a useful preconditioning ingredient, 140 especially for Laplacians and related operators. 141- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`, 142 {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`. 143- Make public accessors for various attributes to facilitate writing composable code. 144- New backend: `/cpu/self/memcheck/serial`. 145- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA 146 backends. (Single source is coming soon for OCCA backends.) 147- Fix some missing edge cases in CUDA backend. 148 149### Performance Improvements 150 151- MAGMA backend performance optimization and non-tensor bases. 152- No-copy optimization in {c:func}`CeedOperatorApply`. 153 154### Interface changes 155 156- Replace {code}`CeedElemRestrictionCreateIdentity` and 157 {code}`CeedElemRestrictionCreateBlocked` with more flexible 158 {c:func}`CeedElemRestrictionCreateStrided` and 159 {c:func}`CeedElemRestrictionCreateBlockedStrided`. 160- Add arguments to {c:func}`CeedQFunctionCreateIdentity`. 161- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification 162 with {cpp:enum}`CeedInterlaceMode`. This is now an attribute of the 163 {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no 164 longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and 165 {c:func}`CeedElemRestrictionApply`. 166 167### Examples 168 169libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`. 170Notable additions include: 171 172- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of 173 a domain in 1, 2, and 3 dimensions by applying a Laplacian. 174 175- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area 176 of domains (like the cube and sphere) by direct integration on a surface mesh; 177 demonstrates geometric dimension different from topological dimension. 178 179- PETSc {ref}`example-petsc-bps`: 180 181 - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support. 182 - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements 183 and transparent CUDA support. 184 - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`): 185 generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric 186 dimension different from topological dimension. 187 188- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid 189 solver with algebraic multigrid coarse solve. 190 191- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly 192 `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`), 193 implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and 194 quasi-2D computational domain support. 195 196- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for 197 linear elasticity, small-strain hyperelasticity, and globalized finite-strain 198 hyperelasticity using p-multigrid with algebraic multigrid coarse solve. 199 200(v0-5)= 201 202## v0.5 (Sep 18, 2019) 203 204For this release, several improvements were made. Two new CUDA backends were added to 205the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art 206performance using single-source {ref}`CeedQFunction`. From this release, users 207can define Q-Functions in a single source code independently of the targeted backend 208with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU 209compilation of the user provided {ref}`CeedQFunction` code. To allow a unified 210declaration, the {ref}`CeedQFunction` API has undergone a slight change: 211the `QFunctionField` parameter `ncomp` has been changed to `size`. This change 212requires setting the previous value of `ncomp` to `ncomp*dim` when adding a 213`QFunctionField` with eval mode `CEED EVAL GRAD`. 214 215Additionally, new CPU backends 216were included in this release, such as the `/cpu/self/opt/*` backends (which are 217written in pure C and use partial **E-vectors** to improve performance) and the 218`/cpu/self/ref/memcheck` backend (which relies upon the 219[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user 220{ref}`CeedQFunction` have no undefined values). 221This release also included various performance improvements, bug fixes, new examples, 222and improved tests. Among these improvements, vectorized instructions for 223{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD` 224instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and 225identity Q-Functions were introduced, and the PETSc benchmark problems were expanded 226to include unstructured meshes handling were. For this expansion, the prior version of 227the PETSc BPs, which only included data associated with structured geometries, were 228renamed `bpsraw`, and the new version of the BPs, which can handle data associated 229with any unstructured geometry, were called `bps`. Additionally, other benchmark 230problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively), 231and BP5 and BP6 (the collocated versions---for which the quadrature points are the same 232as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc 233examples. Furthermoew, another standalone libCEED example, called `ex2`, which 234computes the surface area of a given mesh was added to this release. 235 236Backends available in this release: 237 238| CEED resource (`-ceed`) | Backend | 239|--------------------------|-----------------------------------------------------| 240| `/cpu/self/ref/serial` | Serial reference implementation | 241| `/cpu/self/ref/blocked` | Blocked reference implementation | 242| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks | 243| `/cpu/self/opt/serial` | Serial optimized C implementation | 244| `/cpu/self/opt/blocked` | Blocked optimized C implementation | 245| `/cpu/self/avx/serial` | Serial AVX implementation | 246| `/cpu/self/avx/blocked` | Blocked AVX implementation | 247| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 248| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 249| `/cpu/occa` | Serial OCCA kernels | 250| `/gpu/occa` | CUDA OCCA kernels | 251| `/omp/occa` | OpenMP OCCA kernels | 252| `/ocl/occa` | OpenCL OCCA kernels | 253| `/gpu/cuda/ref` | Reference pure CUDA kernels | 254| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 255| `/gpu/cuda/shared` | Optimized pure CUDA kernels using shared memory | 256| `/gpu/cuda/gen` | Optimized pure CUDA kernels using code generation | 257| `/gpu/magma` | CUDA MAGMA kernels | 258 259Examples available in this release: 260 261:::{list-table} 262:header-rows: 1 263:widths: auto 264* - User code 265 - Example 266* - `ceed` 267 - * ex1 (volume) 268 * ex2 (surface) 269* - `mfem` 270 - * BP1 (scalar mass operator) 271 * BP3 (scalar Laplace operator) 272* - `petsc` 273 - * BP1 (scalar mass operator) 274 * BP2 (vector mass operator) 275 * BP3 (scalar Laplace operator) 276 * BP4 (vector Laplace operator) 277 * BP5 (collocated scalar Laplace operator) 278 * BP6 (collocated vector Laplace operator) 279 * Navier-Stokes 280* - `nek5000` 281 - * BP1 (scalar mass operator) 282 * BP3 (scalar Laplace operator) 283::: 284 285(v0-4)= 286 287## v0.4 (Apr 1, 2019) 288 289libCEED v0.4 was made again publicly available in the second full CEED software 290distribution, release CEED 2.0. This release contained notable features, such as 291four new CPU backends, two new GPU backends, CPU backend optimizations, initial 292support for operator composition, performance benchmarking, and a Navier-Stokes demo. 293The new CPU backends in this release came in two families. The `/cpu/self/*/serial` 294backends process one element at a time and are intended for meshes with a smaller number 295of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of 296eight interlaced elements and are intended for meshes with higher numbers of elements. 297The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU 298performance. The `/cpu/self/xsmm/*` backends rely upon the 299[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU 300performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA. 301The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable 302performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple 303parallelization approach, where each thread treats a finite element. Using just in time 304compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this 305backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend 306achieve good peak performance for 1D, 2D, and low order 3D problems, but performance 307deteriorates very quickly when threads run out of registers. 308 309A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED 310examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`). 311This example solves the time-dependent Navier-Stokes equations of compressible gas 312dynamics in a static Eulerian three-dimensional frame, using structured high-order 313finite/spectral element spatial discretizations and explicit high-order time-stepping 314(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc, 315so that the pointwise physics (defined at quadrature points) is separated from the 316parallelization and meshing concerns. 317 318Backends available in this release: 319 320| CEED resource (`-ceed`) | Backend | 321|--------------------------|-----------------------------------------------------| 322| `/cpu/self/ref/serial` | Serial reference implementation | 323| `/cpu/self/ref/blocked` | Blocked reference implementation | 324| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 325| `/cpu/self/avx/serial` | Serial AVX implementation | 326| `/cpu/self/avx/blocked` | Blocked AVX implementation | 327| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 328| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 329| `/cpu/occa` | Serial OCCA kernels | 330| `/gpu/occa` | CUDA OCCA kernels | 331| `/omp/occa` | OpenMP OCCA kernels | 332| `/ocl/occa` | OpenCL OCCA kernels | 333| `/gpu/cuda/ref` | Reference pure CUDA kernels | 334| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 335| `/gpu/magma` | CUDA MAGMA kernels | 336 337Examples available in this release: 338 339:::{list-table} 340:header-rows: 1 341:widths: auto 342* - User code 343 - Example 344* - `ceed` 345 - * ex1 (volume) 346* - `mfem` 347 - * BP1 (scalar mass operator) 348 * BP3 (scalar Laplace operator) 349* - `petsc` 350 - * BP1 (scalar mass operator) 351 * BP3 (scalar Laplace operator) 352 * Navier-Stokes 353* - `nek5000` 354 - * BP1 (scalar mass operator) 355 * BP3 (scalar Laplace operator) 356::: 357 358(v0-3)= 359 360## v0.3 (Sep 30, 2018) 361 362Notable features in this release include active/passive field interface, support for 363non-tensor bases, backend optimization, and improved Fortran interface. This release 364also focused on providing improved continuous integration, and many new tests with code 365coverage reports of about 90%. This release also provided a significant change to the 366public interface: a {ref}`CeedQFunction` can take any number of named input and output 367arguments while {ref}`CeedOperator` connects them to the actual data, which may be 368supplied explicitly to `CeedOperatorApply()` (active) or separately via 369`CeedOperatorSetField()` (passive). This interface change enables reusable libraries 370of CeedQFunctions and composition of block solvers constructed using 371{ref}`CeedOperator`. A concept of blocked restriction was added to this release and 372used in an optimized CPU backend. Although this is typically not visible to the user, 373it enables effective use of arbitrary-length SIMD while maintaining cache locality. 374This CPU backend also implements an algebraic factorization of tensor product gradients 375to perform fewer operations than standard application of interpolation and 376differentiation from nodes to quadrature points. This algebraic formulation 377automatically supports non-polynomial and non-interpolatory bases, thus is more general 378than the more common derivation in terms of Lagrange polynomials on the quadrature points. 379 380Backends available in this release: 381 382| CEED resource (`-ceed`) | Backend | 383|-------------------------|-----------------------------------------------------| 384| `/cpu/self/blocked` | Blocked reference implementation | 385| `/cpu/self/ref` | Serial reference implementation | 386| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 387| `/cpu/occa` | Serial OCCA kernels | 388| `/gpu/occa` | CUDA OCCA kernels | 389| `/omp/occa` | OpenMP OCCA kernels | 390| `/ocl/occa` | OpenCL OCCA kernels | 391| `/gpu/magma` | CUDA MAGMA kernels | 392 393Examples available in this release: 394 395:::{list-table} 396:header-rows: 1 397:widths: auto 398* - User code 399 - Example 400* - `ceed` 401 - * ex1 (volume) 402* - `mfem` 403 - * BP1 (scalar mass operator) 404 * BP3 (scalar Laplace operator) 405* - `petsc` 406 - * BP1 (scalar mass operator) 407 * BP3 (scalar Laplace operator) 408* - `nek5000` 409 - * BP1 (scalar mass operator) 410 * BP3 (scalar Laplace operator) 411::: 412 413(v0-21)= 414 415## v0.21 (Sep 30, 2018) 416 417A MAGMA backend (which relies upon the 418[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this 419release. This initial integration set up the framework of using MAGMA and provided the 420libCEED functionality through MAGMA kernels as one of libCEED’s computational backends. 421As any other backend, the MAGMA backend provides extended basic data structures for 422{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements 423the fundamental CEED building blocks to work with the new data structures. 424In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data 425but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled 426internally, and thus seamlessly to the user, through the functions/methods that are 427provided to support them. 428 429Backends available in this release: 430 431| CEED resource (`-ceed`) | Backend | 432|-------------------------|---------------------------------| 433| `/cpu/self` | Serial reference implementation | 434| `/cpu/occa` | Serial OCCA kernels | 435| `/gpu/occa` | CUDA OCCA kernels | 436| `/omp/occa` | OpenMP OCCA kernels | 437| `/ocl/occa` | OpenCL OCCA kernels | 438| `/gpu/magma` | CUDA MAGMA kernels | 439 440Examples available in this release: 441 442:::{list-table} 443:header-rows: 1 444:widths: auto 445* - User code 446 - Example 447* - `ceed` 448 - * ex1 (volume) 449* - `mfem` 450 - * BP1 (scalar mass operator) 451 * BP3 (scalar Laplace operator) 452* - `petsc` 453 - * BP1 (scalar mass operator) 454* - `nek5000` 455 - * BP1 (scalar mass operator) 456::: 457 458(v0-2)= 459 460## v0.2 (Mar 30, 2018) 461 462libCEED was made publicly available the first full CEED software distribution, release 463CEED 1.0. The distribution was made available using the Spack package manager to provide 464a common, easy-to-use build environment, where the user can build the CEED distribution 465with all dependencies. This release included a new Fortran interface for the library. 466This release also contained major improvements in the OCCA backend (including a new 467`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to 468compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an 469`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application 470of the Laplace operator) was also added to this release. 471 472Backends available in this release: 473 474| CEED resource (`-ceed`) | Backend | 475|-------------------------|---------------------------------| 476| `/cpu/self` | Serial reference implementation | 477| `/cpu/occa` | Serial OCCA kernels | 478| `/gpu/occa` | CUDA OCCA kernels | 479| `/omp/occa` | OpenMP OCCA kernels | 480| `/ocl/occa` | OpenCL OCCA kernels | 481 482Examples available in this release: 483 484:::{list-table} 485:header-rows: 1 486:widths: auto 487* - User code 488 - Example 489* - `ceed` 490 - * ex1 (volume) 491* - `mfem` 492 - * BP1 (scalar mass operator) 493 * BP3 (scalar Laplace operator) 494* - `petsc` 495 - * BP1 (scalar mass operator) 496* - `nek5000` 497 - * BP1 (scalar mass operator) 498::: 499 500(v0-1)= 501 502## v0.1 (Jan 3, 2018) 503 504Initial low-level API of the CEED project. The low-level API provides a set of Finite 505Elements kernels and components for writing new low-level kernels. Examples include: 506vector and sparse linear algebra, element matrix assembly over a batch of elements, 507partial assembly and action for efficient high-order operators like mass, diffusion, 508advection, etc. The main goal of the low-level API is to establish the basis for the 509high-level API. Also, identifying such low-level kernels and providing a reference 510implementation for them serves as the basis for specialized backend implementations. 511This release contained several backends: `/cpu/self`, and backends which rely upon the 512[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`, 513`/gpu/occa`, and `/omp/occa`. 514It also included several examples, in the `examples` folder: 515A standalone code that shows the usage of libCEED (with no external 516dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1 517(with the application of the mass operator); and a `petsc` example to perform BP1 518(with the application of the mass operator). 519 520Backends available in this release: 521 522| CEED resource (`-ceed`) | Backend | 523|-------------------------|---------------------------------| 524| `/cpu/self` | Serial reference implementation | 525| `/cpu/occa` | Serial OCCA kernels | 526| `/gpu/occa` | CUDA OCCA kernels | 527| `/omp/occa` | OpenMP OCCA kernels | 528 529Examples available in this release: 530 531| User code | Example | 532|-----------------------|-----------------------------------| 533| `ceed` | ex1 (scalar Laplace operator) | 534| `mfem` | BP1 (scalar mass operator) | 535| `petsc` | BP1 (scalar mass operator) | 536``` 537