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