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- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name. 20- Clarify documentation for {c:func}`CeedVectorTakeArray`; this function will error if {c:func}`CeedVectorSetArray` with `copy_mode == CEED_USE_POINTER` was not previously called for the corresponding `CeedMemType`. 21- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitalized arrays; require initalized data for {c:func}`CeedVectorGetArray`. 22- Added {c:func}`CeedQFunctionContextRegisterDouble` and {c:func}`CeedQFunctionContextRegisterInt32` with {c:func}`CeedQFunctionContextSetDouble` and {c:func}`CeedQFunctionContextSetInt32` to facilitate easy updating of {ref}`CeedQFunctionContext` data by user defined field names. 23- Added {c:func}`CeedQFunctionContextGetFieldDescriptions` to retreive user defined descriptions of fields that are registered with `CeedQFunctionContextRegister*`. 24 25### New features 26 27- `CeedScalar` can now be set as `float` or `double` at compile time. 28- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends. 29- Added support for JiT of QFunctions with `#include "relative/path/local-file.h"` statements for additional local files. Note that files included with `""` are searched relative to the current file first, then by compiler paths (as with `<>` includes). To use this feature, one should adhere to relative paths only, not compiler flags like `-I`, which the JiT will not be aware of. 30- Remove need to guard library headers in QFunction source for code generation backends. 31- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present. 32- Added {c:func}`CeedStringAllocCopy` to reduce repeated code for copying strings internally. 33 34### Maintainability 35 36- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage. 37- `Include-what-you-use` makefile target added as `make iwyu`. 38- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase. 39 40(v0-9)= 41 42## v0.9 (Jul 6, 2021) 43 44### Interface changes 45 46- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact. 47 48### New features 49 50- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use. 51- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends. 52- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations. 53- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed). 54 55### Performance improvements 56 57### Examples 58 59- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations. 60- Fluid mechanics example adds GPU support and improves modularity. 61 62### Deprecated backends 63 64- 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. 65 66(v0-8)= 67 68## v0.8 (Mar 31, 2021) 69 70### Interface changes 71 72- Error handling improved to include enumerated error codes for C interface return values. 73- 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. 74 75### New features 76 77- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features. 78- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly. 79- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators. 80 81### Performance improvements 82 83- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`. 84- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`. 85 86### Examples 87 88- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions. 89- {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. 90- {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. 91- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree. 92 93(v0-7)= 94 95## v0.7 (Sep 29, 2020) 96 97### Interface changes 98 99- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors. 100 As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`. 101 These changes improve support for mixed finite element methods. 102- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature. 103- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity. 104- 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. 105- 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. 106- 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. 107- 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. 108 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`. 109- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory. 110- 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. 111 112### New features 113 114- New HIP backend: `/gpu/hip/ref`. 115- CeedQFunction support for user `CUfunction`s in some backends 116 117### Performance improvements 118 119- OCCA backend rebuilt to facilitate future performance enhancements. 120- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`. 121 122### Examples 123 124- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions. 125 126### Deprecated backends 127 128- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`. 129 130(v0-6)= 131 132## v0.6 (Mar 29, 2020) 133 134libCEED v0.6 contains numerous new features and examples, as well as expanded 135documentation in [this new website](https://libceed.readthedocs.io). 136 137### New features 138 139- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly 140 1-1 correspondence with the C interface, plus some convenience features. For instance, 141 data stored in the {cpp:type}`CeedVector` structure are available without copy as 142 {py:class}`numpy.ndarray`. Short tutorials are provided in 143 [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/). 144- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point, 145 {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal 146 ({c:func}`CeedOperatorAssembleLinearDiagonal`). These operations are useful for 147 preconditioning ingredients and are used in the libCEED's multigrid examples. 148- The inverse of separable operators can be obtained using 149 {c:func}`CeedOperatorCreateFDMElementInverse` and applied with 150 {c:func}`CeedOperatorApply`. This is a useful preconditioning ingredient, 151 especially for Laplacians and related operators. 152- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`, 153 {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`. 154- Make public accessors for various attributes to facilitate writing composable code. 155- New backend: `/cpu/self/memcheck/serial`. 156- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA 157 backends. (Single source is coming soon for OCCA backends.) 158- Fix some missing edge cases in CUDA backend. 159 160### Performance Improvements 161 162- MAGMA backend performance optimization and non-tensor bases. 163- No-copy optimization in {c:func}`CeedOperatorApply`. 164 165### Interface changes 166 167- Replace {code}`CeedElemRestrictionCreateIdentity` and 168 {code}`CeedElemRestrictionCreateBlocked` with more flexible 169 {c:func}`CeedElemRestrictionCreateStrided` and 170 {c:func}`CeedElemRestrictionCreateBlockedStrided`. 171- Add arguments to {c:func}`CeedQFunctionCreateIdentity`. 172- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification 173 with {cpp:enum}`CeedInterlaceMode`. This is now an attribute of the 174 {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no 175 longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and 176 {c:func}`CeedElemRestrictionApply`. 177 178### Examples 179 180libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`. 181Notable additions include: 182 183- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of 184 a domain in 1, 2, and 3 dimensions by applying a Laplacian. 185 186- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area 187 of domains (like the cube and sphere) by direct integration on a surface mesh; 188 demonstrates geometric dimension different from topological dimension. 189 190- PETSc {ref}`example-petsc-bps`: 191 192 - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support. 193 - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements 194 and transparent CUDA support. 195 - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`): 196 generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric 197 dimension different from topological dimension. 198 199- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid 200 solver with algebraic multigrid coarse solve. 201 202- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly 203 `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`), 204 implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and 205 quasi-2D computational domain support. 206 207- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for 208 linear elasticity, small-strain hyperelasticity, and globalized finite-strain 209 hyperelasticity using p-multigrid with algebraic multigrid coarse solve. 210 211(v0-5)= 212 213## v0.5 (Sep 18, 2019) 214 215For this release, several improvements were made. Two new CUDA backends were added to 216the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art 217performance using single-source {ref}`CeedQFunction`. From this release, users 218can define Q-Functions in a single source code independently of the targeted backend 219with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU 220compilation of the user provided {ref}`CeedQFunction` code. To allow a unified 221declaration, the {ref}`CeedQFunction` API has undergone a slight change: 222the `QFunctionField` parameter `ncomp` has been changed to `size`. This change 223requires setting the previous value of `ncomp` to `ncomp*dim` when adding a 224`QFunctionField` with eval mode `CEED EVAL GRAD`. 225 226Additionally, new CPU backends 227were included in this release, such as the `/cpu/self/opt/*` backends (which are 228written in pure C and use partial **E-vectors** to improve performance) and the 229`/cpu/self/ref/memcheck` backend (which relies upon the 230[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user 231{ref}`CeedQFunction` have no undefined values). 232This release also included various performance improvements, bug fixes, new examples, 233and improved tests. Among these improvements, vectorized instructions for 234{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD` 235instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and 236identity Q-Functions were introduced, and the PETSc benchmark problems were expanded 237to include unstructured meshes handling were. For this expansion, the prior version of 238the PETSc BPs, which only included data associated with structured geometries, were 239renamed `bpsraw`, and the new version of the BPs, which can handle data associated 240with any unstructured geometry, were called `bps`. Additionally, other benchmark 241problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively), 242and BP5 and BP6 (the collocated versions---for which the quadrature points are the same 243as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc 244examples. Furthermoew, another standalone libCEED example, called `ex2`, which 245computes the surface area of a given mesh was added to this release. 246 247Backends available in this release: 248 249| CEED resource (`-ceed`) | Backend | 250|--------------------------|-----------------------------------------------------| 251| `/cpu/self/ref/serial` | Serial reference implementation | 252| `/cpu/self/ref/blocked` | Blocked reference implementation | 253| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks | 254| `/cpu/self/opt/serial` | Serial optimized C implementation | 255| `/cpu/self/opt/blocked` | Blocked optimized C implementation | 256| `/cpu/self/avx/serial` | Serial AVX implementation | 257| `/cpu/self/avx/blocked` | Blocked AVX implementation | 258| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 259| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 260| `/cpu/occa` | Serial OCCA kernels | 261| `/gpu/occa` | CUDA OCCA kernels | 262| `/omp/occa` | OpenMP OCCA kernels | 263| `/ocl/occa` | OpenCL OCCA kernels | 264| `/gpu/cuda/ref` | Reference pure CUDA kernels | 265| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 266| `/gpu/cuda/shared` | Optimized pure CUDA kernels using shared memory | 267| `/gpu/cuda/gen` | Optimized pure CUDA kernels using code generation | 268| `/gpu/magma` | CUDA MAGMA kernels | 269 270Examples available in this release: 271 272:::{list-table} 273:header-rows: 1 274:widths: auto 275* - User code 276 - Example 277* - `ceed` 278 - * ex1 (volume) 279 * ex2 (surface) 280* - `mfem` 281 - * BP1 (scalar mass operator) 282 * BP3 (scalar Laplace operator) 283* - `petsc` 284 - * BP1 (scalar mass operator) 285 * BP2 (vector mass operator) 286 * BP3 (scalar Laplace operator) 287 * BP4 (vector Laplace operator) 288 * BP5 (collocated scalar Laplace operator) 289 * BP6 (collocated vector Laplace operator) 290 * Navier-Stokes 291* - `nek5000` 292 - * BP1 (scalar mass operator) 293 * BP3 (scalar Laplace operator) 294::: 295 296(v0-4)= 297 298## v0.4 (Apr 1, 2019) 299 300libCEED v0.4 was made again publicly available in the second full CEED software 301distribution, release CEED 2.0. This release contained notable features, such as 302four new CPU backends, two new GPU backends, CPU backend optimizations, initial 303support for operator composition, performance benchmarking, and a Navier-Stokes demo. 304The new CPU backends in this release came in two families. The `/cpu/self/*/serial` 305backends process one element at a time and are intended for meshes with a smaller number 306of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of 307eight interlaced elements and are intended for meshes with higher numbers of elements. 308The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU 309performance. The `/cpu/self/xsmm/*` backends rely upon the 310[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU 311performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA. 312The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable 313performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple 314parallelization approach, where each thread treats a finite element. Using just in time 315compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this 316backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend 317achieve good peak performance for 1D, 2D, and low order 3D problems, but performance 318deteriorates very quickly when threads run out of registers. 319 320A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED 321examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`). 322This example solves the time-dependent Navier-Stokes equations of compressible gas 323dynamics in a static Eulerian three-dimensional frame, using structured high-order 324finite/spectral element spatial discretizations and explicit high-order time-stepping 325(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc, 326so that the pointwise physics (defined at quadrature points) is separated from the 327parallelization and meshing concerns. 328 329Backends available in this release: 330 331| CEED resource (`-ceed`) | Backend | 332|--------------------------|-----------------------------------------------------| 333| `/cpu/self/ref/serial` | Serial reference implementation | 334| `/cpu/self/ref/blocked` | Blocked reference implementation | 335| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 336| `/cpu/self/avx/serial` | Serial AVX implementation | 337| `/cpu/self/avx/blocked` | Blocked AVX implementation | 338| `/cpu/self/xsmm/serial` | Serial LIBXSMM implementation | 339| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation | 340| `/cpu/occa` | Serial OCCA kernels | 341| `/gpu/occa` | CUDA OCCA kernels | 342| `/omp/occa` | OpenMP OCCA kernels | 343| `/ocl/occa` | OpenCL OCCA kernels | 344| `/gpu/cuda/ref` | Reference pure CUDA kernels | 345| `/gpu/cuda/reg` | Pure CUDA kernels using one thread per element | 346| `/gpu/magma` | CUDA MAGMA kernels | 347 348Examples available in this release: 349 350:::{list-table} 351:header-rows: 1 352:widths: auto 353* - User code 354 - Example 355* - `ceed` 356 - * ex1 (volume) 357* - `mfem` 358 - * BP1 (scalar mass operator) 359 * BP3 (scalar Laplace operator) 360* - `petsc` 361 - * BP1 (scalar mass operator) 362 * BP3 (scalar Laplace operator) 363 * Navier-Stokes 364* - `nek5000` 365 - * BP1 (scalar mass operator) 366 * BP3 (scalar Laplace operator) 367::: 368 369(v0-3)= 370 371## v0.3 (Sep 30, 2018) 372 373Notable features in this release include active/passive field interface, support for 374non-tensor bases, backend optimization, and improved Fortran interface. This release 375also focused on providing improved continuous integration, and many new tests with code 376coverage reports of about 90%. This release also provided a significant change to the 377public interface: a {ref}`CeedQFunction` can take any number of named input and output 378arguments while {ref}`CeedOperator` connects them to the actual data, which may be 379supplied explicitly to `CeedOperatorApply()` (active) or separately via 380`CeedOperatorSetField()` (passive). This interface change enables reusable libraries 381of CeedQFunctions and composition of block solvers constructed using 382{ref}`CeedOperator`. A concept of blocked restriction was added to this release and 383used in an optimized CPU backend. Although this is typically not visible to the user, 384it enables effective use of arbitrary-length SIMD while maintaining cache locality. 385This CPU backend also implements an algebraic factorization of tensor product gradients 386to perform fewer operations than standard application of interpolation and 387differentiation from nodes to quadrature points. This algebraic formulation 388automatically supports non-polynomial and non-interpolatory bases, thus is more general 389than the more common derivation in terms of Lagrange polynomials on the quadrature points. 390 391Backends available in this release: 392 393| CEED resource (`-ceed`) | Backend | 394|-------------------------|-----------------------------------------------------| 395| `/cpu/self/blocked` | Blocked reference implementation | 396| `/cpu/self/ref` | Serial reference implementation | 397| `/cpu/self/tmpl` | Backend template, defaults to `/cpu/self/blocked` | 398| `/cpu/occa` | Serial OCCA kernels | 399| `/gpu/occa` | CUDA OCCA kernels | 400| `/omp/occa` | OpenMP OCCA kernels | 401| `/ocl/occa` | OpenCL OCCA kernels | 402| `/gpu/magma` | CUDA MAGMA kernels | 403 404Examples available in this release: 405 406:::{list-table} 407:header-rows: 1 408:widths: auto 409* - User code 410 - Example 411* - `ceed` 412 - * ex1 (volume) 413* - `mfem` 414 - * BP1 (scalar mass operator) 415 * BP3 (scalar Laplace operator) 416* - `petsc` 417 - * BP1 (scalar mass operator) 418 * BP3 (scalar Laplace operator) 419* - `nek5000` 420 - * BP1 (scalar mass operator) 421 * BP3 (scalar Laplace operator) 422::: 423 424(v0-21)= 425 426## v0.21 (Sep 30, 2018) 427 428A MAGMA backend (which relies upon the 429[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this 430release. This initial integration set up the framework of using MAGMA and provided the 431libCEED functionality through MAGMA kernels as one of libCEED’s computational backends. 432As any other backend, the MAGMA backend provides extended basic data structures for 433{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements 434the fundamental CEED building blocks to work with the new data structures. 435In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data 436but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled 437internally, and thus seamlessly to the user, through the functions/methods that are 438provided to support them. 439 440Backends available in this release: 441 442| CEED resource (`-ceed`) | Backend | 443|-------------------------|---------------------------------| 444| `/cpu/self` | Serial reference implementation | 445| `/cpu/occa` | Serial OCCA kernels | 446| `/gpu/occa` | CUDA OCCA kernels | 447| `/omp/occa` | OpenMP OCCA kernels | 448| `/ocl/occa` | OpenCL OCCA kernels | 449| `/gpu/magma` | CUDA MAGMA kernels | 450 451Examples available in this release: 452 453:::{list-table} 454:header-rows: 1 455:widths: auto 456* - User code 457 - Example 458* - `ceed` 459 - * ex1 (volume) 460* - `mfem` 461 - * BP1 (scalar mass operator) 462 * BP3 (scalar Laplace operator) 463* - `petsc` 464 - * BP1 (scalar mass operator) 465* - `nek5000` 466 - * BP1 (scalar mass operator) 467::: 468 469(v0-2)= 470 471## v0.2 (Mar 30, 2018) 472 473libCEED was made publicly available the first full CEED software distribution, release 474CEED 1.0. The distribution was made available using the Spack package manager to provide 475a common, easy-to-use build environment, where the user can build the CEED distribution 476with all dependencies. This release included a new Fortran interface for the library. 477This release also contained major improvements in the OCCA backend (including a new 478`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to 479compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an 480`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application 481of the Laplace operator) was also added to this release. 482 483Backends available in this release: 484 485| CEED resource (`-ceed`) | Backend | 486|-------------------------|---------------------------------| 487| `/cpu/self` | Serial reference implementation | 488| `/cpu/occa` | Serial OCCA kernels | 489| `/gpu/occa` | CUDA OCCA kernels | 490| `/omp/occa` | OpenMP OCCA kernels | 491| `/ocl/occa` | OpenCL OCCA kernels | 492 493Examples available in this release: 494 495:::{list-table} 496:header-rows: 1 497:widths: auto 498* - User code 499 - Example 500* - `ceed` 501 - * ex1 (volume) 502* - `mfem` 503 - * BP1 (scalar mass operator) 504 * BP3 (scalar Laplace operator) 505* - `petsc` 506 - * BP1 (scalar mass operator) 507* - `nek5000` 508 - * BP1 (scalar mass operator) 509::: 510 511(v0-1)= 512 513## v0.1 (Jan 3, 2018) 514 515Initial low-level API of the CEED project. The low-level API provides a set of Finite 516Elements kernels and components for writing new low-level kernels. Examples include: 517vector and sparse linear algebra, element matrix assembly over a batch of elements, 518partial assembly and action for efficient high-order operators like mass, diffusion, 519advection, etc. The main goal of the low-level API is to establish the basis for the 520high-level API. Also, identifying such low-level kernels and providing a reference 521implementation for them serves as the basis for specialized backend implementations. 522This release contained several backends: `/cpu/self`, and backends which rely upon the 523[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`, 524`/gpu/occa`, and `/omp/occa`. 525It also included several examples, in the `examples` folder: 526A standalone code that shows the usage of libCEED (with no external 527dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1 528(with the application of the mass operator); and a `petsc` example to perform BP1 529(with the application of the mass operator). 530 531Backends available in this release: 532 533| CEED resource (`-ceed`) | Backend | 534|-------------------------|---------------------------------| 535| `/cpu/self` | Serial reference implementation | 536| `/cpu/occa` | Serial OCCA kernels | 537| `/gpu/occa` | CUDA OCCA kernels | 538| `/omp/occa` | OpenMP OCCA kernels | 539 540Examples available in this release: 541 542| User code | Example | 543|-----------------------|-----------------------------------| 544| `ceed` | ex1 (scalar Laplace operator) | 545| `mfem` | BP1 (scalar mass operator) | 546| `petsc` | BP1 (scalar mass operator) | 547``` 548