xref: /libCEED/doc/sphinx/source/releasenotes.md (revision 43e1b16f0c7c5dbfae04543c0675e255f265ff67)
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
21### New features
22
23- `CeedScalar` can now be set as `float` or `double` at compile time.
24
25### Maintainability
26
27- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage.
28- `Include-what-you-use` makefile target added as `make iwyu`.
29
30(v0-9)=
31
32## v0.9 (Jul 6, 2021)
33
34### Interface changes
35
36- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact.
37
38### New features
39
40- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use.
41- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends.
42- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations.
43- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed).
44
45### Performance improvements
46
47### Examples
48
49- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations.
50- Fluid mechanics example adds GPU support and improves modularity.
51
52### Deprecated backends
53
54- 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.
55
56(v0-8)=
57
58## v0.8 (Mar 31, 2021)
59
60### Interface changes
61
62- Error handling improved to include enumerated error codes for C interface return values.
63- 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.
64
65### New features
66
67- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features.
68- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly.
69- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators.
70
71### Performance improvements
72
73- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`.
74- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`.
75
76### Examples
77
78- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions.
79- {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.
80- {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.
81- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree.
82
83(v0-7)=
84
85## v0.7 (Sep 29, 2020)
86
87### Interface changes
88
89- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors.
90  As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`.
91  These changes improve support for mixed finite element methods.
92- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature.
93- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity.
94- 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.
95- 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.
96- 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.
97- 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.
98  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`.
99- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory.
100- 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.
101
102### New features
103
104- New HIP backend: `/gpu/hip/ref`.
105- CeedQFunction support for user `CUfunction`s in some backends
106
107### Performance improvements
108
109- OCCA backend rebuilt to facilitate future performance enhancements.
110- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`.
111
112### Examples
113
114- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions.
115
116### Deprecated backends
117
118- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`.
119
120(v0-6)=
121
122## v0.6 (Mar 29, 2020)
123
124libCEED v0.6 contains numerous new features and examples, as well as expanded
125documentation in [this new website](https://libceed.readthedocs.io).
126
127### New features
128
129- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly
130  1-1 correspondence with the C interface, plus some convenience features.  For instance,
131  data stored in the {cpp:type}`CeedVector` structure are available without copy as
132  {py:class}`numpy.ndarray`.  Short tutorials are provided in
133  [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/).
134- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point,
135  {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal
136  ({c:func}`CeedOperatorAssembleLinearDiagonal`).  These operations are useful for
137  preconditioning ingredients and are used in the libCEED's multigrid examples.
138- The inverse of separable operators can be obtained using
139  {c:func}`CeedOperatorCreateFDMElementInverse` and applied with
140  {c:func}`CeedOperatorApply`.  This is a useful preconditioning ingredient,
141  especially for Laplacians and related operators.
142- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`,
143  {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`.
144- Make public accessors for various attributes to facilitate writing composable code.
145- New backend: `/cpu/self/memcheck/serial`.
146- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA
147  backends.  (Single source is coming soon for OCCA backends.)
148- Fix some missing edge cases in CUDA backend.
149
150### Performance Improvements
151
152- MAGMA backend performance optimization and non-tensor bases.
153- No-copy optimization in {c:func}`CeedOperatorApply`.
154
155### Interface changes
156
157- Replace {code}`CeedElemRestrictionCreateIdentity` and
158  {code}`CeedElemRestrictionCreateBlocked` with more flexible
159  {c:func}`CeedElemRestrictionCreateStrided` and
160  {c:func}`CeedElemRestrictionCreateBlockedStrided`.
161- Add arguments to {c:func}`CeedQFunctionCreateIdentity`.
162- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification
163  with {cpp:enum}`CeedInterlaceMode`.  This is now an attribute of the
164  {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no
165  longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and
166  {c:func}`CeedElemRestrictionApply`.
167
168### Examples
169
170libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`.
171Notable additions include:
172
173- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of
174  a domain in 1, 2, and 3 dimensions by applying a Laplacian.
175
176- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area
177  of domains (like the cube and sphere) by direct integration on a surface mesh;
178  demonstrates geometric dimension different from topological dimension.
179
180- PETSc {ref}`example-petsc-bps`:
181
182  - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support.
183  - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements
184    and transparent CUDA support.
185  - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`):
186    generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric
187    dimension different from topological dimension.
188
189- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid
190  solver with algebraic multigrid coarse solve.
191
192- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly
193  `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`),
194  implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and
195  quasi-2D computational domain support.
196
197- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for
198  linear elasticity, small-strain hyperelasticity, and globalized finite-strain
199  hyperelasticity using p-multigrid with algebraic multigrid coarse solve.
200
201(v0-5)=
202
203## v0.5 (Sep 18, 2019)
204
205For this release, several improvements were made. Two new CUDA backends were added to
206the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art
207performance using single-source {ref}`CeedQFunction`. From this release, users
208can define Q-Functions in a single source code independently of the targeted backend
209with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU
210compilation of the user provided {ref}`CeedQFunction` code. To allow a unified
211declaration, the {ref}`CeedQFunction` API has undergone a slight change:
212the `QFunctionField` parameter `ncomp` has been changed to `size`. This change
213requires setting the previous value of `ncomp` to `ncomp*dim` when adding a
214`QFunctionField` with eval mode `CEED EVAL GRAD`.
215
216Additionally, new CPU backends
217were included in this release, such as the `/cpu/self/opt/*` backends (which are
218written in pure C and use partial **E-vectors** to improve performance) and the
219`/cpu/self/ref/memcheck` backend (which relies upon the
220[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user
221{ref}`CeedQFunction` have no undefined values).
222This release also included various performance improvements, bug fixes, new examples,
223and improved tests. Among these improvements, vectorized instructions for
224{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD`
225instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and
226identity Q-Functions were introduced, and the PETSc benchmark problems were expanded
227to include unstructured meshes handling were. For this expansion, the prior version of
228the PETSc BPs, which only included data associated with structured geometries, were
229renamed `bpsraw`, and the new version of the BPs, which can handle data associated
230with any unstructured geometry, were called `bps`. Additionally, other benchmark
231problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively),
232and BP5 and BP6 (the collocated versions---for which the quadrature points are the same
233as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc
234examples. Furthermoew, another standalone libCEED example, called `ex2`, which
235computes the surface area of a given mesh was added to this release.
236
237Backends available in this release:
238
239| CEED resource (`-ceed`)  | Backend                                             |
240|--------------------------|-----------------------------------------------------|
241| `/cpu/self/ref/serial`   | Serial reference implementation                     |
242| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
243| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks            |
244| `/cpu/self/opt/serial`   | Serial optimized C implementation                   |
245| `/cpu/self/opt/blocked`  | Blocked optimized C implementation                  |
246| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
247| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
248| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
249| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
250| `/cpu/occa`              | Serial OCCA kernels                                 |
251| `/gpu/occa`              | CUDA OCCA kernels                                   |
252| `/omp/occa`              | OpenMP OCCA kernels                                 |
253| `/ocl/occa`              | OpenCL OCCA kernels                                 |
254| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
255| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
256| `/gpu/cuda/shared`       | Optimized pure CUDA kernels using shared memory     |
257| `/gpu/cuda/gen`          | Optimized pure CUDA kernels using code generation   |
258| `/gpu/magma`             | CUDA MAGMA kernels                                  |
259
260Examples available in this release:
261
262:::{list-table}
263:header-rows: 1
264:widths: auto
265* - User code
266  - Example
267* - `ceed`
268  - * ex1 (volume)
269    * ex2 (surface)
270* - `mfem`
271  - * BP1 (scalar mass operator)
272    * BP3 (scalar Laplace operator)
273* - `petsc`
274  - * BP1 (scalar mass operator)
275    * BP2 (vector mass operator)
276    * BP3 (scalar Laplace operator)
277    * BP4 (vector Laplace operator)
278    * BP5 (collocated scalar Laplace operator)
279    * BP6 (collocated vector Laplace operator)
280    * Navier-Stokes
281* - `nek5000`
282  - * BP1 (scalar mass operator)
283    * BP3 (scalar Laplace operator)
284:::
285
286(v0-4)=
287
288## v0.4 (Apr 1, 2019)
289
290libCEED v0.4 was made again publicly available in the second full CEED software
291distribution, release CEED 2.0. This release contained notable features, such as
292four new CPU backends, two new GPU backends, CPU backend optimizations, initial
293support for operator composition, performance benchmarking, and a Navier-Stokes demo.
294The new CPU backends in this release came in two families. The `/cpu/self/*/serial`
295backends process one element at a time and are intended for meshes with a smaller number
296of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of
297eight interlaced elements and are intended for meshes with higher numbers of elements.
298The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU
299performance. The `/cpu/self/xsmm/*` backends rely upon the
300[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU
301performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA.
302The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable
303performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple
304parallelization approach, where each thread treats a finite element. Using just in time
305compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this
306backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend
307achieve good peak performance for 1D, 2D, and low order 3D problems, but performance
308deteriorates very quickly when threads run out of registers.
309
310A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED
311examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`).
312This example solves the time-dependent Navier-Stokes equations of compressible gas
313dynamics in a static Eulerian three-dimensional frame, using structured high-order
314finite/spectral element spatial discretizations and explicit high-order time-stepping
315(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc,
316so that the pointwise physics (defined at quadrature points) is separated from the
317parallelization and meshing concerns.
318
319Backends available in this release:
320
321| CEED resource (`-ceed`)  | Backend                                             |
322|--------------------------|-----------------------------------------------------|
323| `/cpu/self/ref/serial`   | Serial reference implementation                     |
324| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
325| `/cpu/self/tmpl`         | Backend template, defaults to `/cpu/self/blocked`   |
326| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
327| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
328| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
329| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
330| `/cpu/occa`              | Serial OCCA kernels                                 |
331| `/gpu/occa`              | CUDA OCCA kernels                                   |
332| `/omp/occa`              | OpenMP OCCA kernels                                 |
333| `/ocl/occa`              | OpenCL OCCA kernels                                 |
334| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
335| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
336| `/gpu/magma`             | CUDA MAGMA kernels                                  |
337
338Examples available in this release:
339
340:::{list-table}
341:header-rows: 1
342:widths: auto
343* - User code
344  - Example
345* - `ceed`
346  - * ex1 (volume)
347* - `mfem`
348  - * BP1 (scalar mass operator)
349    * BP3 (scalar Laplace operator)
350* - `petsc`
351  - * BP1 (scalar mass operator)
352    * BP3 (scalar Laplace operator)
353    * Navier-Stokes
354* - `nek5000`
355  - * BP1 (scalar mass operator)
356    * BP3 (scalar Laplace operator)
357:::
358
359(v0-3)=
360
361## v0.3 (Sep 30, 2018)
362
363Notable features in this release include active/passive field interface, support for
364non-tensor bases, backend optimization, and improved Fortran interface. This release
365also focused on providing improved continuous integration, and many new tests with code
366coverage reports of about 90%. This release also provided a significant change to the
367public interface: a {ref}`CeedQFunction` can take any number of named input and output
368arguments while {ref}`CeedOperator` connects them to the actual data, which may be
369supplied explicitly to `CeedOperatorApply()` (active) or separately via
370`CeedOperatorSetField()` (passive). This interface change enables reusable libraries
371of CeedQFunctions and composition of block solvers constructed using
372{ref}`CeedOperator`. A concept of blocked restriction was added to this release and
373used in an optimized CPU backend. Although this is typically not visible to the user,
374it enables effective use of arbitrary-length SIMD while maintaining cache locality.
375This CPU backend also implements an algebraic factorization of tensor product gradients
376to perform fewer operations than standard application of interpolation and
377differentiation from nodes to quadrature points. This algebraic formulation
378automatically supports non-polynomial and non-interpolatory bases, thus is more general
379than the more common derivation in terms of Lagrange polynomials on the quadrature points.
380
381Backends available in this release:
382
383| CEED resource (`-ceed`) | Backend                                             |
384|-------------------------|-----------------------------------------------------|
385| `/cpu/self/blocked`     | Blocked reference implementation                    |
386| `/cpu/self/ref`         | Serial reference implementation                     |
387| `/cpu/self/tmpl`        | Backend template, defaults to `/cpu/self/blocked`   |
388| `/cpu/occa`             | Serial OCCA kernels                                 |
389| `/gpu/occa`             | CUDA OCCA kernels                                   |
390| `/omp/occa`             | OpenMP OCCA kernels                                 |
391| `/ocl/occa`             | OpenCL OCCA kernels                                 |
392| `/gpu/magma`            | CUDA MAGMA kernels                                  |
393
394Examples available in this release:
395
396:::{list-table}
397:header-rows: 1
398:widths: auto
399* - User code
400  - Example
401* - `ceed`
402  - * ex1 (volume)
403* - `mfem`
404  - * BP1 (scalar mass operator)
405    * BP3 (scalar Laplace operator)
406* - `petsc`
407  - * BP1 (scalar mass operator)
408    * BP3 (scalar Laplace operator)
409* - `nek5000`
410  - * BP1 (scalar mass operator)
411    * BP3 (scalar Laplace operator)
412:::
413
414(v0-21)=
415
416## v0.21 (Sep 30, 2018)
417
418A MAGMA backend (which relies upon the
419[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this
420release. This initial integration set up the framework of using MAGMA and provided the
421libCEED functionality through MAGMA kernels as one of libCEED’s computational backends.
422As any other backend, the MAGMA backend provides extended basic data structures for
423{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements
424the fundamental CEED building blocks to work with the new data structures.
425In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data
426but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled
427internally, and thus seamlessly to the user, through the functions/methods that are
428provided to support them.
429
430Backends available in this release:
431
432| CEED resource (`-ceed`) | Backend                         |
433|-------------------------|---------------------------------|
434| `/cpu/self`             | Serial reference implementation |
435| `/cpu/occa`             | Serial OCCA kernels             |
436| `/gpu/occa`             | CUDA OCCA kernels               |
437| `/omp/occa`             | OpenMP OCCA kernels             |
438| `/ocl/occa`             | OpenCL OCCA kernels             |
439| `/gpu/magma`            | CUDA MAGMA kernels              |
440
441Examples available in this release:
442
443:::{list-table}
444:header-rows: 1
445:widths: auto
446* - User code
447  - Example
448* - `ceed`
449  - * ex1 (volume)
450* - `mfem`
451  - * BP1 (scalar mass operator)
452    * BP3 (scalar Laplace operator)
453* - `petsc`
454  - * BP1 (scalar mass operator)
455* - `nek5000`
456  - * BP1 (scalar mass operator)
457:::
458
459(v0-2)=
460
461## v0.2 (Mar 30, 2018)
462
463libCEED was made publicly available the first full CEED software distribution, release
464CEED 1.0. The distribution was made available using the Spack package manager to provide
465a common, easy-to-use build environment, where the user can build the CEED distribution
466with all dependencies. This release included a new Fortran interface for the library.
467This release also contained major improvements in the OCCA backend (including a new
468`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to
469compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an
470`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application
471of the Laplace operator) was also added to this release.
472
473Backends available in this release:
474
475| CEED resource (`-ceed`) | Backend                         |
476|-------------------------|---------------------------------|
477| `/cpu/self`             | Serial reference implementation |
478| `/cpu/occa`             | Serial OCCA kernels             |
479| `/gpu/occa`             | CUDA OCCA kernels               |
480| `/omp/occa`             | OpenMP OCCA kernels             |
481| `/ocl/occa`             | OpenCL OCCA kernels             |
482
483Examples available in this release:
484
485:::{list-table}
486:header-rows: 1
487:widths: auto
488* - User code
489  - Example
490* - `ceed`
491  - * ex1 (volume)
492* - `mfem`
493  - * BP1 (scalar mass operator)
494    * BP3 (scalar Laplace operator)
495* - `petsc`
496  - * BP1 (scalar mass operator)
497* - `nek5000`
498  - * BP1 (scalar mass operator)
499:::
500
501(v0-1)=
502
503## v0.1 (Jan 3, 2018)
504
505Initial low-level API of the CEED project. The low-level API provides a set of Finite
506Elements kernels and components for writing new low-level kernels. Examples include:
507vector and sparse linear algebra, element matrix assembly over a batch of elements,
508partial assembly and action for efficient high-order operators like mass, diffusion,
509advection, etc. The main goal of the low-level API is to establish the basis for the
510high-level API. Also, identifying such low-level kernels and providing a reference
511implementation for them serves as the basis for specialized backend implementations.
512This release contained several backends: `/cpu/self`, and backends which rely upon the
513[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`,
514`/gpu/occa`, and `/omp/occa`.
515It also included several examples, in the `examples` folder:
516A standalone code that shows the usage of libCEED (with no external
517dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1
518(with the application of the mass operator); and a `petsc` example to perform BP1
519(with the application of the mass operator).
520
521Backends available in this release:
522
523| CEED resource (`-ceed`) | Backend                         |
524|-------------------------|---------------------------------|
525| `/cpu/self`             | Serial reference implementation |
526| `/cpu/occa`             | Serial OCCA kernels             |
527| `/gpu/occa`             | CUDA OCCA kernels               |
528| `/omp/occa`             | OpenMP OCCA kernels             |
529
530Examples available in this release:
531
532| User code             | Example                           |
533|-----------------------|-----------------------------------|
534| `ceed`                | ex1 (scalar Laplace operator)     |
535| `mfem`                | BP1 (scalar mass operator)        |
536| `petsc`               | BP1 (scalar mass operator)        |
537```
538