xref: /libCEED/doc/sphinx/source/releasenotes.md (revision 49a40c8a2d720db341b0b117b89656b473cbebfb)
1# Changes/Release Notes
2
3On this page we provide a summary of the main API changes, new features and examples for each release of libCEED.
4
5(main)=
6
7## Current `main` branch
8
9### Interface changes
10
11### New features
12
13### Examples
14
15(v0-12)=
16
17## v0.12 (Oct 31, 2023)
18
19### Interface changes
20
21- Update `CeedOperatorContext*` functions to `CeedOperator*Context*` functions for consistency.
22For example, `CeedOperatorContextGetFieldLabel` was renamed to `CeedOperatorGetContextFieldLabel`.
23- Removed `CeedBasisSetNumQuadraturePoints` as redundant and bug-prone interface.
24
25### New features
26
27- Added {c:func}`CeedOperatorGetFieldByName` to access a specific `CeedOperatorField` by its name.
28- Update `/cpu/self/memcheck/*` backends to help verify `CeedVector` array access assumptions and `CeedQFunction` user output assumptions.
29- Update {c:func}`CeedOperatorLinearAssembleDiagonal` to provide default implementation that supports `CeedOperator` with multiple active bases.
30- Added Sycl backends `/gpu/sycl/ref`, `/gpu/sycl/shared`, and `/gpu/sycl/gen`.
31- Added {c:func}`CeedBasisApplyAtPoints` for evaluation of values and derivatives at arbitrary points inside elements.
32- Added support for non-tensor $H(\text{curl})$ finite element spaces with {c:func}`CeedBasisCreateHcurl`.
33- Added {c:func}`CeedElemRestrictionCreateCurlOriented`, similar to {c:func}`CeedElemRestrictionCreateOriented`, for element restrictions requiring more general element transformations such as those for high-order $H(\text{curl})$ spaces on tetrahedra (see [https://dl.acm.org/doi/pdf/10.1145/3524456](https://dl.acm.org/doi/pdf/10.1145/3524456)).
34- Added {c:func}`CeedOperatorLinearAssemblePointBlockDiagonalSymbolic` to create COO mapping for mapping out of {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`.
35- Added support for application codes which manage multiple {ref}`Ceed` objects, parallelized across OpenMP threads.
36
37### Examples
38
39- Add `DMSwarm` example demonstrating interpolation from background mesh to swarm points and projection from swarm points to background mesh.
40
41#### {ref}`example-petsc-bps`
42
43- Requires PETSc version 3.19 or later.
44
45#### {ref}`example-petsc-navier-stokes`
46
47- Updated restart and checkpointing interface.
48- Add data-driven subgrid-stress model.
49- Add differential filtering of solution.
50- Add turbulence statistics collection over spanwise-symmetric geometries.
51- Add Taylor-Green vortex initial condition.
52- Add Riemann-based outflow boundary conditions.
53- Added vortex shedding and flow past cylinder example, including calculations for lift, drag, and heat transfer.
54- Add Internal Damping Layer (IDL) for helping turbulent simulation stability.
55- Derive `CeedBasis` from `PetscFE`, and various other internal maintainability updates.
56
57(v0-11)=
58
59## v0.11 (Dec 24, 2022)
60
61### Interface changes
62
63- Added {c:func}`CeedOperatorSetName` for more readable {c:func}`CeedOperatorView` output.
64- Added {c:func}`CeedBasisCreateProjection` to facilitate interpolation between nodes for separate `CeedBases`.
65- Rename and move {c:func}`CeedCompositeOperatorGetNumSub` and {c:func}`CeedCompositeOperatorGetSubList` to public interface.
66- Renamed `CEED_BASIS_COLLOCATED` to `CEED_BASIS_NONE` for clarity.
67Some users previously misinterpreted a `CeedOperator` field using `CEED_BASIS_COLLOCATED` as meaning that the entire `CeedOperator` used a quadrature space that is collocated with the nodal space of the active bases.
68
69### New features
70
71- Update `/cpu/self/memcheck/*` backends to help verify `CeedQFunctionContext` data sizes provided by user.
72- Improved support for $H(\text{div})$ bases.
73- Added `CeedInt_FMT` to support potential future use of larger integer sizes.
74- Added `CEED_QFUNCTION_ATTR` for setting compiler attributes/pragmas to `CEED_QFUNCTION_HELPER` and `CEED_QFUNCTION`.
75- OCCA backend updated to latest OCCA release; DPC++ and OMP OCCA modes enabled.
76Due to a limitation of the OCCA parser, typedefs are required to use pointers to arrays in QFunctions with the OCCA backend.
77This issue will be fixed in a future OCCA release.
78
79### Bugfix
80
81- Fix bug in setting device id for GPU backends.
82- Fix storing of indices for `CeedElemRestriction` on the host with GPU backends.
83- Fix `CeedElemRestriction` sizing for {c:func}`CeedOperatorAssemblePointBlockDiagonal`.
84- Fix bugs in CPU implementation of {c:func}`CeedOperatorLinearAssemble` when there are different number of active input modes and active output modes.
85
86### Examples
87
88#### {ref}`example-petsc-navier-stokes`
89
90- Various performance enhancements, analytic matrix-free and assembled Jacobian, and PETSc solver configurations for GPUs.
91- Refactored to improve code reuse and modularity.
92- Support for primitive variables for more accurate boundary layers and all-speed flow.
93- Added $YZ\beta$ shock capturing scheme and Shock Tube example.
94- Added Channel example, with comparison to analytic solutions.
95- Added Flat Plate with boundary layer mesh and compressible Blasius inflow condition based on Chebyshev collocation solution of the Blasius equations.
96- Added strong and weak synthetic turbulence generation (STG) inflow boundary conditions.
97- Added "freestream" boundary conditions based on HLLC Riemann solver.
98- Automated stabilization coefficients for different basis degree.
99
100#### {ref}`example-petsc-bps`
101
102- Support for convergence studies.
103
104### Maintainability
105
106- Refactored `/gpu/cuda/shared` and `/gpu/cuda/gen` as well as `/gpu/hip/shared` and `/gpu/hip/gen` backend to improve maintainablity and reduce duplicated code.
107- Enabled support for `p > 8` for `/gpu/*/shared` backends.
108- Switch to `clang-format` over `astyle` for automatic formatting; Makefile command changed to `make format` from `make style`.
109- Improved test harness.
110
111(v0-10-1)=
112
113## v0.10.1 (Apr 11, 2022)
114
115### Interface changes
116
117- Added {c:func}`CeedQFunctionSetUserFlopsEstimate` and {c:func}`CeedOperatorGetFlopsEstimate` to facilitate estimating FLOPs in operator application.
118
119### New features
120
121- Switched MAGMA backends to use runtime compilation for tensor basis kernels (and element restriction kernels, in non-deterministic `/gpu/*/magma` backends).
122This reduces time to compile the library and increases the range of parameters for which the MAGMA tensor basis kernels will work.
123
124### Bugfix
125
126- Install JiT source files in install directory to fix GPU functionality for installed libCEED.
127
128(v0-10)=
129
130## v0.10 (Mar 21, 2022)
131
132### Interface changes
133
134- Update {c:func}`CeedQFunctionGetFields` and {c:func}`CeedOperatorGetFields` to include number of fields.
135- 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`.
136- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added.
137- Add {c:func}`CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients.
138- Promote {c:func}`CeedOperatorCheckReady`to the public API to facilitate interactive interfaces.
139- Warning added when compiling OCCA backend to alert users that this backend is experimental.
140- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`.
141- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name.
142- 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`.
143- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitialized arrays; require initialized data for {c:func}`CeedVectorGetArray`.
144- Added {c:func}`CeedQFunctionContextRegisterDouble` and {c:func}`CeedQFunctionContextRegisterInt32` with {c:func}`CeedQFunctionContextSetDouble` and {c:func}`CeedQFunctionContextSetInt32` to facilitate easy updating of {c:struct}`CeedQFunctionContext` data by user defined field names.
145- Added {c:func}`CeedQFunctionContextGetFieldDescriptions` to retrieve user defined descriptions of fields that are registered with `CeedQFunctionContextRegister*`.
146- Renamed `CeedElemTopology` entries for clearer namespacing between libCEED enums.
147- Added type `CeedSize` equivalent to `ptrdiff_t` for array sizes in {c:func}`CeedVectorCreate`, {c:func}`CeedVectorGetLength`, `CeedElemRestrictionCreate*`, {c:func}`CeedElemRestrictionGetLVectorSize`, and {c:func}`CeedOperatorLinearAssembleSymbolic`. This is a breaking change.
148- Added {c:func}`CeedOperatorSetQFunctionUpdated` to facilitate QFunction data re-use between operators sharing the same quadrature space, such as in a multigrid hierarchy.
149- Added {c:func}`CeedOperatorGetActiveVectorLengths` to get shape of CeedOperator.
150
151### New features
152
153- `CeedScalar` can now be set as `float` or `double` at compile time.
154- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends.
155- 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.
156- Remove need to guard library headers in QFunction source for code generation backends.
157- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present.
158- Added {c:func}`CeedStringAllocCopy` to reduce repeated code for copying strings internally.
159- Added {c:func}`CeedPathConcatenate` to facilitate loading kernel source files with a path relative to the current file.
160- Added support for non-tensor $H(\text{div})$ elements, to include CPU backend implementations and {c:func}`CeedBasisCreateHdiv` convenience constructor.
161- Added {c:func}`CeedQFunctionSetContextWritable` and read-only access to `CeedQFunctionContext` data as an optional feature to improve GPU performance. By default, calling the `CeedQFunctionUser` during {c:func}`CeedQFunctionApply` is assumed to write into the `CeedQFunctionContext` data, consistent with the previous behavior. Note that if a user asserts that their `CeedQFunctionUser` does not write into the `CeedQFunctionContext` data, they are responsible for the validity of this assertion.
162- Added support for element matrix assembly in GPU backends.
163
164### Maintainability
165
166- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage.
167- `Include-what-you-use` makefile target added as `make iwyu`.
168- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase.
169- Put GPU JiTed kernel source code into separate files.
170- Dropped legacy version support in PETSc based examples to better utilize PETSc DMPlex and Mat updates to support libCEED; current minimum PETSc version for the examples is v3.17.
171
172(v0-9)=
173
174## v0.9 (Jul 6, 2021)
175
176### Interface changes
177
178- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact.
179
180### New features
181
182- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use.
183- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends.
184- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations.
185- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed).
186
187### Performance improvements
188
189### Examples
190
191- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations.
192- Fluid mechanics example adds GPU support and improves modularity.
193
194### Deprecated backends
195
196- 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.
197
198(v0-8)=
199
200## v0.8 (Mar 31, 2021)
201
202### Interface changes
203
204- Error handling improved to include enumerated error codes for C interface return values.
205- 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.
206
207### New features
208
209- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features.
210- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly.
211- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators.
212
213### Performance improvements
214
215- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`.
216- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`.
217
218### Examples
219
220- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions.
221- {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.
222- {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.
223- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree.
224
225(v0-7)=
226
227## v0.7 (Sep 29, 2020)
228
229### Interface changes
230
231- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors.
232  As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`.
233  These changes improve support for mixed finite element methods.
234- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature.
235- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity.
236- 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.
237- 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.
238- 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.
239- 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.
240  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`.
241- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory.
242- 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.
243
244### New features
245
246- New HIP backend: `/gpu/hip/ref`.
247- CeedQFunction support for user `CUfunction`s in some backends
248
249### Performance improvements
250
251- OCCA backend rebuilt to facilitate future performance enhancements.
252- Petsc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`.
253
254### Examples
255
256- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions.
257
258### Deprecated backends
259
260- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`.
261
262(v0-6)=
263
264## v0.6 (Mar 29, 2020)
265
266libCEED v0.6 contains numerous new features and examples, as well as expanded
267documentation in [this new website](https://libceed.org).
268
269### New features
270
271- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly
272  1-1 correspondence with the C interface, plus some convenience features.  For instance,
273  data stored in the {cpp:type}`CeedVector` structure are available without copy as
274  {py:class}`numpy.ndarray`.  Short tutorials are provided in
275  [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/).
276- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point,
277  {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal
278  ({c:func}`CeedOperatorAssembleLinearDiagonal`).  These operations are useful for
279  preconditioning ingredients and are used in the libCEED's multigrid examples.
280- The inverse of separable operators can be obtained using
281  {c:func}`CeedOperatorCreateFDMElementInverse` and applied with
282  {c:func}`CeedOperatorApply`.  This is a useful preconditioning ingredient,
283  especially for Laplacians and related operators.
284- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`,
285  {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`.
286- Make public accessors for various attributes to facilitate writing composable code.
287- New backend: `/cpu/self/memcheck/serial`.
288- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA
289  backends.  (Single source is coming soon for OCCA backends.)
290- Fix some missing edge cases in CUDA backend.
291
292### Performance Improvements
293
294- MAGMA backend performance optimization and non-tensor bases.
295- No-copy optimization in {c:func}`CeedOperatorApply`.
296
297### Interface changes
298
299- Replace {code}`CeedElemRestrictionCreateIdentity` and
300  {code}`CeedElemRestrictionCreateBlocked` with more flexible
301  {c:func}`CeedElemRestrictionCreateStrided` and
302  {c:func}`CeedElemRestrictionCreateBlockedStrided`.
303- Add arguments to {c:func}`CeedQFunctionCreateIdentity`.
304- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification
305  with {cpp:enum}`CeedInterlaceMode`.  This is now an attribute of the
306  {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no
307  longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and
308  {c:func}`CeedElemRestrictionApply`.
309
310### Examples
311
312libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`.
313Notable additions include:
314
315- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of
316  a domain in 1, 2, and 3 dimensions by applying a Laplacian.
317
318- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area
319  of domains (like the cube and sphere) by direct integration on a surface mesh;
320  demonstrates geometric dimension different from topological dimension.
321
322- PETSc {ref}`example-petsc-bps`:
323
324  - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support.
325  - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements
326    and transparent CUDA support.
327  - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`):
328    generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric
329    dimension different from topological dimension.
330
331- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid
332  solver with algebraic multigrid coarse solve.
333
334- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly
335  `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`),
336  implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and
337  quasi-2D computational domain support.
338
339- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for
340  linear elasticity, small-strain hyperelasticity, and globalized finite-strain
341  hyperelasticity using p-multigrid with algebraic multigrid coarse solve.
342
343(v0-5)=
344
345## v0.5 (Sep 18, 2019)
346
347For this release, several improvements were made. Two new CUDA backends were added to
348the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art
349performance using single-source {ref}`CeedQFunction`. From this release, users
350can define Q-Functions in a single source code independently of the targeted backend
351with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU
352compilation of the user provided {ref}`CeedQFunction` code. To allow a unified
353declaration, the {ref}`CeedQFunction` API has undergone a slight change:
354the `QFunctionField` parameter `ncomp` has been changed to `size`. This change
355requires setting the previous value of `ncomp` to `ncomp*dim` when adding a
356`QFunctionField` with eval mode `CEED EVAL GRAD`.
357
358Additionally, new CPU backends
359were included in this release, such as the `/cpu/self/opt/*` backends (which are
360written in pure C and use partial **E-vectors** to improve performance) and the
361`/cpu/self/ref/memcheck` backend (which relies upon the
362[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user
363{ref}`CeedQFunction` have no undefined values).
364This release also included various performance improvements, bug fixes, new examples,
365and improved tests. Among these improvements, vectorized instructions for
366{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD`
367instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and
368identity Q-Functions were introduced, and the PETSc benchmark problems were expanded
369to include unstructured meshes handling were. For this expansion, the prior version of
370the PETSc BPs, which only included data associated with structured geometries, were
371renamed `bpsraw`, and the new version of the BPs, which can handle data associated
372with any unstructured geometry, were called `bps`. Additionally, other benchmark
373problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively),
374and BP5 and BP6 (the collocated versions---for which the quadrature points are the same
375as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc
376examples. Furthermoew, another standalone libCEED example, called `ex2`, which
377computes the surface area of a given mesh was added to this release.
378
379Backends available in this release:
380
381| CEED resource (`-ceed`)  | Backend                                             |
382|--------------------------|-----------------------------------------------------|
383| `/cpu/self/ref/serial`   | Serial reference implementation                     |
384| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
385| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks            |
386| `/cpu/self/opt/serial`   | Serial optimized C implementation                   |
387| `/cpu/self/opt/blocked`  | Blocked optimized C implementation                  |
388| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
389| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
390| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
391| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
392| `/cpu/occa`              | Serial OCCA kernels                                 |
393| `/gpu/occa`              | CUDA OCCA kernels                                   |
394| `/omp/occa`              | OpenMP OCCA kernels                                 |
395| `/ocl/occa`              | OpenCL OCCA kernels                                 |
396| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
397| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
398| `/gpu/cuda/shared`       | Optimized pure CUDA kernels using shared memory     |
399| `/gpu/cuda/gen`          | Optimized pure CUDA kernels using code generation   |
400| `/gpu/magma`             | CUDA MAGMA kernels                                  |
401
402Examples available in this release:
403
404:::{list-table}
405:header-rows: 1
406:widths: auto
407* - User code
408  - Example
409* - `ceed`
410  - * ex1 (volume)
411    * ex2 (surface)
412* - `mfem`
413  - * BP1 (scalar mass operator)
414    * BP3 (scalar Laplace operator)
415* - `petsc`
416  - * BP1 (scalar mass operator)
417    * BP2 (vector mass operator)
418    * BP3 (scalar Laplace operator)
419    * BP4 (vector Laplace operator)
420    * BP5 (collocated scalar Laplace operator)
421    * BP6 (collocated vector Laplace operator)
422    * Navier-Stokes
423* - `nek5000`
424  - * BP1 (scalar mass operator)
425    * BP3 (scalar Laplace operator)
426:::
427
428(v0-4)=
429
430## v0.4 (Apr 1, 2019)
431
432libCEED v0.4 was made again publicly available in the second full CEED software
433distribution, release CEED 2.0. This release contained notable features, such as
434four new CPU backends, two new GPU backends, CPU backend optimizations, initial
435support for operator composition, performance benchmarking, and a Navier-Stokes demo.
436The new CPU backends in this release came in two families. The `/cpu/self/*/serial`
437backends process one element at a time and are intended for meshes with a smaller number
438of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of
439eight interlaced elements and are intended for meshes with higher numbers of elements.
440The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU
441performance. The `/cpu/self/xsmm/*` backends rely upon the
442[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU
443performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA.
444The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable
445performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple
446parallelization approach, where each thread treats a finite element. Using just in time
447compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this
448backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend
449achieve good peak performance for 1D, 2D, and low order 3D problems, but performance
450deteriorates very quickly when threads run out of registers.
451
452A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED
453examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`).
454This example solves the time-dependent Navier-Stokes equations of compressible gas
455dynamics in a static Eulerian three-dimensional frame, using structured high-order
456finite/spectral element spatial discretizations and explicit high-order time-stepping
457(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc,
458so that the pointwise physics (defined at quadrature points) is separated from the
459parallelization and meshing concerns.
460
461Backends available in this release:
462
463| CEED resource (`-ceed`)  | Backend                                             |
464|--------------------------|-----------------------------------------------------|
465| `/cpu/self/ref/serial`   | Serial reference implementation                     |
466| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
467| `/cpu/self/tmpl`         | Backend template, defaults to `/cpu/self/blocked`   |
468| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
469| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
470| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
471| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM 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| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
477| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
478| `/gpu/magma`             | CUDA MAGMA kernels                                  |
479
480Examples available in this release:
481
482:::{list-table}
483:header-rows: 1
484:widths: auto
485* - User code
486  - Example
487* - `ceed`
488  - * ex1 (volume)
489* - `mfem`
490  - * BP1 (scalar mass operator)
491    * BP3 (scalar Laplace operator)
492* - `petsc`
493  - * BP1 (scalar mass operator)
494    * BP3 (scalar Laplace operator)
495    * Navier-Stokes
496* - `nek5000`
497  - * BP1 (scalar mass operator)
498    * BP3 (scalar Laplace operator)
499:::
500
501(v0-3)=
502
503## v0.3 (Sep 30, 2018)
504
505Notable features in this release include active/passive field interface, support for
506non-tensor bases, backend optimization, and improved Fortran interface. This release
507also focused on providing improved continuous integration, and many new tests with code
508coverage reports of about 90%. This release also provided a significant change to the
509public interface: a {ref}`CeedQFunction` can take any number of named input and output
510arguments while {ref}`CeedOperator` connects them to the actual data, which may be
511supplied explicitly to `CeedOperatorApply()` (active) or separately via
512`CeedOperatorSetField()` (passive). This interface change enables reusable libraries
513of CeedQFunctions and composition of block solvers constructed using
514{ref}`CeedOperator`. A concept of blocked restriction was added to this release and
515used in an optimized CPU backend. Although this is typically not visible to the user,
516it enables effective use of arbitrary-length SIMD while maintaining cache locality.
517This CPU backend also implements an algebraic factorization of tensor product gradients
518to perform fewer operations than standard application of interpolation and
519differentiation from nodes to quadrature points. This algebraic formulation
520automatically supports non-polynomial and non-interpolatory bases, thus is more general
521than the more common derivation in terms of Lagrange polynomials on the quadrature points.
522
523Backends available in this release:
524
525| CEED resource (`-ceed`) | Backend                                             |
526|-------------------------|-----------------------------------------------------|
527| `/cpu/self/blocked`     | Blocked reference implementation                    |
528| `/cpu/self/ref`         | Serial reference implementation                     |
529| `/cpu/self/tmpl`        | Backend template, defaults to `/cpu/self/blocked`   |
530| `/cpu/occa`             | Serial OCCA kernels                                 |
531| `/gpu/occa`             | CUDA OCCA kernels                                   |
532| `/omp/occa`             | OpenMP OCCA kernels                                 |
533| `/ocl/occa`             | OpenCL OCCA kernels                                 |
534| `/gpu/magma`            | CUDA MAGMA kernels                                  |
535
536Examples available in this release:
537
538:::{list-table}
539:header-rows: 1
540:widths: auto
541* - User code
542  - Example
543* - `ceed`
544  - * ex1 (volume)
545* - `mfem`
546  - * BP1 (scalar mass operator)
547    * BP3 (scalar Laplace operator)
548* - `petsc`
549  - * BP1 (scalar mass operator)
550    * BP3 (scalar Laplace operator)
551* - `nek5000`
552  - * BP1 (scalar mass operator)
553    * BP3 (scalar Laplace operator)
554:::
555
556(v0-21)=
557
558## v0.21 (Sep 30, 2018)
559
560A MAGMA backend (which relies upon the
561[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this
562release. This initial integration set up the framework of using MAGMA and provided the
563libCEED functionality through MAGMA kernels as one of libCEED’s computational backends.
564As any other backend, the MAGMA backend provides extended basic data structures for
565{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements
566the fundamental CEED building blocks to work with the new data structures.
567In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data
568but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled
569internally, and thus seamlessly to the user, through the functions/methods that are
570provided to support them.
571
572Backends available in this release:
573
574| CEED resource (`-ceed`) | Backend                         |
575|-------------------------|---------------------------------|
576| `/cpu/self`             | Serial reference implementation |
577| `/cpu/occa`             | Serial OCCA kernels             |
578| `/gpu/occa`             | CUDA OCCA kernels               |
579| `/omp/occa`             | OpenMP OCCA kernels             |
580| `/ocl/occa`             | OpenCL OCCA kernels             |
581| `/gpu/magma`            | CUDA MAGMA kernels              |
582
583Examples available in this release:
584
585:::{list-table}
586:header-rows: 1
587:widths: auto
588* - User code
589  - Example
590* - `ceed`
591  - * ex1 (volume)
592* - `mfem`
593  - * BP1 (scalar mass operator)
594    * BP3 (scalar Laplace operator)
595* - `petsc`
596  - * BP1 (scalar mass operator)
597* - `nek5000`
598  - * BP1 (scalar mass operator)
599:::
600
601(v0-2)=
602
603## v0.2 (Mar 30, 2018)
604
605libCEED was made publicly available the first full CEED software distribution, release
606CEED 1.0. The distribution was made available using the Spack package manager to provide
607a common, easy-to-use build environment, where the user can build the CEED distribution
608with all dependencies. This release included a new Fortran interface for the library.
609This release also contained major improvements in the OCCA backend (including a new
610`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to
611compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an
612`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application
613of the Laplace operator) was also added to this release.
614
615Backends available in this release:
616
617| CEED resource (`-ceed`) | Backend                         |
618|-------------------------|---------------------------------|
619| `/cpu/self`             | Serial reference implementation |
620| `/cpu/occa`             | Serial OCCA kernels             |
621| `/gpu/occa`             | CUDA OCCA kernels               |
622| `/omp/occa`             | OpenMP OCCA kernels             |
623| `/ocl/occa`             | OpenCL OCCA kernels             |
624
625Examples available in this release:
626
627:::{list-table}
628:header-rows: 1
629:widths: auto
630* - User code
631  - Example
632* - `ceed`
633  - * ex1 (volume)
634* - `mfem`
635  - * BP1 (scalar mass operator)
636    * BP3 (scalar Laplace operator)
637* - `petsc`
638  - * BP1 (scalar mass operator)
639* - `nek5000`
640  - * BP1 (scalar mass operator)
641:::
642
643(v0-1)=
644
645## v0.1 (Jan 3, 2018)
646
647Initial low-level API of the CEED project. The low-level API provides a set of Finite
648Elements kernels and components for writing new low-level kernels. Examples include:
649vector and sparse linear algebra, element matrix assembly over a batch of elements,
650partial assembly and action for efficient high-order operators like mass, diffusion,
651advection, etc. The main goal of the low-level API is to establish the basis for the
652high-level API. Also, identifying such low-level kernels and providing a reference
653implementation for them serves as the basis for specialized backend implementations.
654This release contained several backends: `/cpu/self`, and backends which rely upon the
655[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`,
656`/gpu/occa`, and `/omp/occa`.
657It also included several examples, in the `examples` folder:
658A standalone code that shows the usage of libCEED (with no external
659dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1
660(with the application of the mass operator); and a `petsc` example to perform BP1
661(with the application of the mass operator).
662
663Backends available in this release:
664
665| CEED resource (`-ceed`) | Backend                         |
666|-------------------------|---------------------------------|
667| `/cpu/self`             | Serial reference implementation |
668| `/cpu/occa`             | Serial OCCA kernels             |
669| `/gpu/occa`             | CUDA OCCA kernels               |
670| `/omp/occa`             | OpenMP OCCA kernels             |
671
672Examples available in this release:
673
674| User code             | Example                           |
675|-----------------------|-----------------------------------|
676| `ceed`                | ex1 (scalar Laplace operator)     |
677| `mfem`                | BP1 (scalar mass operator)        |
678| `petsc`               | BP1 (scalar mass operator)        |
679```
680