xref: /libCEED/doc/sphinx/source/releasenotes.md (revision 3e551a327d6c97f9de071b988b42ffdb7bed19a7)
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- Add `bool` field type for `CeedQFunctionContext` and related interfaces to use `bool` fields.
12- `CEED_BASIS_COLLOCATED` removed; users should only use `CEED_BASIS_NONE`.
13- Remove unneeded pointer for `CeedElemRestrictionGetELayout`.
14- Change QFunction source include file handling in JiT compilers
15    - Add `CEED_RUNNING_JIT_PASS` compiler definition for wrapping header files that device JiT compilers cannot read
16    - Users should now prefer `#include <ceed/types.h>` rather than `#include <ceed.h>` in QFunction source files
17- Require use of `Ceed*Destroy()` on Ceed objects returned from `Ceed*Get*()`.
18
19### New features
20
21- Add `CeedOperatorCreateAtPoints` which evaluates the `CeedQFunction` at arbitrary locations in each element, for use in Particle in Cell, Material Point Method, and similar methods.
22- Add `CeedElemRestrictionGetLLayout` to provide L-vector layout for strided `CeedElemRestriction` created with `CEED_BACKEND_STRIDES`.
23- Add `CeedVectorReturnCeed` and similar when parent `Ceed` context for a libCEED object is only needed once in a calling scope.
24- Enable `#pragma once` for all JiT source; remove duplicate includes in JiT source string before compilation.
25- Allow user to set additional compiler options for CUDA and HIP JiT.
26Specifically, directories set with `CeedAddJitSourceRoot(ceed, "foo/bar")` will be used to set `-Ifoo/bar` and defines set with `CeedAddJitDefine(ceed, "foo=bar")` will be used to set `-Dfoo=bar`.
27
28### Examples
29
30- Add deal.II example with CEED BP suite.
31
32(v0-12)=
33
34## v0.12 (Oct 31, 2023)
35
36### Interface changes
37
38- Update `CeedOperatorContext*` functions to `CeedOperator*Context*` functions for consistency.
39For example, `CeedOperatorContextGetFieldLabel` was renamed to `CeedOperatorGetContextFieldLabel`.
40- Removed `CeedBasisSetNumQuadraturePoints` as redundant and bug-prone interface.
41
42### New features
43
44- Added {c:func}`CeedOperatorGetFieldByName` to access a specific `CeedOperatorField` by its name.
45- Update `/cpu/self/memcheck/*` backends to help verify `CeedVector` array access assumptions and `CeedQFunction` user output assumptions.
46- Update {c:func}`CeedOperatorLinearAssembleDiagonal` to provide default implementation that supports `CeedOperator` with multiple active bases.
47- Added Sycl backends `/gpu/sycl/ref`, `/gpu/sycl/shared`, and `/gpu/sycl/gen`.
48- Added {c:func}`CeedBasisApplyAtPoints` for evaluation of values and derivatives at arbitrary points inside elements.
49- Added support for non-tensor $H(\text{curl})$ finite element spaces with {c:func}`CeedBasisCreateHcurl`.
50- 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)).
51- Added {c:func}`CeedOperatorLinearAssemblePointBlockDiagonalSymbolic` to create COO mapping for mapping out of {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`.
52- Added support for application codes which manage multiple {ref}`Ceed` objects, parallelized across OpenMP threads.
53
54### Examples
55
56- Add `DMSwarm` example demonstrating interpolation from background mesh to swarm points and projection from swarm points to background mesh.
57
58#### {ref}`example-petsc-bps`
59
60- Requires PETSc version 3.19 or later.
61
62#### {ref}`example-petsc-navier-stokes`
63
64- Updated restart and checkpointing interface.
65- Add data-driven subgrid-stress model.
66- Add differential filtering of solution.
67- Add turbulence statistics collection over spanwise-symmetric geometries.
68- Add Taylor-Green vortex initial condition.
69- Add Riemann-based outflow boundary conditions.
70- Added vortex shedding and flow past cylinder example, including calculations for lift, drag, and heat transfer.
71- Add Internal Damping Layer (IDL) for helping turbulent simulation stability.
72- Derive `CeedBasis` from `PetscFE`, and various other internal maintainability updates.
73
74(v0-11)=
75
76## v0.11 (Dec 24, 2022)
77
78### Interface changes
79
80- Added {c:func}`CeedOperatorSetName` for more readable {c:func}`CeedOperatorView` output.
81- Added {c:func}`CeedBasisCreateProjection` to facilitate interpolation between nodes for separate `CeedBases`.
82- Rename and move {c:func}`CeedCompositeOperatorGetNumSub` and {c:func}`CeedCompositeOperatorGetSubList` to public interface.
83- Renamed `CEED_BASIS_COLLOCATED` to `CEED_BASIS_NONE` for clarity.
84Some 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.
85
86### New features
87
88- Update `/cpu/self/memcheck/*` backends to help verify `CeedQFunctionContext` data sizes provided by user.
89- Improved support for $H(\text{div})$ bases.
90- Added `CeedInt_FMT` to support potential future use of larger integer sizes.
91- Added `CEED_QFUNCTION_ATTR` for setting compiler attributes/pragmas to `CEED_QFUNCTION_HELPER` and `CEED_QFUNCTION`.
92- OCCA backend updated to latest OCCA release; DPC++ and OMP OCCA modes enabled.
93Due to a limitation of the OCCA parser, typedefs are required to use pointers to arrays in QFunctions with the OCCA backend.
94This issue will be fixed in a future OCCA release.
95
96### Bugfix
97
98- Fix bug in setting device id for GPU backends.
99- Fix storing of indices for `CeedElemRestriction` on the host with GPU backends.
100- Fix `CeedElemRestriction` sizing for {c:func}`CeedOperatorAssemblePointBlockDiagonal`.
101- Fix bugs in CPU implementation of {c:func}`CeedOperatorLinearAssemble` when there are different number of active input modes and active output modes.
102
103### Examples
104
105#### {ref}`example-petsc-navier-stokes`
106
107- Various performance enhancements, analytic matrix-free and assembled Jacobian, and PETSc solver configurations for GPUs.
108- Refactored to improve code reuse and modularity.
109- Support for primitive variables for more accurate boundary layers and all-speed flow.
110- Added $YZ\beta$ shock capturing scheme and Shock Tube example.
111- Added Channel example, with comparison to analytic solutions.
112- Added Flat Plate with boundary layer mesh and compressible Blasius inflow condition based on Chebyshev collocation solution of the Blasius equations.
113- Added strong and weak synthetic turbulence generation (STG) inflow boundary conditions.
114- Added "freestream" boundary conditions based on HLLC Riemann solver.
115- Automated stabilization coefficients for different basis degree.
116
117#### {ref}`example-petsc-bps`
118
119- Support for convergence studies.
120
121### Maintainability
122
123- 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.
124- Enabled support for `p > 8` for `/gpu/*/shared` backends.
125- Switch to `clang-format` over `astyle` for automatic formatting; Makefile command changed to `make format` from `make style`.
126- Improved test harness.
127
128(v0-10-1)=
129
130## v0.10.1 (Apr 11, 2022)
131
132### Interface changes
133
134- Added {c:func}`CeedQFunctionSetUserFlopsEstimate` and {c:func}`CeedOperatorGetFlopsEstimate` to facilitate estimating FLOPs in operator application.
135
136### New features
137
138- Switched MAGMA backends to use runtime compilation for tensor basis kernels (and element restriction kernels, in non-deterministic `/gpu/*/magma` backends).
139This reduces time to compile the library and increases the range of parameters for which the MAGMA tensor basis kernels will work.
140
141### Bugfix
142
143- Install JiT source files in install directory to fix GPU functionality for installed libCEED.
144
145(v0-10)=
146
147## v0.10 (Mar 21, 2022)
148
149### Interface changes
150
151- Update {c:func}`CeedQFunctionGetFields` and {c:func}`CeedOperatorGetFields` to include number of fields.
152- 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`.
153- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added.
154- Add {c:func}`CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients.
155- Promote {c:func}`CeedOperatorCheckReady`to the public API to facilitate interactive interfaces.
156- Warning added when compiling OCCA backend to alert users that this backend is experimental.
157- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`.
158- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name.
159- 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`.
160- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitialized arrays; require initialized data for {c:func}`CeedVectorGetArray`.
161- 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.
162- Added {c:func}`CeedQFunctionContextGetFieldDescriptions` to retrieve user defined descriptions of fields that are registered with `CeedQFunctionContextRegister*`.
163- Renamed `CeedElemTopology` entries for clearer namespacing between libCEED enums.
164- 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.
165- Added {c:func}`CeedOperatorSetQFunctionUpdated` to facilitate QFunction data re-use between operators sharing the same quadrature space, such as in a multigrid hierarchy.
166- Added {c:func}`CeedOperatorGetActiveVectorLengths` to get shape of CeedOperator.
167
168### New features
169
170- `CeedScalar` can now be set as `float` or `double` at compile time.
171- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends.
172- 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.
173- Remove need to guard library headers in QFunction source for code generation backends.
174- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present.
175- Added {c:func}`CeedStringAllocCopy` to reduce repeated code for copying strings internally.
176- Added {c:func}`CeedPathConcatenate` to facilitate loading kernel source files with a path relative to the current file.
177- Added support for non-tensor $H(\text{div})$ elements, to include CPU backend implementations and {c:func}`CeedBasisCreateHdiv` convenience constructor.
178- 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.
179- Added support for element matrix assembly in GPU backends.
180
181### Maintainability
182
183- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage.
184- `Include-what-you-use` makefile target added as `make iwyu`.
185- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase.
186- Put GPU JiTed kernel source code into separate files.
187- 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.
188
189(v0-9)=
190
191## v0.9 (Jul 6, 2021)
192
193### Interface changes
194
195- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact.
196
197### New features
198
199- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use.
200- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends.
201- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations.
202- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed).
203
204### Performance improvements
205
206### Examples
207
208- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations.
209- Fluid mechanics example adds GPU support and improves modularity.
210
211### Deprecated backends
212
213- 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.
214
215(v0-8)=
216
217## v0.8 (Mar 31, 2021)
218
219### Interface changes
220
221- Error handling improved to include enumerated error codes for C interface return values.
222- 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.
223
224### New features
225
226- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features.
227- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly.
228- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators.
229
230### Performance improvements
231
232- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`.
233- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`.
234
235### Examples
236
237- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions.
238- {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.
239- {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.
240- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree.
241
242(v0-7)=
243
244## v0.7 (Sep 29, 2020)
245
246### Interface changes
247
248- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors.
249  As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`.
250  These changes improve support for mixed finite element methods.
251- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature.
252- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity.
253- 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.
254- 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.
255- 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.
256- 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.
257  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`.
258- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory.
259- 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.
260
261### New features
262
263- New HIP backend: `/gpu/hip/ref`.
264- CeedQFunction support for user `CUfunction`s in some backends
265
266### Performance improvements
267
268- OCCA backend rebuilt to facilitate future performance enhancements.
269- PETSc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`.
270
271### Examples
272
273- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions.
274
275### Deprecated backends
276
277- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`.
278
279(v0-6)=
280
281## v0.6 (Mar 29, 2020)
282
283libCEED v0.6 contains numerous new features and examples, as well as expanded
284documentation in [this new website](https://libceed.org).
285
286### New features
287
288- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly
289  1-1 correspondence with the C interface, plus some convenience features.  For instance,
290  data stored in the {cpp:type}`CeedVector` structure are available without copy as
291  {py:class}`numpy.ndarray`.  Short tutorials are provided in
292  [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/).
293- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point,
294  {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal
295  ({c:func}`CeedOperatorAssembleLinearDiagonal`).  These operations are useful for
296  preconditioning ingredients and are used in the libCEED's multigrid examples.
297- The inverse of separable operators can be obtained using
298  {c:func}`CeedOperatorCreateFDMElementInverse` and applied with
299  {c:func}`CeedOperatorApply`.  This is a useful preconditioning ingredient,
300  especially for Laplacians and related operators.
301- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`,
302  {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`.
303- Make public accessors for various attributes to facilitate writing composable code.
304- New backend: `/cpu/self/memcheck/serial`.
305- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA
306  backends.  (Single source is coming soon for OCCA backends.)
307- Fix some missing edge cases in CUDA backend.
308
309### Performance Improvements
310
311- MAGMA backend performance optimization and non-tensor bases.
312- No-copy optimization in {c:func}`CeedOperatorApply`.
313
314### Interface changes
315
316- Replace {code}`CeedElemRestrictionCreateIdentity` and
317  {code}`CeedElemRestrictionCreateBlocked` with more flexible
318  {c:func}`CeedElemRestrictionCreateStrided` and
319  {c:func}`CeedElemRestrictionCreateBlockedStrided`.
320- Add arguments to {c:func}`CeedQFunctionCreateIdentity`.
321- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification
322  with {cpp:enum}`CeedInterlaceMode`.  This is now an attribute of the
323  {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no
324  longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and
325  {c:func}`CeedElemRestrictionApply`.
326
327### Examples
328
329libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`.
330Notable additions include:
331
332- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of
333  a domain in 1, 2, and 3 dimensions by applying a Laplacian.
334
335- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area
336  of domains (like the cube and sphere) by direct integration on a surface mesh;
337  demonstrates geometric dimension different from topological dimension.
338
339- PETSc {ref}`example-petsc-bps`:
340
341  - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support.
342  - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements
343    and transparent CUDA support.
344  - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`):
345    generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric
346    dimension different from topological dimension.
347
348- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid
349  solver with algebraic multigrid coarse solve.
350
351- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly
352  `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`),
353  implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and
354  quasi-2D computational domain support.
355
356- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for
357  linear elasticity, small-strain hyperelasticity, and globalized finite-strain
358  hyperelasticity using p-multigrid with algebraic multigrid coarse solve.
359
360(v0-5)=
361
362## v0.5 (Sep 18, 2019)
363
364For this release, several improvements were made. Two new CUDA backends were added to
365the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art
366performance using single-source {ref}`CeedQFunction`. From this release, users
367can define Q-Functions in a single source code independently of the targeted backend
368with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU
369compilation of the user provided {ref}`CeedQFunction` code. To allow a unified
370declaration, the {ref}`CeedQFunction` API has undergone a slight change:
371the `QFunctionField` parameter `ncomp` has been changed to `size`. This change
372requires setting the previous value of `ncomp` to `ncomp*dim` when adding a
373`QFunctionField` with eval mode `CEED EVAL GRAD`.
374
375Additionally, new CPU backends
376were included in this release, such as the `/cpu/self/opt/*` backends (which are
377written in pure C and use partial **E-vectors** to improve performance) and the
378`/cpu/self/ref/memcheck` backend (which relies upon the
379[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user
380{ref}`CeedQFunction` have no undefined values).
381This release also included various performance improvements, bug fixes, new examples,
382and improved tests. Among these improvements, vectorized instructions for
383{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD`
384instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and
385identity Q-Functions were introduced, and the PETSc benchmark problems were expanded
386to include unstructured meshes handling were. For this expansion, the prior version of
387the PETSc BPs, which only included data associated with structured geometries, were
388renamed `bpsraw`, and the new version of the BPs, which can handle data associated
389with any unstructured geometry, were called `bps`. Additionally, other benchmark
390problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively),
391and BP5 and BP6 (the collocated versions---for which the quadrature points are the same
392as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc
393examples. Furthermoew, another standalone libCEED example, called `ex2`, which
394computes the surface area of a given mesh was added to this release.
395
396Backends available in this release:
397
398| CEED resource (`-ceed`)  | Backend                                             |
399|--------------------------|-----------------------------------------------------|
400| `/cpu/self/ref/serial`   | Serial reference implementation                     |
401| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
402| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks            |
403| `/cpu/self/opt/serial`   | Serial optimized C implementation                   |
404| `/cpu/self/opt/blocked`  | Blocked optimized C implementation                  |
405| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
406| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
407| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
408| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
409| `/cpu/occa`              | Serial OCCA kernels                                 |
410| `/gpu/occa`              | CUDA OCCA kernels                                   |
411| `/omp/occa`              | OpenMP OCCA kernels                                 |
412| `/ocl/occa`              | OpenCL OCCA kernels                                 |
413| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
414| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
415| `/gpu/cuda/shared`       | Optimized pure CUDA kernels using shared memory     |
416| `/gpu/cuda/gen`          | Optimized pure CUDA kernels using code generation   |
417| `/gpu/magma`             | CUDA MAGMA kernels                                  |
418
419Examples available in this release:
420
421:::{list-table}
422:header-rows: 1
423:widths: auto
424* - User code
425  - Example
426* - `ceed`
427  - * ex1 (volume)
428    * ex2 (surface)
429* - `mfem`
430  - * BP1 (scalar mass operator)
431    * BP3 (scalar Laplace operator)
432* - `petsc`
433  - * BP1 (scalar mass operator)
434    * BP2 (vector mass operator)
435    * BP3 (scalar Laplace operator)
436    * BP4 (vector Laplace operator)
437    * BP5 (collocated scalar Laplace operator)
438    * BP6 (collocated vector Laplace operator)
439    * Navier-Stokes
440* - `nek5000`
441  - * BP1 (scalar mass operator)
442    * BP3 (scalar Laplace operator)
443:::
444
445(v0-4)=
446
447## v0.4 (Apr 1, 2019)
448
449libCEED v0.4 was made again publicly available in the second full CEED software
450distribution, release CEED 2.0. This release contained notable features, such as
451four new CPU backends, two new GPU backends, CPU backend optimizations, initial
452support for operator composition, performance benchmarking, and a Navier-Stokes demo.
453The new CPU backends in this release came in two families. The `/cpu/self/*/serial`
454backends process one element at a time and are intended for meshes with a smaller number
455of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of
456eight interlaced elements and are intended for meshes with higher numbers of elements.
457The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU
458performance. The `/cpu/self/xsmm/*` backends rely upon the
459[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU
460performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA.
461The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable
462performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple
463parallelization approach, where each thread treats a finite element. Using just in time
464compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this
465backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend
466achieve good peak performance for 1D, 2D, and low order 3D problems, but performance
467deteriorates very quickly when threads run out of registers.
468
469A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED
470examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`).
471This example solves the time-dependent Navier-Stokes equations of compressible gas
472dynamics in a static Eulerian three-dimensional frame, using structured high-order
473finite/spectral element spatial discretizations and explicit high-order time-stepping
474(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc,
475so that the pointwise physics (defined at quadrature points) is separated from the
476parallelization and meshing concerns.
477
478Backends available in this release:
479
480| CEED resource (`-ceed`)  | Backend                                             |
481|--------------------------|-----------------------------------------------------|
482| `/cpu/self/ref/serial`   | Serial reference implementation                     |
483| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
484| `/cpu/self/tmpl`         | Backend template, defaults to `/cpu/self/blocked`   |
485| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
486| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
487| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
488| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
489| `/cpu/occa`              | Serial OCCA kernels                                 |
490| `/gpu/occa`              | CUDA OCCA kernels                                   |
491| `/omp/occa`              | OpenMP OCCA kernels                                 |
492| `/ocl/occa`              | OpenCL OCCA kernels                                 |
493| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
494| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
495| `/gpu/magma`             | CUDA MAGMA kernels                                  |
496
497Examples available in this release:
498
499:::{list-table}
500:header-rows: 1
501:widths: auto
502* - User code
503  - Example
504* - `ceed`
505  - * ex1 (volume)
506* - `mfem`
507  - * BP1 (scalar mass operator)
508    * BP3 (scalar Laplace operator)
509* - `petsc`
510  - * BP1 (scalar mass operator)
511    * BP3 (scalar Laplace operator)
512    * Navier-Stokes
513* - `nek5000`
514  - * BP1 (scalar mass operator)
515    * BP3 (scalar Laplace operator)
516:::
517
518(v0-3)=
519
520## v0.3 (Sep 30, 2018)
521
522Notable features in this release include active/passive field interface, support for
523non-tensor bases, backend optimization, and improved Fortran interface. This release
524also focused on providing improved continuous integration, and many new tests with code
525coverage reports of about 90%. This release also provided a significant change to the
526public interface: a {ref}`CeedQFunction` can take any number of named input and output
527arguments while {ref}`CeedOperator` connects them to the actual data, which may be
528supplied explicitly to `CeedOperatorApply()` (active) or separately via
529`CeedOperatorSetField()` (passive). This interface change enables reusable libraries
530of CeedQFunctions and composition of block solvers constructed using
531{ref}`CeedOperator`. A concept of blocked restriction was added to this release and
532used in an optimized CPU backend. Although this is typically not visible to the user,
533it enables effective use of arbitrary-length SIMD while maintaining cache locality.
534This CPU backend also implements an algebraic factorization of tensor product gradients
535to perform fewer operations than standard application of interpolation and
536differentiation from nodes to quadrature points. This algebraic formulation
537automatically supports non-polynomial and non-interpolatory bases, thus is more general
538than the more common derivation in terms of Lagrange polynomials on the quadrature points.
539
540Backends available in this release:
541
542| CEED resource (`-ceed`) | Backend                                             |
543|-------------------------|-----------------------------------------------------|
544| `/cpu/self/blocked`     | Blocked reference implementation                    |
545| `/cpu/self/ref`         | Serial reference implementation                     |
546| `/cpu/self/tmpl`        | Backend template, defaults to `/cpu/self/blocked`   |
547| `/cpu/occa`             | Serial OCCA kernels                                 |
548| `/gpu/occa`             | CUDA OCCA kernels                                   |
549| `/omp/occa`             | OpenMP OCCA kernels                                 |
550| `/ocl/occa`             | OpenCL OCCA kernels                                 |
551| `/gpu/magma`            | CUDA MAGMA kernels                                  |
552
553Examples available in this release:
554
555:::{list-table}
556:header-rows: 1
557:widths: auto
558* - User code
559  - Example
560* - `ceed`
561  - * ex1 (volume)
562* - `mfem`
563  - * BP1 (scalar mass operator)
564    * BP3 (scalar Laplace operator)
565* - `petsc`
566  - * BP1 (scalar mass operator)
567    * BP3 (scalar Laplace operator)
568* - `nek5000`
569  - * BP1 (scalar mass operator)
570    * BP3 (scalar Laplace operator)
571:::
572
573(v0-21)=
574
575## v0.21 (Sep 30, 2018)
576
577A MAGMA backend (which relies upon the
578[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this
579release. This initial integration set up the framework of using MAGMA and provided the
580libCEED functionality through MAGMA kernels as one of libCEED’s computational backends.
581As any other backend, the MAGMA backend provides extended basic data structures for
582{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements
583the fundamental CEED building blocks to work with the new data structures.
584In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data
585but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled
586internally, and thus seamlessly to the user, through the functions/methods that are
587provided to support them.
588
589Backends available in this release:
590
591| CEED resource (`-ceed`) | Backend                         |
592|-------------------------|---------------------------------|
593| `/cpu/self`             | Serial reference implementation |
594| `/cpu/occa`             | Serial OCCA kernels             |
595| `/gpu/occa`             | CUDA OCCA kernels               |
596| `/omp/occa`             | OpenMP OCCA kernels             |
597| `/ocl/occa`             | OpenCL OCCA kernels             |
598| `/gpu/magma`            | CUDA MAGMA kernels              |
599
600Examples available in this release:
601
602:::{list-table}
603:header-rows: 1
604:widths: auto
605* - User code
606  - Example
607* - `ceed`
608  - * ex1 (volume)
609* - `mfem`
610  - * BP1 (scalar mass operator)
611    * BP3 (scalar Laplace operator)
612* - `petsc`
613  - * BP1 (scalar mass operator)
614* - `nek5000`
615  - * BP1 (scalar mass operator)
616:::
617
618(v0-2)=
619
620## v0.2 (Mar 30, 2018)
621
622libCEED was made publicly available the first full CEED software distribution, release
623CEED 1.0. The distribution was made available using the Spack package manager to provide
624a common, easy-to-use build environment, where the user can build the CEED distribution
625with all dependencies. This release included a new Fortran interface for the library.
626This release also contained major improvements in the OCCA backend (including a new
627`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to
628compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an
629`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application
630of the Laplace operator) was also added to this release.
631
632Backends available in this release:
633
634| CEED resource (`-ceed`) | Backend                         |
635|-------------------------|---------------------------------|
636| `/cpu/self`             | Serial reference implementation |
637| `/cpu/occa`             | Serial OCCA kernels             |
638| `/gpu/occa`             | CUDA OCCA kernels               |
639| `/omp/occa`             | OpenMP OCCA kernels             |
640| `/ocl/occa`             | OpenCL OCCA kernels             |
641
642Examples available in this release:
643
644:::{list-table}
645:header-rows: 1
646:widths: auto
647* - User code
648  - Example
649* - `ceed`
650  - * ex1 (volume)
651* - `mfem`
652  - * BP1 (scalar mass operator)
653    * BP3 (scalar Laplace operator)
654* - `petsc`
655  - * BP1 (scalar mass operator)
656* - `nek5000`
657  - * BP1 (scalar mass operator)
658:::
659
660(v0-1)=
661
662## v0.1 (Jan 3, 2018)
663
664Initial low-level API of the CEED project. The low-level API provides a set of Finite
665Elements kernels and components for writing new low-level kernels. Examples include:
666vector and sparse linear algebra, element matrix assembly over a batch of elements,
667partial assembly and action for efficient high-order operators like mass, diffusion,
668advection, etc. The main goal of the low-level API is to establish the basis for the
669high-level API. Also, identifying such low-level kernels and providing a reference
670implementation for them serves as the basis for specialized backend implementations.
671This release contained several backends: `/cpu/self`, and backends which rely upon the
672[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`,
673`/gpu/occa`, and `/omp/occa`.
674It also included several examples, in the `examples` folder:
675A standalone code that shows the usage of libCEED (with no external
676dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1
677(with the application of the mass operator); and a `petsc` example to perform BP1
678(with the application of the mass operator).
679
680Backends available in this release:
681
682| CEED resource (`-ceed`) | Backend                         |
683|-------------------------|---------------------------------|
684| `/cpu/self`             | Serial reference implementation |
685| `/cpu/occa`             | Serial OCCA kernels             |
686| `/gpu/occa`             | CUDA OCCA kernels               |
687| `/omp/occa`             | OpenMP OCCA kernels             |
688
689Examples available in this release:
690
691| User code             | Example                           |
692|-----------------------|-----------------------------------|
693| `ceed`                | ex1 (scalar Laplace operator)     |
694| `mfem`                | BP1 (scalar mass operator)        |
695| `petsc`               | BP1 (scalar mass operator)        |
696```
697