xref: /libCEED/doc/sphinx/source/releasenotes.md (revision 8b7d3340714d052d4785a048802399d76a15146b)
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- Added non-tensor basis support to code generation backends `/gpu/cuda/gen` and `/gpu/hip/gen`.
28- Added support to code generation backends `/gpu/cuda/gen` and `/gpu/hip/gen` for operators with both tensor and non-tensor bases.
29- Add `CeedGetGitVersion()` to access the Git commit and dirty state of the repository at build time.
30- Add `CeedGetBuildConfiguration()` to access compilers, flags, and related information about the build environment.
31
32### Examples
33
34- Add deal.II example with CEED BP suite.
35
36(v0-12)=
37
38## v0.12 (Oct 31, 2023)
39
40### Interface changes
41
42- Update `CeedOperatorContext*` functions to `CeedOperator*Context*` functions for consistency.
43For example, `CeedOperatorContextGetFieldLabel` was renamed to `CeedOperatorGetContextFieldLabel`.
44- Removed `CeedBasisSetNumQuadraturePoints` as redundant and bug-prone interface.
45
46### New features
47
48- Added {c:func}`CeedOperatorGetFieldByName` to access a specific `CeedOperatorField` by its name.
49- Update `/cpu/self/memcheck/*` backends to help verify `CeedVector` array access assumptions and `CeedQFunction` user output assumptions.
50- Update {c:func}`CeedOperatorLinearAssembleDiagonal` to provide default implementation that supports `CeedOperator` with multiple active bases.
51- Added Sycl backends `/gpu/sycl/ref`, `/gpu/sycl/shared`, and `/gpu/sycl/gen`.
52- Added {c:func}`CeedBasisApplyAtPoints` for evaluation of values and derivatives at arbitrary points inside elements.
53- Added support for non-tensor $H(\text{curl})$ finite element spaces with {c:func}`CeedBasisCreateHcurl`.
54- 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)).
55- Added {c:func}`CeedOperatorLinearAssemblePointBlockDiagonalSymbolic` to create COO mapping for mapping out of {c:func}`CeedOperatorLinearAssemblePointBlockDiagonal`.
56- Added support for application codes which manage multiple {ref}`Ceed` objects, parallelized across OpenMP threads.
57
58### Examples
59
60- Add `DMSwarm` example demonstrating interpolation from background mesh to swarm points and projection from swarm points to background mesh.
61
62#### {ref}`example-petsc-bps`
63
64- Requires PETSc version 3.19 or later.
65
66#### {ref}`example-petsc-navier-stokes`
67
68- Updated restart and checkpointing interface.
69- Add data-driven subgrid-stress model.
70- Add differential filtering of solution.
71- Add turbulence statistics collection over spanwise-symmetric geometries.
72- Add Taylor-Green vortex initial condition.
73- Add Riemann-based outflow boundary conditions.
74- Added vortex shedding and flow past cylinder example, including calculations for lift, drag, and heat transfer.
75- Add Internal Damping Layer (IDL) for helping turbulent simulation stability.
76- Derive `CeedBasis` from `PetscFE`, and various other internal maintainability updates.
77
78(v0-11)=
79
80## v0.11 (Dec 24, 2022)
81
82### Interface changes
83
84- Added {c:func}`CeedOperatorSetName` for more readable {c:func}`CeedOperatorView` output.
85- Added {c:func}`CeedBasisCreateProjection` to facilitate interpolation between nodes for separate `CeedBases`.
86- Rename and move {c:func}`CeedCompositeOperatorGetNumSub` and {c:func}`CeedCompositeOperatorGetSubList` to public interface.
87- Renamed `CEED_BASIS_COLLOCATED` to `CEED_BASIS_NONE` for clarity.
88Some 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.
89
90### New features
91
92- Update `/cpu/self/memcheck/*` backends to help verify `CeedQFunctionContext` data sizes provided by user.
93- Improved support for $H(\text{div})$ bases.
94- Added `CeedInt_FMT` to support potential future use of larger integer sizes.
95- Added `CEED_QFUNCTION_ATTR` for setting compiler attributes/pragmas to `CEED_QFUNCTION_HELPER` and `CEED_QFUNCTION`.
96- OCCA backend updated to latest OCCA release; DPC++ and OMP OCCA modes enabled.
97Due to a limitation of the OCCA parser, typedefs are required to use pointers to arrays in QFunctions with the OCCA backend.
98This issue will be fixed in a future OCCA release.
99
100### Bugfix
101
102- Fix bug in setting device id for GPU backends.
103- Fix storing of indices for `CeedElemRestriction` on the host with GPU backends.
104- Fix `CeedElemRestriction` sizing for {c:func}`CeedOperatorAssemblePointBlockDiagonal`.
105- Fix bugs in CPU implementation of {c:func}`CeedOperatorLinearAssemble` when there are different number of active input modes and active output modes.
106
107### Examples
108
109#### {ref}`example-petsc-navier-stokes`
110
111- Various performance enhancements, analytic matrix-free and assembled Jacobian, and PETSc solver configurations for GPUs.
112- Refactored to improve code reuse and modularity.
113- Support for primitive variables for more accurate boundary layers and all-speed flow.
114- Added $YZ\beta$ shock capturing scheme and Shock Tube example.
115- Added Channel example, with comparison to analytic solutions.
116- Added Flat Plate with boundary layer mesh and compressible Blasius inflow condition based on Chebyshev collocation solution of the Blasius equations.
117- Added strong and weak synthetic turbulence generation (STG) inflow boundary conditions.
118- Added "freestream" boundary conditions based on HLLC Riemann solver.
119- Automated stabilization coefficients for different basis degree.
120
121#### {ref}`example-petsc-bps`
122
123- Support for convergence studies.
124
125### Maintainability
126
127- 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.
128- Enabled support for `p > 8` for `/gpu/*/shared` backends.
129- Switch to `clang-format` over `astyle` for automatic formatting; Makefile command changed to `make format` from `make style`.
130- Improved test harness.
131
132(v0-10-1)=
133
134## v0.10.1 (Apr 11, 2022)
135
136### Interface changes
137
138- Added {c:func}`CeedQFunctionSetUserFlopsEstimate` and {c:func}`CeedOperatorGetFlopsEstimate` to facilitate estimating FLOPs in operator application.
139
140### New features
141
142- Switched MAGMA backends to use runtime compilation for tensor basis kernels (and element restriction kernels, in non-deterministic `/gpu/*/magma` backends).
143This reduces time to compile the library and increases the range of parameters for which the MAGMA tensor basis kernels will work.
144
145### Bugfix
146
147- Install JiT source files in install directory to fix GPU functionality for installed libCEED.
148
149(v0-10)=
150
151## v0.10 (Mar 21, 2022)
152
153### Interface changes
154
155- Update {c:func}`CeedQFunctionGetFields` and {c:func}`CeedOperatorGetFields` to include number of fields.
156- 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`.
157- Clarify and document conditions where `CeedQFunction` and `CeedOperator` become immutable and no further fields or suboperators can be added.
158- Add {c:func}`CeedOperatorLinearAssembleQFunctionBuildOrUpdate` to reduce object creation overhead in assembly of CeedOperator preconditioning ingredients.
159- Promote {c:func}`CeedOperatorCheckReady`to the public API to facilitate interactive interfaces.
160- Warning added when compiling OCCA backend to alert users that this backend is experimental.
161- `ceed-backend.h`, `ceed-hash.h`, and `ceed-khash.h` removed. Users should use `ceed/backend.h`, `ceed/hash.h`, and `ceed/khash.h`.
162- Added {c:func}`CeedQFunctionGetKernelName`; refactored {c:func}`CeedQFunctionGetSourcePath` to exclude function kernel name.
163- 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`.
164- Added {c:func}`CeedVectorGetArrayWrite` that allows access to uninitialized arrays; require initialized data for {c:func}`CeedVectorGetArray`.
165- 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.
166- Added {c:func}`CeedQFunctionContextGetFieldDescriptions` to retrieve user defined descriptions of fields that are registered with `CeedQFunctionContextRegister*`.
167- Renamed `CeedElemTopology` entries for clearer namespacing between libCEED enums.
168- 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.
169- Added {c:func}`CeedOperatorSetQFunctionUpdated` to facilitate QFunction data re-use between operators sharing the same quadrature space, such as in a multigrid hierarchy.
170- Added {c:func}`CeedOperatorGetActiveVectorLengths` to get shape of CeedOperator.
171
172### New features
173
174- `CeedScalar` can now be set as `float` or `double` at compile time.
175- Added JiT utilities in `ceed/jit-tools.h` to reduce duplicated code in GPU backends.
176- 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.
177- Remove need to guard library headers in QFunction source for code generation backends.
178- `CeedDebugEnv()` macro created to provide debugging outputs when Ceed context is not present.
179- Added {c:func}`CeedStringAllocCopy` to reduce repeated code for copying strings internally.
180- Added {c:func}`CeedPathConcatenate` to facilitate loading kernel source files with a path relative to the current file.
181- Added support for non-tensor $H(\text{div})$ elements, to include CPU backend implementations and {c:func}`CeedBasisCreateHdiv` convenience constructor.
182- 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.
183- Added support for element matrix assembly in GPU backends.
184
185### Maintainability
186
187- Refactored preconditioner support internally to facilitate future development and improve GPU completeness/test coverage.
188- `Include-what-you-use` makefile target added as `make iwyu`.
189- Create backend constant `CEED_FIELD_MAX` to reduce magic numbers in codebase.
190- Put GPU JiTed kernel source code into separate files.
191- 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.
192
193(v0-9)=
194
195## v0.9 (Jul 6, 2021)
196
197### Interface changes
198
199- Minor modification in error handling macro to silence pedantic warnings when compiling with Clang, but no functional impact.
200
201### New features
202
203- Add {c:func}`CeedVectorAXPY` and {c:func}`CeedVectorPointwiseMult` as a convenience for stand-alone testing and internal use.
204- Add `CEED_QFUNCTION_HELPER` macro to properly annotate QFunction helper functions for code generation backends.
205- Add `CeedPragmaOptimizeOff` macro for code that is sensitive to floating point errors from fast math optimizations.
206- Rust support: split `libceed-sys` crate out of `libceed` and [publish both on crates.io](https://crates.io/crates/libceed).
207
208### Performance improvements
209
210### Examples
211
212- Solid mechanics mini-app updated to explore the performance impacts of various formulations in the initial and current configurations.
213- Fluid mechanics example adds GPU support and improves modularity.
214
215### Deprecated backends
216
217- 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.
218
219(v0-8)=
220
221## v0.8 (Mar 31, 2021)
222
223### Interface changes
224
225- Error handling improved to include enumerated error codes for C interface return values.
226- 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.
227
228### New features
229
230- Julia and Rust interfaces added, providing a nearly 1-1 correspondence with the C interface, plus some convenience features.
231- Static libraries can be built with `make STATIC=1` and the pkg-config file is installed accordingly.
232- Add {c:func}`CeedOperatorLinearAssembleSymbolic` and {c:func}`CeedOperatorLinearAssemble` to support full assembly of libCEED operators.
233
234### Performance improvements
235
236- New HIP MAGMA backends for hipMAGMA library users: `/gpu/hip/magma` and `/gpu/hip/magma/det`.
237- New HIP backends for improved tensor basis performance: `/gpu/hip/shared` and `/gpu/hip/gen`.
238
239### Examples
240
241- {ref}`example-petsc-elasticity` example updated with traction boundary conditions and improved Dirichlet boundary conditions.
242- {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.
243- {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.
244- {ref}`example-petsc-navier-stokes` example updated with support for performing convergence study and plotting order of convergence by polynomial degree.
245
246(v0-7)=
247
248## v0.7 (Sep 29, 2020)
249
250### Interface changes
251
252- Replace limited {code}`CeedInterlaceMode` with more flexible component stride {code}`compstride` in {code}`CeedElemRestriction` constructors.
253  As a result, the {code}`indices` parameter has been replaced with {code}`offsets` and the {code}`nnodes` parameter has been replaced with {code}`lsize`.
254  These changes improve support for mixed finite element methods.
255- Replace various uses of {code}`Ceed*Get*Status` with {code}`Ceed*Is*` in the backend API to match common nomenclature.
256- Replace {code}`CeedOperatorAssembleLinearDiagonal` with {c:func}`CeedOperatorLinearAssembleDiagonal` for clarity.
257- 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.
258- 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.
259- 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.
260- 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.
261  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`.
262- Added {code}`CeedQFunctionContext` object to manage user QFunction context data and reduce copies between device and host memory.
263- 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.
264
265### New features
266
267- New HIP backend: `/gpu/hip/ref`.
268- CeedQFunction support for user `CUfunction`s in some backends
269
270### Performance improvements
271
272- OCCA backend rebuilt to facilitate future performance enhancements.
273- PETSc BPs suite improved to reduce noise due to multiple calls to {code}`mpiexec`.
274
275### Examples
276
277- {ref}`example-petsc-elasticity` example updated with strain energy computation and more flexible boundary conditions.
278
279### Deprecated backends
280
281- The `/gpu/cuda/reg` backend has been removed, with its core features moved into `/gpu/cuda/ref` and `/gpu/cuda/shared`.
282
283(v0-6)=
284
285## v0.6 (Mar 29, 2020)
286
287libCEED v0.6 contains numerous new features and examples, as well as expanded
288documentation in [this new website](https://libceed.org).
289
290### New features
291
292- New Python interface using [CFFI](https://cffi.readthedocs.io/) provides a nearly
293  1-1 correspondence with the C interface, plus some convenience features.  For instance,
294  data stored in the {cpp:type}`CeedVector` structure are available without copy as
295  {py:class}`numpy.ndarray`.  Short tutorials are provided in
296  [Binder](https://mybinder.org/v2/gh/CEED/libCEED/main?urlpath=lab/tree/examples/tutorials/).
297- Linear QFunctions can be assembled as block-diagonal matrices (per quadrature point,
298  {c:func}`CeedOperatorAssembleLinearQFunction`) or to evaluate the diagonal
299  ({c:func}`CeedOperatorAssembleLinearDiagonal`).  These operations are useful for
300  preconditioning ingredients and are used in the libCEED's multigrid examples.
301- The inverse of separable operators can be obtained using
302  {c:func}`CeedOperatorCreateFDMElementInverse` and applied with
303  {c:func}`CeedOperatorApply`.  This is a useful preconditioning ingredient,
304  especially for Laplacians and related operators.
305- New functions: {c:func}`CeedVectorNorm`, {c:func}`CeedOperatorApplyAdd`,
306  {c:func}`CeedQFunctionView`, {c:func}`CeedOperatorView`.
307- Make public accessors for various attributes to facilitate writing composable code.
308- New backend: `/cpu/self/memcheck/serial`.
309- QFunctions using variable-length array (VLA) pointer constructs can be used with CUDA
310  backends.  (Single source is coming soon for OCCA backends.)
311- Fix some missing edge cases in CUDA backend.
312
313### Performance Improvements
314
315- MAGMA backend performance optimization and non-tensor bases.
316- No-copy optimization in {c:func}`CeedOperatorApply`.
317
318### Interface changes
319
320- Replace {code}`CeedElemRestrictionCreateIdentity` and
321  {code}`CeedElemRestrictionCreateBlocked` with more flexible
322  {c:func}`CeedElemRestrictionCreateStrided` and
323  {c:func}`CeedElemRestrictionCreateBlockedStrided`.
324- Add arguments to {c:func}`CeedQFunctionCreateIdentity`.
325- Replace ambiguous uses of {cpp:enum}`CeedTransposeMode` for L-vector identification
326  with {cpp:enum}`CeedInterlaceMode`.  This is now an attribute of the
327  {cpp:type}`CeedElemRestriction` (see {c:func}`CeedElemRestrictionCreate`) and no
328  longer passed as `lmode` arguments to {c:func}`CeedOperatorSetField` and
329  {c:func}`CeedElemRestrictionApply`.
330
331### Examples
332
333libCEED-0.6 contains greatly expanded examples with {ref}`new documentation <Examples>`.
334Notable additions include:
335
336- Standalone {ref}`ex2-surface` ({file}`examples/ceed/ex2-surface`): compute the area of
337  a domain in 1, 2, and 3 dimensions by applying a Laplacian.
338
339- PETSc {ref}`example-petsc-area` ({file}`examples/petsc/area.c`): computes surface area
340  of domains (like the cube and sphere) by direct integration on a surface mesh;
341  demonstrates geometric dimension different from topological dimension.
342
343- PETSc {ref}`example-petsc-bps`:
344
345  - {file}`examples/petsc/bpsraw.c` (formerly `bps.c`): transparent CUDA support.
346  - {file}`examples/petsc/bps.c` (formerly `bpsdmplex.c`): performance improvements
347    and transparent CUDA support.
348  - {ref}`example-petsc-bps-sphere` ({file}`examples/petsc/bpssphere.c`):
349    generalizations of all CEED BPs to the surface of the sphere; demonstrates geometric
350    dimension different from topological dimension.
351
352- {ref}`example-petsc-multigrid` ({file}`examples/petsc/multigrid.c`): new p-multigrid
353  solver with algebraic multigrid coarse solve.
354
355- {ref}`example-petsc-navier-stokes` ({file}`examples/fluids/navierstokes.c`; formerly
356  `examples/navier-stokes`): unstructured grid support (using PETSc's `DMPlex`),
357  implicit time integration, SU/SUPG stabilization, free-slip boundary conditions, and
358  quasi-2D computational domain support.
359
360- {ref}`example-petsc-elasticity` ({file}`examples/solids/elasticity.c`): new solver for
361  linear elasticity, small-strain hyperelasticity, and globalized finite-strain
362  hyperelasticity using p-multigrid with algebraic multigrid coarse solve.
363
364(v0-5)=
365
366## v0.5 (Sep 18, 2019)
367
368For this release, several improvements were made. Two new CUDA backends were added to
369the family of backends, of which, the new `cuda-gen` backend achieves state-of-the-art
370performance using single-source {ref}`CeedQFunction`. From this release, users
371can define Q-Functions in a single source code independently of the targeted backend
372with the aid of a new macro `CEED QFUNCTION` to support JIT (Just-In-Time) and CPU
373compilation of the user provided {ref}`CeedQFunction` code. To allow a unified
374declaration, the {ref}`CeedQFunction` API has undergone a slight change:
375the `QFunctionField` parameter `ncomp` has been changed to `size`. This change
376requires setting the previous value of `ncomp` to `ncomp*dim` when adding a
377`QFunctionField` with eval mode `CEED EVAL GRAD`.
378
379Additionally, new CPU backends
380were included in this release, such as the `/cpu/self/opt/*` backends (which are
381written in pure C and use partial **E-vectors** to improve performance) and the
382`/cpu/self/ref/memcheck` backend (which relies upon the
383[Valgrind](http://valgrind.org/) Memcheck tool to help verify that user
384{ref}`CeedQFunction` have no undefined values).
385This release also included various performance improvements, bug fixes, new examples,
386and improved tests. Among these improvements, vectorized instructions for
387{ref}`CeedQFunction` code compiled for CPU were enhanced by using `CeedPragmaSIMD`
388instead of `CeedPragmaOMP`, implementation of a {ref}`CeedQFunction` gallery and
389identity Q-Functions were introduced, and the PETSc benchmark problems were expanded
390to include unstructured meshes handling were. For this expansion, the prior version of
391the PETSc BPs, which only included data associated with structured geometries, were
392renamed `bpsraw`, and the new version of the BPs, which can handle data associated
393with any unstructured geometry, were called `bps`. Additionally, other benchmark
394problems, namely BP2 and BP4 (the vector-valued versions of BP1 and BP3, respectively),
395and BP5 and BP6 (the collocated versions---for which the quadrature points are the same
396as the Gauss Lobatto nodes---of BP3 and BP4 respectively) were added to the PETSc
397examples. Furthermoew, another standalone libCEED example, called `ex2`, which
398computes the surface area of a given mesh was added to this release.
399
400Backends available in this release:
401
402| CEED resource (`-ceed`)  | Backend                                             |
403|--------------------------|-----------------------------------------------------|
404| `/cpu/self/ref/serial`   | Serial reference implementation                     |
405| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
406| `/cpu/self/ref/memcheck` | Memcheck backend, undefined value checks            |
407| `/cpu/self/opt/serial`   | Serial optimized C implementation                   |
408| `/cpu/self/opt/blocked`  | Blocked optimized C implementation                  |
409| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
410| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
411| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
412| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
413| `/cpu/occa`              | Serial OCCA kernels                                 |
414| `/gpu/occa`              | CUDA OCCA kernels                                   |
415| `/omp/occa`              | OpenMP OCCA kernels                                 |
416| `/ocl/occa`              | OpenCL OCCA kernels                                 |
417| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
418| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
419| `/gpu/cuda/shared`       | Optimized pure CUDA kernels using shared memory     |
420| `/gpu/cuda/gen`          | Optimized pure CUDA kernels using code generation   |
421| `/gpu/magma`             | CUDA MAGMA kernels                                  |
422
423Examples available in this release:
424
425:::{list-table}
426:header-rows: 1
427:widths: auto
428* - User code
429  - Example
430* - `ceed`
431  - * ex1 (volume)
432    * ex2 (surface)
433* - `mfem`
434  - * BP1 (scalar mass operator)
435    * BP3 (scalar Laplace operator)
436* - `petsc`
437  - * BP1 (scalar mass operator)
438    * BP2 (vector mass operator)
439    * BP3 (scalar Laplace operator)
440    * BP4 (vector Laplace operator)
441    * BP5 (collocated scalar Laplace operator)
442    * BP6 (collocated vector Laplace operator)
443    * Navier-Stokes
444* - `nek5000`
445  - * BP1 (scalar mass operator)
446    * BP3 (scalar Laplace operator)
447:::
448
449(v0-4)=
450
451## v0.4 (Apr 1, 2019)
452
453libCEED v0.4 was made again publicly available in the second full CEED software
454distribution, release CEED 2.0. This release contained notable features, such as
455four new CPU backends, two new GPU backends, CPU backend optimizations, initial
456support for operator composition, performance benchmarking, and a Navier-Stokes demo.
457The new CPU backends in this release came in two families. The `/cpu/self/*/serial`
458backends process one element at a time and are intended for meshes with a smaller number
459of high order elements. The `/cpu/self/*/blocked` backends process blocked batches of
460eight interlaced elements and are intended for meshes with higher numbers of elements.
461The `/cpu/self/avx/*` backends rely upon AVX instructions to provide vectorized CPU
462performance. The `/cpu/self/xsmm/*` backends rely upon the
463[LIBXSMM](http://github.com/hfp/libxsmm) package to provide vectorized CPU
464performance. The `/gpu/cuda/*` backends provide GPU performance strictly using CUDA.
465The `/gpu/cuda/ref` backend is a reference CUDA backend, providing reasonable
466performance for most problem configurations. The `/gpu/cuda/reg` backend uses a simple
467parallelization approach, where each thread treats a finite element. Using just in time
468compilation, provided by nvrtc (NVidia Runtime Compiler), and runtime parameters, this
469backend unroll loops and map memory address to registers. The `/gpu/cuda/reg` backend
470achieve good peak performance for 1D, 2D, and low order 3D problems, but performance
471deteriorates very quickly when threads run out of registers.
472
473A new explicit time-stepping Navier-Stokes solver was added to the family of libCEED
474examples in the `examples/petsc` directory (see {ref}`example-petsc-navier-stokes`).
475This example solves the time-dependent Navier-Stokes equations of compressible gas
476dynamics in a static Eulerian three-dimensional frame, using structured high-order
477finite/spectral element spatial discretizations and explicit high-order time-stepping
478(available in PETSc). Moreover, the Navier-Stokes example was developed using PETSc,
479so that the pointwise physics (defined at quadrature points) is separated from the
480parallelization and meshing concerns.
481
482Backends available in this release:
483
484| CEED resource (`-ceed`)  | Backend                                             |
485|--------------------------|-----------------------------------------------------|
486| `/cpu/self/ref/serial`   | Serial reference implementation                     |
487| `/cpu/self/ref/blocked`  | Blocked reference implementation                    |
488| `/cpu/self/tmpl`         | Backend template, defaults to `/cpu/self/blocked`   |
489| `/cpu/self/avx/serial`   | Serial AVX implementation                           |
490| `/cpu/self/avx/blocked`  | Blocked AVX implementation                          |
491| `/cpu/self/xsmm/serial`  | Serial LIBXSMM implementation                       |
492| `/cpu/self/xsmm/blocked` | Blocked LIBXSMM implementation                      |
493| `/cpu/occa`              | Serial OCCA kernels                                 |
494| `/gpu/occa`              | CUDA OCCA kernels                                   |
495| `/omp/occa`              | OpenMP OCCA kernels                                 |
496| `/ocl/occa`              | OpenCL OCCA kernels                                 |
497| `/gpu/cuda/ref`          | Reference pure CUDA kernels                         |
498| `/gpu/cuda/reg`          | Pure CUDA kernels using one thread per element      |
499| `/gpu/magma`             | CUDA MAGMA kernels                                  |
500
501Examples available in this release:
502
503:::{list-table}
504:header-rows: 1
505:widths: auto
506* - User code
507  - Example
508* - `ceed`
509  - * ex1 (volume)
510* - `mfem`
511  - * BP1 (scalar mass operator)
512    * BP3 (scalar Laplace operator)
513* - `petsc`
514  - * BP1 (scalar mass operator)
515    * BP3 (scalar Laplace operator)
516    * Navier-Stokes
517* - `nek5000`
518  - * BP1 (scalar mass operator)
519    * BP3 (scalar Laplace operator)
520:::
521
522(v0-3)=
523
524## v0.3 (Sep 30, 2018)
525
526Notable features in this release include active/passive field interface, support for
527non-tensor bases, backend optimization, and improved Fortran interface. This release
528also focused on providing improved continuous integration, and many new tests with code
529coverage reports of about 90%. This release also provided a significant change to the
530public interface: a {ref}`CeedQFunction` can take any number of named input and output
531arguments while {ref}`CeedOperator` connects them to the actual data, which may be
532supplied explicitly to `CeedOperatorApply()` (active) or separately via
533`CeedOperatorSetField()` (passive). This interface change enables reusable libraries
534of CeedQFunctions and composition of block solvers constructed using
535{ref}`CeedOperator`. A concept of blocked restriction was added to this release and
536used in an optimized CPU backend. Although this is typically not visible to the user,
537it enables effective use of arbitrary-length SIMD while maintaining cache locality.
538This CPU backend also implements an algebraic factorization of tensor product gradients
539to perform fewer operations than standard application of interpolation and
540differentiation from nodes to quadrature points. This algebraic formulation
541automatically supports non-polynomial and non-interpolatory bases, thus is more general
542than the more common derivation in terms of Lagrange polynomials on the quadrature points.
543
544Backends available in this release:
545
546| CEED resource (`-ceed`) | Backend                                             |
547|-------------------------|-----------------------------------------------------|
548| `/cpu/self/blocked`     | Blocked reference implementation                    |
549| `/cpu/self/ref`         | Serial reference implementation                     |
550| `/cpu/self/tmpl`        | Backend template, defaults to `/cpu/self/blocked`   |
551| `/cpu/occa`             | Serial OCCA kernels                                 |
552| `/gpu/occa`             | CUDA OCCA kernels                                   |
553| `/omp/occa`             | OpenMP OCCA kernels                                 |
554| `/ocl/occa`             | OpenCL OCCA kernels                                 |
555| `/gpu/magma`            | CUDA MAGMA kernels                                  |
556
557Examples available in this release:
558
559:::{list-table}
560:header-rows: 1
561:widths: auto
562* - User code
563  - Example
564* - `ceed`
565  - * ex1 (volume)
566* - `mfem`
567  - * BP1 (scalar mass operator)
568    * BP3 (scalar Laplace operator)
569* - `petsc`
570  - * BP1 (scalar mass operator)
571    * BP3 (scalar Laplace operator)
572* - `nek5000`
573  - * BP1 (scalar mass operator)
574    * BP3 (scalar Laplace operator)
575:::
576
577(v0-21)=
578
579## v0.21 (Sep 30, 2018)
580
581A MAGMA backend (which relies upon the
582[MAGMA](https://bitbucket.org/icl/magma) package) was integrated in libCEED for this
583release. This initial integration set up the framework of using MAGMA and provided the
584libCEED functionality through MAGMA kernels as one of libCEED’s computational backends.
585As any other backend, the MAGMA backend provides extended basic data structures for
586{ref}`CeedVector`, {ref}`CeedElemRestriction`, and {ref}`CeedOperator`, and implements
587the fundamental CEED building blocks to work with the new data structures.
588In general, the MAGMA-specific data structures keep the libCEED pointers to CPU data
589but also add corresponding device (e.g., GPU) pointers to the data. Coherency is handled
590internally, and thus seamlessly to the user, through the functions/methods that are
591provided to support them.
592
593Backends available in this release:
594
595| CEED resource (`-ceed`) | Backend                         |
596|-------------------------|---------------------------------|
597| `/cpu/self`             | Serial reference implementation |
598| `/cpu/occa`             | Serial OCCA kernels             |
599| `/gpu/occa`             | CUDA OCCA kernels               |
600| `/omp/occa`             | OpenMP OCCA kernels             |
601| `/ocl/occa`             | OpenCL OCCA kernels             |
602| `/gpu/magma`            | CUDA MAGMA kernels              |
603
604Examples available in this release:
605
606:::{list-table}
607:header-rows: 1
608:widths: auto
609* - User code
610  - Example
611* - `ceed`
612  - * ex1 (volume)
613* - `mfem`
614  - * BP1 (scalar mass operator)
615    * BP3 (scalar Laplace operator)
616* - `petsc`
617  - * BP1 (scalar mass operator)
618* - `nek5000`
619  - * BP1 (scalar mass operator)
620:::
621
622(v0-2)=
623
624## v0.2 (Mar 30, 2018)
625
626libCEED was made publicly available the first full CEED software distribution, release
627CEED 1.0. The distribution was made available using the Spack package manager to provide
628a common, easy-to-use build environment, where the user can build the CEED distribution
629with all dependencies. This release included a new Fortran interface for the library.
630This release also contained major improvements in the OCCA backend (including a new
631`/ocl/occa` backend) and new examples. The standalone libCEED example was modified to
632compute the volume volume of a given mesh (in 1D, 2D, or 3D) and placed in an
633`examples/ceed` subfolder. A new `mfem` example to perform BP3 (with the application
634of the Laplace operator) was also added to this release.
635
636Backends available in this release:
637
638| CEED resource (`-ceed`) | Backend                         |
639|-------------------------|---------------------------------|
640| `/cpu/self`             | Serial reference implementation |
641| `/cpu/occa`             | Serial OCCA kernels             |
642| `/gpu/occa`             | CUDA OCCA kernels               |
643| `/omp/occa`             | OpenMP OCCA kernels             |
644| `/ocl/occa`             | OpenCL OCCA kernels             |
645
646Examples available in this release:
647
648:::{list-table}
649:header-rows: 1
650:widths: auto
651* - User code
652  - Example
653* - `ceed`
654  - * ex1 (volume)
655* - `mfem`
656  - * BP1 (scalar mass operator)
657    * BP3 (scalar Laplace operator)
658* - `petsc`
659  - * BP1 (scalar mass operator)
660* - `nek5000`
661  - * BP1 (scalar mass operator)
662:::
663
664(v0-1)=
665
666## v0.1 (Jan 3, 2018)
667
668Initial low-level API of the CEED project. The low-level API provides a set of Finite
669Elements kernels and components for writing new low-level kernels. Examples include:
670vector and sparse linear algebra, element matrix assembly over a batch of elements,
671partial assembly and action for efficient high-order operators like mass, diffusion,
672advection, etc. The main goal of the low-level API is to establish the basis for the
673high-level API. Also, identifying such low-level kernels and providing a reference
674implementation for them serves as the basis for specialized backend implementations.
675This release contained several backends: `/cpu/self`, and backends which rely upon the
676[OCCA](http://github.com/libocca/occa) package, such as `/cpu/occa`,
677`/gpu/occa`, and `/omp/occa`.
678It also included several examples, in the `examples` folder:
679A standalone code that shows the usage of libCEED (with no external
680dependencies) to apply the Laplace operator, `ex1`; an `mfem` example to perform BP1
681(with the application of the mass operator); and a `petsc` example to perform BP1
682(with the application of the mass operator).
683
684Backends available in this release:
685
686| CEED resource (`-ceed`) | Backend                         |
687|-------------------------|---------------------------------|
688| `/cpu/self`             | Serial reference implementation |
689| `/cpu/occa`             | Serial OCCA kernels             |
690| `/gpu/occa`             | CUDA OCCA kernels               |
691| `/omp/occa`             | OpenMP OCCA kernels             |
692
693Examples available in this release:
694
695| User code             | Example                           |
696|-----------------------|-----------------------------------|
697| `ceed`                | ex1 (scalar Laplace operator)     |
698| `mfem`                | BP1 (scalar mass operator)        |
699| `petsc`               | BP1 (scalar mass operator)        |
700```
701