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