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