vec - add device impl of axpy and pointwisemult

Update device ID selection for HIP/CUDA backends; add for MAGMA backends

Install install backend headers under include/ceed/

This makes it possible to distribute source plugins that provide
additional backends. It's also used in MFEM, perhaps temporarily.

Deprecate ceed

Install install backend headers under include/ceed/

This makes it possible to distribute source plugins that provide
additional backends. It's also used in MFEM, perhaps temporarily.

Deprecate ceed-backend.h, which was not previously installed, but some
users accessed it from an in-place build.

Also install CUDA and HIP headers that allow users to provide CUfunction
and hipFunction_t.

Co-authored-by: Jeremy L. Thompson <jeremy.thompson@colorado.edu>
Requested-by: Andrew T. Barker <barker29@llnl.gov>

show more ...

Error Handling Improvement [fix #696] (#691)

* Operator - add operator/qfunction field compatibility checks

* QFunction - do not allow adding fields to QFunction in use with an operator

* Exam

Error Handling Improvement [fix #696] (#691)

* Operator - add operator/qfunction field compatibility checks

* QFunction - do not allow adding fields to QFunction in use with an operator

* Examples - add some extra exclusion markers in ceed example

* error - add error enum

* error - update error enum names and numbering

* error - use CEED_ERROR_BACKEND in all backend errors

* error - begin classifying interface errors

* error - update backends to use CEED_ERROR_SUCCESS and CeedChkBackend

* error - use new errors in gallery

* error - add some unsaved modifications

* error - improve documentation

* error - define CEED_ERROR_SUCCESS in GPU JiT; we really should have a common header to pipe defines to the JiT code

* error - more error code editing

* error - fix error string

* operator - fix setting field qpts

* basis - add input/output dimension error checking

* python - move basis utility methods to ceed object, no basis required or used

* python - force exit with negative error code

* make style-py

* rust - initial work to add error handling logic

* rust - add ceed.resource method

* rust - add results for methods that may fail

* rust - also format doctests

* minor - drop unused CeedChk()

* error - rename terminal/nonterminal to major/minor

* rust - set ErrorStore as default errorhandler

* python - revert error handing change for python

* python - use success error code from C bindings

* error - only upgrade error code in backend if positive

show more ...

headers - clearify includes to not rely on transitive includes (#701)

* headers - clearify includes to not rely on transitive includes

* style - add header recommendations from 'include-what-you-

headers - clearify includes to not rely on transitive includes (#701)

* headers - clearify includes to not rely on transitive includes

* style - add header recommendations from 'include-what-you-use'

* style - apply 'include-what-you-use' changes to CUDA backends

* style - 'include-what-you-use' for hip backends

* style - drop ceed.h includes in gallery qf source

* docs - add dev notes for header files

* style - header style and alphabetize

show more ...

Skip hipSetDevice/cudaSetDevice unless specified by user

Some applications will set the device before using libCEED (perhaps
according to a policy based on MPI rank). In such cases, we should leave

Skip hipSetDevice/cudaSetDevice unless specified by user

Some applications will set the device before using libCEED (perhaps
according to a policy based on MPI rank). In such cases, we should leave
it by default rather than override it back to zero.

show more ...

Add static library (libceed.a) [resolve #670]

We no longer use __attribute__((constructor)) to register backends and
gallery implementations because we can't ensure that the symbols are
linked into

Add static library (libceed.a) [resolve #670]

We no longer use __attribute__((constructor)) to register backends and
gallery implementations because we can't ensure that the symbols are
linked into applications that link the static library. We've switched to
creating CeedRegisterAll() and CeedQFunctionRegisterAll(), which are
called automatically by the library, and call weak symbols to register
all the backend/gallery implementations. This strategy was partly
motivated by not wanting to have preprocessor macros describing what is
available, and the associated need to recompile rather than just relink
when those macros change.

So we now have backends/ceed-backend-list.h that declares all the
backends wrapped in a macro. It is included by
backends/ceed-backend-weak.c to create weak definitions of all the
backends. In the makefile, we sort so this comes last when linking a
shared or static library, and thus these weak symbols will only be
picked up if they were not defined by the actual backend source files.
The same header is included (with different macro wrapping) in
interface/ceed-register.c, where CeedRegisterAll() is defined.

To add a new backend, one must do essentially the same registration
strategy as in the past, plus add one line to the common
ceed-backend-list.h.

show more ...

make style

Error checking for user CUfunction in CeedCudaBuildQFunction

Add SetCUDAUserFunction to /gpu/cuda/ref backend

Consolidate CUDA backends (#623)

* Travis - allow icc failure for now

* move cuda-reg basis init kernels to cuda-shared

* move cuda-reg restrictions to cuda-ref

* change delegate ceeds for

Consolidate CUDA backends (#623)

* Travis - allow icc failure for now

* move cuda-reg basis init kernels to cuda-shared

* move cuda-reg restrictions to cuda-ref

* change delegate ceeds for previous uses of cuda-reg

* remove cuda-reg backend

* update hip restrictions to match cuda

* update backends list in README

* make style

* update release notes for removal of cuda-reg

Co-authored-by: jeremylt <thompson.jeremy.luke@gmail.com>

show more ...

QFunction Context Data Object (#596)

* Ctx - create context object for QFunction context data

* Context - rename UserContext -> QFunctionContext

* Ctx - add lcov markers

* Ctx - fix leak in

QFunction Context Data Object (#596)

* Ctx - create context object for QFunction context data

* Context - rename UserContext -> QFunctionContext

* Ctx - add lcov markers

* Ctx - fix leak in identity QFunctions

* Hip/Cuda - rename sync functions for vector/context

* Tests - lcov marker update

* QFunction - drop unused function

* Python - fix copy-paste errors

* Ctx - update notes for Fortran usage

* Fortran - drop unneeded cast

Co-authored-by: Jed Brown <jed@jedbrown.org>

* Interface - use void* for SetData interfaces

* Make - use call quiet for NVCC

* Interface - use void* for GetData interfaces

* Make - add quiet call option for examples

* Makefile - create common makefile to reduce duplication/complexity in example makefiles

Co-authored-by: Jed Brown <jed@jedbrown.org>

show more ...

Cuda/Hip - name QFunctions for easier profiling

Fix CodeCov Reports (#597)

* Tests - use qfunction headers for Fortran tests to improve bypass gcov issue

* Gitlab - use latest gcc on Noether

* Cuda/Hip - add case in reciprocal for completen

Fix CodeCov Reports (#597)

* Tests - use qfunction headers for Fortran tests to improve bypass gcov issue

* Gitlab - use latest gcc on Noether

* Cuda/Hip - add case in reciprocal for completeness

* Cuda - remove duplicate case

* Makefile - exclude fortran test headers from make style

* Travis - update to Focal

* Cov - adjust style to be consistent and avoid false misses

* Travis - update comments and style

show more ...

Op - add interface for multigrid level creation (#579)

* Op - add interface for multigrid level creation

* Op - add implementation for OperatorMultigridLevelCreate

* make style

* make tidy

Op - add interface for multigrid level creation (#579)

* Op - add interface for multigrid level creation

* Op - add implementation for OperatorMultigridLevelCreate

* make style

* make tidy

* Op - add test t550, fix errors

* Tests - add Fortran version of t550

* Tests - add t511 for testing tensor basis multigrid level setup

* make style and tidy

* Tests - fix t55* memory leaks

* Tests - add t552 for non-tensor basis multigrid levels

* CUDA - use CeedIntMax in shared CUDA backend

* Tests - add OCCA test exception for t55*

* Op - add lvector global prolongation multiplicity, simplifies user interface

* Solids - convert example to new interface

* make style

* Tests - convert t550 to multicomponent

* Solids - drop unused ceed_fine

* Python - add new multigrid level interface

* Python - fix operator wrap, use ceed python obj rather than ceed pointer

* Gallery - update comment slightly

* Tests - remove accidental duplicate test

* Multigrid - add size=2 case as well

* Operator - drop unneeded inline

* QFunction - simplify context ownership to match vector

* make style

* Python - update multigrid function signature

* Operator - refactor prolong/restrict qfunctions as scaling qfunctions

* Vector - add testing for reciprocal and add to Fortran/Python interfaces

* CUDA - add VectorReciprocal on device

* Gallery - drop specalized versions for 'Scale', wil fix performance hit later

* Hip - add vector reciprocal

* Operator - add more flexible prologation basis creation interface

* Vec - make sure data is set for VectorReciprocal

* Tests - drop ncomp for t550/1 so kernel is not too large for Magma backend

* Tests - add missing lcov markers

* make style

* Travis - allow ARM job to fail

* Travis - fix intel install

* Travis - try different install dir name for inteloneapi

* Travis - add ifort, ipp packages

* Tests - add missing lcov marker

show more ...

Cuda - add cublasGetErrorName exclusion to match hip

Hip/Cuda - expand QFunction LCOV exception for failing to open qf source file

Device - put device kernels in separate 'kernels' folder in backends

Cuda - add LCOV markers

fix spacing for CeedHostSetValue_Cuda

add hipblas norm to hip-ref

Merge branch 'main' into natalie/hip-ref-v2

Add _Cuda or _Hip to Ceed[Host/Device]SetValue functions

Style - update backend operator function names to use Add as needed

CUDA - change LinearAssemble* to LinearAssembleAdd* when setting backend function