* py-chainer: Add test method for ChainerMN (continued #21848, #21940)
* py-chainer: Fixed the word in the message
* py-chainer: Delete unnecessary imports
* py-chainer: Incorporation of the measures pointed out in #21940 was insufficient.
Adds several EpetraExt_BUILD_* options as well as an Amesos2_ENABLE_Basker option. Adds `none` as an option to `gotype=`, which should be among the options since 'none' is specifically handled later in the package definition.
Adds `stokhos` and `trilinoscouplings` as options in spack which already are available in CMake for Trilinos (e.g. Trilinos_ENABLE_Stokhos:BOOL=)
* py-pytest-html recipe
* added missing deps + copyright
* Update var/spack/repos/builtin/packages/py-pytest-html/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update var/spack/repos/builtin/packages/py-pytest-metadata/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Patch eospac's Makefile.-linux-gnu.hashes to consider only `$(notdir
$(F90))` when constructing a key to look up compiler flags in the
_F90-CPUINFO_COMP_FLAGS associative array.
This patch was accepted into eospac itself after the release of
6.4.2beta, so apply it only to 6.4.2beta and earlier releases.
- Fix faulty patch
- Only use GEARSHIFFT_BACKEND_FFTW_PTHREADS if ~openmp
- Explicitly disable float16 support
- Use correct minimum required Boost version
- Add variants for Intel MKL and ROCm rocfft
In the past, we only had the binutils variant, which included the
bootstrapping flag. Now that we have a separate bootstrap variant, fix
the nvptx conflict accordingly.
* Add intel cluster package update2 for 2020
* add pacifica cli tools, and pager
* remove boilerplate code
* update flake8 lints
* update flake8 lint, missed one
* add a description for pager
* Shorten a line
* Remove whitespace
* check on dependencies and move urls to proper place
* Remove import package as it seems it is not required
* add requests to the uploader config
* remove blank Line
* change to build and run for packages
* add run and build to the packages
* move from url method to pypi method
* adjust requirements based on feedback from adamjstewart
* remove python 3 requirement, and add setuptools-scm
* remove dependence on python
Co-authored-by: Evan Felix <evan.felix@pnnl.gov>
Cray added necessary functionality for CMake to support fortran preprocessing using crayftn. This patch is necessary for the current release of cmake (3.19), with this patched expected to be in the 3.20 release of Cmake. The included patch is from kitware.
see https://gitlab.kitware.com/cmake/cmake/-/merge_requests/5882
Co-authored-by: James Elliott <jjellio@sandia.govv>
* adding package for libabigail, which we likely will need to use it for an analysis!
* includes variant for documentation (doxygen and pysphinx are associated dependencies)
Remove casacore's old version of the file with a package patch()
function, and depend on a modern CMake for the build.
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
1. Add version 2021.03.01.
2. Cleanup the binutils dependencies now that 2.35.2 and 2.36 are available.
3. Require gcc 7.x or later for current 2021 version.
4. Simplify the xz depends to always require +pic.
This works around a glitch in the original concretizer.
For opt-in packages in Spack, its common that the `cuda` variant
is disabled by default.
This also simplifies downstream usage in multi-variants for
backends in user code.
- as outlined in merge-request #21336 some clang compilers
can trigger erroneous floating point exceptions.
OpenFOAM normally traps FPE, but disable this in the etc/controlDict
for specific compilers:
change "trapFpe digit;" -> "trapFpe 0;"
Eliminate previous use of FOAM_SGIFPE env variable in favour of
using the etc/controlDict setting - cleaner and robuster.
Co-authored-by: Mark Olesen <Mark.Olesen@esi-group.com>
* py-importlib: Python 2.7 is needed to build.
added depends_on('python@2.7.0:2.7.99')
* Update var/spack/repos/builtin/packages/py-importlib/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
This check is performed in cmake_args rather than with a 'conflicts'
statement because matching on !clang (i.e. any compiler that is not
clang) cannot currently be done with our spec syntax.
* asciidoc: current sourceforge a2x needs python2, new github release python3
* asciidoc: current sourceforge a2x needs python2, new github release python3
* asciidoc: making python 2.3 to 2.7 able to cope with asciidoc
* asciidoc: current sourceforge a2x needs python2, new github release python3
* asciidoc: current sourceforge a2x needs python2, new github release python3
* asciidoc: current sourceforge a2x needs python2, new github release python3
* asciidoc: current sourceforge a2x needs python2, new github release python3
* Fix sensei@develop
Should work with all options but libsim.
Current releases don't work with ~catalyst
See
https://gitlab.kitware.com/sensei/sensei/-/merge_requests/240
for the fix for develop.
Current releases work only with paraview 5.7 and 5.6
See
https://gitlab.kitware.com/sensei/sensei/-/merge_requests/239
for the fix for develop (which works with 5.9)
* Fix libsim.
* Fix warnings.
* Fix python runtime.
* Many changes:
* Reworked cmake options top use the CMakePackage option helpers
* Simplified and consolidated options
* Replaced adios with adios2 variant
* Added vtkm variant (not yet working)
* paraview: Fix downstream consumers getting the wrong FindMPI
* vtk: Fix downstream consumers getting the wrong FindMPI
* Add +ascent, +adios2; remove +adios; variants off by default
* Fix catalyst python logic
* sensei: cleanup formatting
Co-authored-by: Chuck Atkins <chuck.atkins@kitware.com>
When using an external package with the old concretizer, all
dependencies of that external package were severed. This was not
performed bidirectionally though, so for an external package W with
a dependency on Z, if some other package Y depended on Z, Z could
still pull properties (e.g. compiler) from W since it was not
severed as a parent dependency.
This performs the severing bidirectionally, and adds tests to
confirm expected behavior when using config from DAG-adjacent
packages during concretization.
Allow libfuse to build without setuid binary and bump versions of both
libfuse and fuse-overlayfs.
Still doesn't solve the issue where this package tries to install things
into /etc/init.d though.
kcov CMakeLists.txt generates the "kcov" executable only if
certain dependencies are found. These dependencies are
"libbfd", "libopcodes" and "libiberty", hence the dependency
on binutils.
* n2p2: Add new package
* remove ,
* Resurrection of , and changed " to single
* changed example.command to example.co
* n2p2: Added v2.1.1
* n2p2: Changed the type of depends_on.
Since there are many variables being set I thought it would be a good idea to document them better and slightly simplify the logic for external vs not-external.
* Spec.splice feature
Construct a new spec with a dependency swapped out. Currently can only swap dependencies of the same name, and can only apply to concrete specs.
This feature is not yet attached to any install functionality, but will eventually allow us to "rewire" a package to depend on a different set of dependencies.
Docstring is reformatted for git below
Splices dependency "other" into this ("target") Spec, and return the result as a concrete Spec.
If transitive, then other and its dependencies will be extrapolated to a list of Specs and spliced in accordingly.
For example, let there exist a dependency graph as follows:
T
| \
Z<-H
In this example, Spec T depends on H and Z, and H also depends on Z.
Suppose, however, that we wish to use a differently-built H, known as H'. This function will splice in the new H' in one of two ways:
1. transitively, where H' depends on the Z' it was built with, and the new T* also directly depends on this new Z', or
2. intransitively, where the new T* and H' both depend on the original Z.
Since the Spec returned by this splicing function is no longer deployed the same way it was built, any such changes are tracked by setting the build_spec to point to the corresponding dependency from the original Spec.
Co-authored-by: Nathan Hanford <hanford1@llnl.gov>
* Made DiHydrogen a required dependencies on newer versions of LBANN.
Added an explicit variant for enabling Boost-dependent callbacks.
Updated the separation for embedded Python and the Python front end
code and associated dependencies.
* Bugfix on ROCm include in DiHydrogen
BlasPP by ECP SLATE will fail to install by default
(`spack install blaspp`) because:
- the default BLAS installation in Spack is OpenBLAS
- BlasPP conflicts with `threads=none` for all recent OpenBLAS releases
OpenBLAS introduced a threadsafe compile option
with 0.3.7+ aka `USE_LOCKING`:
```
61 # If you want to build a single-threaded OpenBLAS, but expect to call this
62 # from several concurrent threads in some other program, comment this in for
63 # thread safety. (This is done automatically for USE_THREAD=1 , and should not
64 # be necessary when USE_OPENMP=1)
65 # USE_LOCKING = 1
```
According to tests, with `spack install --test root blaspp`,
this exactly addresses the issues in BlasPP tests.
It also seems to be a good option to set by default for OpenBLAS and
users that do not need this safety net can always disable it.
Solve issues with newer OpenBLAS by requiring
`+locking` over none-default threading options.
