* [gromacs] Fix intel (classic) libstdc++ path
Gromacs's `cmake` run will look for `--gcc-toolchain` (e.g. LLVM, icpx) or
`--gcc-name` (e.g. icpc) in `CMAKE_CXX_FLAGS`. Only if it does not find a good
g++ candidate there it will look for `GMX_GPLUSPLUS_PATH`:
cb6b311c39/cmake/FindLibStdCpp.cmake (L97)
Spack installed intel compilers already define a g++ for std libs. But in
`icp{c,x}.cfg` instead of the compile line. If we use the pre-defined g++ we not
only have less chance of mixing g++ versions, but also don't need to explicitly
add `gcc` as dependency to `gromacs`.
* Fix format
* Use a variant
As there is no way to check if a file exists at depends_on stage
* Fix format
* New name and fail if variant is used with other compiler
* Line too long.
The pcluster image has am internal buildcache without an index.
Also, we need to force reuse to avoid rebuilding GCC, since the default is
to only reuse dependencies - and that is subject to changes in the GCC
recipe.
* bedtools2: patching to build with gcc@13
* bedtools2: patching to build with gcc@13
* Update var/spack/repos/builtin/packages/bedtools2/package.py
Yep, sure. Makes sense.
Co-authored-by: Alec Scott <alec@bcs.sh>
---------
Co-authored-by: LMS Bioinformatics <bioinformatics@lms.mrc.ac.uk>
Co-authored-by: Alec Scott <alec@bcs.sh>
* ensmallen: new package
ensmallen is a high-quality C++ library for non-linear numerical optimization.
* r-rcppensmallen: new package
---------
Co-authored-by: wdconinc <wdconinc@users.noreply.github.com>
* Improve lib/spack/spack/test/cmd/compiler.py
* Use "tmp_path" in the "mock_executable" fixture
* Return a pathlib.Path from mock_executable
* Fix mock_executable fixture on Windows
"mock_gcc" was very similar to mock_executable, so use the latter to reduce code duplication
* Remove wrong compiler cache, fix compiler removal
fixes#37996
_CACHE_CONFIG_FILES was both unneeded and wrong, if called
subsequently with different scopes.
Here we remove that cache, and we fix an issue with compiler
removal triggered by having the same compiler spec in multiple
scopes.
* e4s oneapi ci: use official intel oneapi-derived runner image
* update oneapi image
* tau builds ok, but only with libdrm - comment out for now, follow up with pr later
* Guard for define in netcdf 4.9.0 and later.
This code is already available in ParaView 5.11.0 so no patching
needed there.
* Add latest needed version (even if not in spack).
---------
Co-authored-by: Dan Lipsa <dan.lipsa@khq.kitware.com>
* llvm: add new versions and set default for libomptarget according to os
modified: var/spack/repos/builtin/packages/llvm/package.py
* Incorporate reviewer suggestions
Co-authored-by: Sergey Kosukhin <skosukhin@gmail.com>
---------
Co-authored-by: Sergey Kosukhin <skosukhin@gmail.com>
* e4s cray ci stack
* e4s ci: add cray
* add zen4 tag
* WIP: new defintions just for cray
* updates
* remove ci signing job overrride, not necessary
* echo $PATH and show modules loaded
* add mirror
* add external def for cray-libsci
* comment out quantum-espresso
* use /etc/protected-runner as key path
* cray ci stack: do not remove tags: [spack, public]
* make cray stack composable
* generate job should run on public tagged runner, override default config:install_tree:root
* CI: Use relative path in default script
* CI: Use relative includes paths for shell runners
* Use concrete_env_dir for relpath
* ml-darwin-aarch64-mps: jax has bazel codesign issue
---------
Co-authored-by: Scott Wittenburg <scott.wittenburg@kitware.com>
Co-authored-by: Ryan Krattiger <ryan.krattiger@kitware.com>
#37592 updated cached cmake packages to set CMAKE_CUDA_ARCHITECTURES.
The condition `if archs != "none"` lead to `CMAKE_CUDA_ARCHITECTURES=none`
when cuda_arch=none (incorrect check on the value of a multi-valued
variant), i.e. CMAKE_CUDA_ARCHITECTURES is always set. This PR udpates
the condition to if archs[0] != "none" to ensure CMAKE_CUDA_ARCHITECTURES
is only set if cuda_arch is not none (which seems to be the pattern used
in other packages).
This does the same for HIP (although in general ROCmPackage disallows
amdgpu_target=none when +rocm).
