This properly sets PATH/CPATH/LIBRARY_PATH etc. to make the
Spack-generated module file for intel-oneapi-compilers useful
(without this, 'icx' would not be found after loading the module
file for intel-oneapi-compilers).
The C-Library for the current compiler should already be used by the compiler. So there is no point in returning any libs for this package.
Without this patch: if one uses this as an external package (as intended), then this will can inject system library paths into the build process at the wrong place.
* Update recipe for AOMP.
Reduced repitition with version hashes.
Expanded dependency versioning.
Reduced repitition with cmake args.
Added version 3.10.0
* Update dependency versions and remove uneeded quotes.
* Update var/spack/repos/builtin/packages/aomp/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Update of Eccodes to 2.19.1
* PEP8
* PEP8
* PEP8-whitespace
* Update var/spack/repos/builtin/packages/eccodes/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Michael Blaschek <michael.blaschek@univie.ac.at>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* OpenMPI: Depends on hwlock & libevent
Both hwlock & libevent are required dependencies of Open MPI.
While they are also shipped internally, newer releases (>=4.0)
will start looking for external packages by default.
This caused build issues of Open MPI 4.0.5 with Fortran on macOS
10.15.
* Open MPI 4.0: libevent external
Internally shipped libevent just works fine for prior releases.
#20076 moved Cray-specific MPICH support from the Spack MPICH package
to a new cray-mpich Package. This broke existing package installs
using external mpich on Cray systems. This PR keeps the cray-mpich
package but restores the Cray-specific MPICH support for older
installations.
In the future this support should be removed from the Spack mpich
package and users should be directed to use cray-mpich on Cray.
- [x] add `concretize.lp`, `spack.yaml`, etc. to licensed files
- [x] update all licensed files to say 2013-2021 using
`spack license update-copyright-year`
- [x] appease mypy with some additions to package.py that needed
for oneapi.py
* hdf-eos5: new package (HDF for Earth Observing Sytem using hdf v5)
* hdf-eos5: flake8 fixes
* hdf-eos5: trying to fix flake8 errors
* hdf-eos5: flake8 fix
* hdf-eos5: Fix to support Fortran codes
The -Df2cFortran compilation flag needed to support Fortran
* hdf-eos2: new package (HDF for Earth Observing System using hdf5)
* hdf-eos2: flake8 fixes
* hdf-eos2: fix to support Fortran
Need the compilation flag -Df2cFortran to allow support for Fortran
codes
libuuid is currently contained in util-linux, libuuid and uuid. This
change introduces a new virtual provider `uuid` and renames the existing
`uuid` package to `ossp-uuid`.
util-linux's libuuid is provided in the form of a separate package
util-linux-uuid to make sure that packages depending on uuid and
util-linux can use a separate uuid implementation, which the concretizer
does not allow if libuuid is contained in util-linux.
- added several patches
- added some missing dependencies
- remove unneeded dependencies
- add CUDA support
- disable queue support, which was limited, and broken anyway
- move package text that was specific to the package to a comment, so it
does not show up the environment module
- set conflicts for cuda and compilers
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
* OpenMPI: Add version 4.1.0
* OpenMPI: Prefer version 4.0.5.
* OpenMPI: Update links
The download links changed, there is currently a redirection but it might not work forever. The website also switched to https.
Previously compiler-rt didn't correctly passthrough cmake
variables for python when building the various santizers.
This patch passes these variables through.
This patch may also correctly apply to any version of LLVM
to any version of LLVM that uses the newer monorepo style organization,
and any older llvm newer than 7.0.0 as long as the paths were set
appropriately. However, this was not done so because it was not
tested with older LLVM releases.
Fixes#19908
See also: https://bugs.llvm.org/show_bug.cgi?id=48180
This updates the UnifyFS packages to account for the latest v0.9.1
release.
Updates required and optional dependencies for the respective
releases.
Locks margo and mercury dependencies at specific versions while
integration with their latest versions is still in progress.
* PGI compiler has trouble with avx2 SIMD support
(https://github.com/FFTW/fftw3/issues/78)
* Hew to the project's preferred indentation standard.
* Expand '%nvhpc' logic to include '%pgi'.
* Exceeded the max line-length.
* Break up the long compound statement into nested if's.
* Inadvertently picked up an extraneous file.
* PGI compiler has trouble with avx2/avx-512 SIMD support, too.
* Add PGI runtime libs to LDFLAGS when '%pgi' in spec.
* Revert "Add PGI runtime libs to LDFLAGS when '%pgi' in spec."
This reverts commit 31c3ef8ea2.
* Add PGI runtime libs to LDFLAGS when '%pgi' in spec.
* Adding support for the CMake flags in LBANN that are missing.
* Added new flag to OpenCV dependency and removed negative variants
since OpenCV no longer turns on everything by default. Removed CMake
flags in LBANN that have been deprecated.
* Removed type='build' flags from dependencies so that they get linked
into a environment's view.
* Removed type='build' flags from dependencies so that they get linked
into a environment's view. Fixed DiHydrogen variant to enable
DistConv feature, renamed to +distconv from +legacy. Added conflicts
line to indicated that DistConv and ROCm don't work with +half
support.
* Fixed Flake8 and cleaned up ordering of variants.
* Flake8
* Backed out changes to not mark and cmake and ninja as build
dependencies, which was introduced to make sure that they appear in
a spack environment.
* Backed out changes to not mark doc related packages as build
dependencies, which was introduced to make sure that they appear
in a spack environment.
* Fixed how recipe communicates the intent to build and run tests to the
package CMake.
This is to make sure that the build system doesn't pick up a library that
would happen to be available.
Co-authored-by: Baptiste Jonglez <git@bitsofnetworks.org>
This creates a set of packages which all use the same script to install
components of Intel oneAPI. This includes:
* An inheritable IntelOneApiPackage which knows how to invoke the
installation script based on which components are requested
* For components which include headers/libraries, an inheritable
IntelOneApiLibraryPackage is provided to locate them
* Individual packages for DAL, DNN, TBB, etc.
* A package for the Intel oneAPI compilers (icx/ifx). This also includes
icc/ifortran but these are not currently detected in this PR
This patch logic resovles a linking issue with ncurses in the mesa
package. This appears to be a recurring problem that was identified in
the mesa gitlab issues here:
https://gitlab.freedesktop.org/mesa/mesa/-/issues/2843
Using `_llvm_method = 'auto'` is broken. This patch replaces that with
`_llvm_method = 'config-tool'`, which is a hack, but makes it possible
to build.
I have commented on the closed issue (2843), referencing the original
author of the bug, and one of the mesa developers, so perhaps they will
fix the problem.
* Dependencies of Go will now correctly set the GOPATH for the
appropriate spec to avoid using the user's default path.
* Bumped version to latest releases(1.15.6 & 1.14.13).
Most people installing `clingo` with Spack are going to be doing it to
use the new concretizer, and that requires the `master` branch.
- [x] make `master` the default so we don't have to keep telling people
to install `clingo@master`. We'll update the preferred version when
there's a new release.
Cray's version of MPICH uses a different versioning system than
MPICH, so it has been split into its own package. It is an
external-only package (always provided by the system, never
installed by Spack).
* Kluge to get the gfortran linker to work correctly on Big Sur.
* Fixed formatting error; stetting the other.
* Removed spaces.
* Added comment, mainly to re-trigger Spack CI.
* ParaView: add new ParaView-5.9.0-RC2 release
Signed-off-by: Vicente Adolfo Bolea Sanchez <vicente.bolea@kitware.com>
* Update var/spack/repos/builtin/packages/paraview/package.py
Indeed, I misunderstood the previous review. This looks good to me too.
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Fixed hard tab in flux-sched edit and unbound hwloc in flux-core after
testing to better support modern MPIs in spack environments
Verified that flux-core@0.17 is when hwloc@2: became viable
Track all the variant values mentioned when emitting constraints, validate them
and emit a fact that allows them as possible values.
This modification ensures that open-ended variants (variants accepting any string
or any integer) are projected to the finite set of values that are relevant for this
concretization.
2020.10.0 is the latest stable release, and the preferred version
for general use (when the user does not specify otherwise).
2020.11.0 is a prototype for the memory kinds feature that is also
available when requested.
* [cmd versions] add spack versions --new flag to only fetch new versions
format
[cmd versions] rename --latest to --newest and add --remote-only
[cmd versions] add tests for --remote-only and --new
format
[cmd versions] update shell tab completion
[cmd versions] remove test for --remote-only --new which gives empty output
[cmd versions] final rename
format
* add brillig mock package
* add test for spack versions --new
* [brillig] format
* [versions] increase test coverage
* Update lib/spack/spack/cmd/versions.py
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
* Update lib/spack/spack/cmd/versions.py
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
Update geant4-data and individual datasets for Geant4 versions 10.6.3
and 10.7.0.
Update geant4 package with new versions 10.6.3 and 10.7.0. Update
dependencies on CLHEP and VecGeom with versions required for Geant4
10.7.
Add GEANT4_INSTALL_PACKAGE_CACHE=OFF to CMake args for 10.6 onwards.
Prevents install of the "package cahce" file that contains hard-coded
paths for dependencies, improving relocatability. It relies on Spack
setting CMAKE_PREFIX_PATH correctly in build/use environments that
consume the geant4 package.
`cmake @3.17:` is necessary to handle `cuda @11:` correctly. Earlier versions of `cmake` do not know that `cuda @11:` does not support `compute_30` any more, and list that compute capability as supported. This is handled in `cmake`'s file `Modules/FindCUDA/select_compute_arch.cmake`.
The bowtie2 Makefile uses `prefix`, not `PREFIX`, for versions before v2.4.
Credit to @tkameyama
Co-authored-by: george.hartzell <george.hartzell@sana.com>
* bump up version for rocm-3.10.0 release
* bump up version for rocm-3.10.0
* remove duplicate version addition for 3.9.0
* bump up version for rocm-3.10.0 release
* bump up version for rocm-3.10.0 release
* bump up version for rocm-debug-agent and rocm-dbgapi
* bump up version for rocm-bandwidth-test,rocm-gdb,rocprofiler,roctracer for rocm-3.10.0
* add smoke test
* remove whitespaces
* fix minimum version issue
* reorder decorators & replace make with cmake build
* merge cmake build into one line
* reorganize smoke test function
Co-authored-by: Jieyang Chen <chenj3@ornl.gov>
* New package: py-qsymm
* py-qsymm: Convert to using tarballs from PyPi instead of git checkouts
* py-qsymm: add missing dependencies
* Update var/spack/repos/builtin/packages/py-qsymm/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* py-qsymm: Fix url to use pypi hidden download interface
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* AOCC-2.3.0 is now added to spack
Change-Id: I18fd9606e6fd9a288cc7dc6c6ead11ea17839a7c
* Added flag and version tests for AOCC-2.3.0
* Addressed review comments
Co-authored-by: vkallesh <Vijay-teekinavar.Kallesh@amd.com>
fixes#20040
Matching compilers among nodes has been prioritized
in #20020. Selection of default variants has been
tuned in #20182. With this setup there is no need
to have an ad-hoc rule for external packages. On
the contrary it should be removed to prefer having
default variant values over more external nodes in
the DAG.
refers #20040
Before this PR optimization rules would have selected default
providers at a higher priority than default variants. Here we
swap this priority and we consider variants that are forced by
any means (root spec or spec in depends_on clause) the same as
if they were with a default value.
This prevents the solver from avoiding expected configurations
just because they contain directives like:
depends_on('pkg+foo')
and `+foo` is not the default variant value for pkg.
* OpenBLAS: More Precise GCC Conflicts
Add more precise GCC conflicts so e.g. GCC 6 and GCC 7.5 don't fail.
* Compact syntax
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
As part of pull request #19452, a patch method was added to the mfem
package to delete byte order marks from 3 mfem source files. These
files first appeared in a stable release of mfem as of version
4.1. Consequently, attempts to install mfem 3.4 or mfem 4.0 fail
because no files exist at the path arguments of the filter_file
commands used to execute this operation. Decorating the patch method
so it runs only on mfem versions 4.1 and later resolves the errors
that were thrown due to files not found.
This commit adds that decorator.
* Qt: add options to disable docs and gui
- Add `~gui` option for minimal build
- Add `+doc` option to install docs, and attempt to disable the implicit
llvm dependency if not
- Removes the 'freetype' option which hasn't worked reliably in qt5, as
many of the gui components implicitly rely on freetype.
- Add and test version 5.15 (and skip qtlocation if disabling opengl)
- Refactor some of the dependency logic
I've tested this on linux with 5.15.2 and 4.8.7 in a couple of different
configurations.
* Address reviewer feedback and correctly disable llvm
* Fix qt doc generation
* py-rosdep: add new package
* setuptools needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
* py-rospkg: add new package
* setuptools needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
* py-catkin-pkg: add new package
* setuptools is needed at run-time
Co-authored-by: Andrew W Elble <aweits@rit.edu>
Co-authored-by: Andrew W Elble <aweits@rit.edu>
fixes#19981
This commit adds support for target ranges in directives,
for instance:
conflicts('+foo', when='target=x86_64:,aarch64:')
If any target in a spec body is not a known target the
following clause will be emitted:
node_target_satisfies(Package, TargetConstraint)
when traversing the spec and a definition of
the clause will then be printed at the end similarly
to what is done for package and compiler versions.
* spack recipe for gromacs with aocc compiler support
Change-Id: I364aab4a0aa2dcd44bc47eb50c81b2d94c99cfbd
* Removed arch and other associated compilers flags
Added cycle_subcounters variant
Co-authored-by: vkallesh <Vijay-teekinavar.Kallesh@amd.com>
fixes#20019
Before this modification having a newer version of a node came
at higher priority in the optimization than having matching
compilers. This could result in unexpected configurations for
packages with conflict directives on compilers of the type:
conflicts('%gcc@X.Y:', when='@:A.B')
where changing the compiler for just that node is preferred to
lower the node version to less than 'A.B'. Now the priority has
been switched so the solver will try to lower the version of the
nodes in question before changing their compiler.
* llvm-amdgpu: fix the build for version 3.9.0
Adapt the fix-system-zlib-ncurses.patch for version 3.9.0. Without
the patch, llvm-amdgpu builds, but then rocm-device-libs fails with
"cannot find -ltinfo."
Tighten the version requirements for cmake according to the
llvm/CMakeLists.txt file.
* Add a conflict for cmake 3.19.0.
refers #20079
Added docstrings to 'concretize' and 'concretized' to
document the format for tests.
Added tests for the activation of test dependencies.
refers #20040
This modification emits rules like:
provides_virtual("netlib-lapack","blas") :- variant_value("netlib-lapack","external-blas","False").
for packages that provide virtual dependencies conditionally instead
of a fact that doesn't account for the condition.
* intel-tbb: patch for arm64 on macOS
as submitted upstream and used in homebrew
* intel-tbb: check patchable versions
* intel-tbb: avoid patch breakage when 2021.1 is released
2021.1-beta05 would be considered newer than 2021.1
* Add the 'exciting' package.
Version 14 (latest available) is defined.
An as-of-yet unpublished patch (dfgather.patch) from the developers is also
included.
* fixed flake8 errors (I *thought* I had already gotten them! OOPS!)
* Update var/spack/repos/builtin/packages/exciting/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* fixed install method to just do the install, and no build method is needed.
* *Actually* added the lapack dependency!
* removed variant from blas dependency
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* h5cpp: Correct checksum and build instructions
Closes https://github.com/spack/spack/issues/20046.
* h5cpp: Update to 1.10.4-6
* Update var/spack/repos/builtin/packages/h5cpp/package.py
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* h5cpp: Correct formatting
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* fix: leading . is not needed in extension kwarg
* mfem: add support for NVIDIA AmgX
fix: proper spacing
* mfem: use conflict to indicate that AmgX is expected to depend on CUDA
