Age | Commit message (Collapse) | Author | Files | Lines |
|
By default `openmpi` needs `rsh` from `openssh`, which is a somewhat
redundant dependency for clusters using slurm. This PR adds a toggle to
allow users to disable the ssh/rsh plm altogether.
|
|
|
|
|
|
This package was not setting FFTW when +mklfft was used with +cuda.
Since both were set to 'True', the default build was not linked to
any FFTW, leading to a run time error. It seems MKL support was
conflated with alternative CPU acceleration support. This PR does the
following:
- adds the altcpu variant to specify non-GPU/CPU acceleration
- sets a conflict between +altcpu and +cuda
- sets an FFTW implementation
- sets fltk+xft when +gui to get a decent looking GUI interface
- sets tbb dependency only when +altcpu
- adds dependency on ctffind
- adds variant and dependency on motioncor2
- sets defaults for
- qsub template location
- ctffind location
- motioncor2 location
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
|
|
|
|
fixes #28259
This commit discard specs from unknown namespaces from the
ones that can be "reused" during concretization. Previously
Spack would just error out when encountering them.
|
|
|
|
|
|
(#30307)
|
|
|
|
* fossil: new package
* fossil: Format code
|
|
Allow re-use of run_test() in install_time_test_callbacks
Co-authored-by: Greg Becker <becker33@llnl.gov>
|
|
|
|
* hadoop: Update url and add version 3.3.2
* Update url
|
|
|
|
|
|
|
|
|
|
|
|
Only tested against the latest version, which is why it only applies the patch
when using 4.1.14.
Fix comes from:
https://grigoriefflab.umassmed.edu/forum/software/ctffind_ctftilt/installing_ctffind_4114_ibm_power9
|
|
|
|
|
|
1. Add version 2022.04.17 (new numbering scheme) and update mbuild resource.
2. Branch 'master' is now 'main'.
3. Old rev 10.2019.03 needs a patch for python vs python3.
|
|
See release notes https://github.com/NOAA-EMC/NCEPLIBS-ncio/releases/tag/v1.1.0
|
|
The parent thread in the process stdout redirection logic on Windows
was closing a file that was being read in child thread, which lead to
error-based termination of the reader thread. This updates the
interaction to avoid the error.
|
|
* Add checksum for py-ipywidgets@7.7.0
* Correct py-widgetsnbextension and py-ipython dependencies
* Update widgetsnbextension dependency to 3.6
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Allow requirement to next versions
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Revert ipyhton dependencies
* Add widgetsnbextension@3.6.0 checksum
Co-authored-by: aandvalenzuela <andrea.valenzuela.ramirez@cern.ch>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
|
|
|
|
|
|
* opendatadetector: add new package
* style
* opendatadetector: add hep tag
|
|
|
|
|
|
|
|
Co-authored-by: Tim Haines <thaines.astro@gmail.com>
|
|
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
|
|
|
|
|
|
There's no master branch on GitHub for googletest.
|
|
Co-authored-by: David Kuehn <las_dkuehn@iastate.edu>
|
|
|
|
|
|
|
|
Reduces the number of packages to bootstrap gcc (which actually has
--disable-nls).
|
|
* ASP-based solver: allow configuring target selection
This commit adds a new "concretizer:targets" configuration
section, and two options under it.
- "concretizer:targets:granularity" allows switching from
considering only generic targets to consider all possible
microarchitectures.
- "concretizer:targets:host_compatible" instead controls
whether we can concretize for microarchitectures that
are incompatible with the current host.
* Add documentation
* Add unit-tests
|
|
With an active environment, you can now run "spack concretize --quiet"
and it will suppress printing the concretized specs.
|
|
* MAINT: Add a debug flag
* MAINT: Apply suggestions from code review
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
|
|
appropriate (#30259)
* exago +rocm: pass amdgpu_target to raja, hiop dependencies
* +cuda builds: pass cuda_arch to dependencies
* hiop: pass cuda_arch to dependencies
|
|
Don't apply a Darwin-specific patch that doesn't work on 11.3 any more.
Closes https://github.com/spack/spack/issues/30264.
|
|
Don't apply a Darwin-specific patch that doesn't work on 11.3 any more.
Closes https://github.com/spack/spack/issues/30264.
|
|
* openmpi: always require pmix for 4:
`~pmix` is not applicable to version 4+, since it always builds a vendored
copy of pmix (currently 3.2.3).
* pmix: relax version requirements
When the version range was specified, newer versions didn't exist.
Also use normalized spack versions rather than artificial .9.9 /.0.0.
* openmpi: restrict pmix versions
pmix option isn't available for OpenMPI@1, and according to
https://github.com/open-mpi/ompi/issues/7988 , OpenMPI 4.0.1 will not
build with pmix@3.1.5.
* pmix: add newer versions
* OpenMPI: re-express conflicts/configure logic as conditional variants
This relies partly on `self.enable_or_disable` and its ilk to emit an
empty list when the variant isn't applicable.
|
|
* ASP-based solver: always consider version of installed packages
fixes #29201
Explicitly add facts for versions of installed software when
using the --reuse option, so that we could consider versions
that are not declared in package.py
|