Use spack¶
List the available packages¶
$ spack list
==> 3247 packages.
abinit py-fiona
abyss py-fiscalyear
…
or to filter for certain packages, for example
$ spack list numpy
==> 2 packages.
py-numpy py-numpydoc
List the installed packages¶
$ spack find
==> 17 installed packages
-- darwin-mojave-x86_64 / clang@10.0.1-apple --------------------
bzip2@1.0.8 libffi@3.2.1 perl@5.26.2 python@3.7.4 zlib@1.2.11
diffutils@3.7 ncurses@6.1 pkgconf@1.6.1 readline@7.0
expat@2.2.5 openblas@0.3.6 py-numpy@1.16.4 sqlite@3.28.0
gdbm@1.18.1 openssl@1.1.1b py-setuptools@41.0.1 xz@5.2.4
spack info¶
$ spack info py-numpy
PythonPackage: py-numpy
Description:
NumPy is the fundamental package for scientific computing with Python.
It contains among other things: a powerful N-dimensional array object,
sophisticated (broadcasting) functions, tools for integrating C/C++ and
Fortran code, and useful linear algebra, Fourier transform, and random
number capabilities
Homepage: http://www.numpy.org/
Tags:
None
Preferred version:
1.16.4 https://pypi.io/packages/source/n/numpy/numpy-1.16.4.zip
Safe versions:
1.16.4 https://pypi.io/packages/source/n/numpy/numpy-1.16.4.zip
1.16.3 https://pypi.io/packages/source/n/numpy/numpy-1.16.3.zip
1.16.2 https://pypi.io/packages/source/n/numpy/numpy-1.16.2.zip
1.16.1 https://pypi.io/packages/source/n/numpy/numpy-1.16.1.zip
1.16.0 https://pypi.io/packages/source/n/numpy/numpy-1.16.0.zip
1.15.4 https://pypi.io/packages/source/n/numpy/numpy-1.15.4.zip
1.15.3 https://pypi.io/packages/source/n/numpy/numpy-1.15.3.zip
1.15.2 https://pypi.io/packages/source/n/numpy/numpy-1.15.2.zip
1.15.1 https://pypi.io/packages/source/n/numpy/numpy-1.15.1.zip
1.15.0 https://pypi.io/packages/source/n/numpy/numpy-1.15.0.zip
1.14.6 https://pypi.io/packages/source/n/numpy/numpy-1.14.6.zip
1.14.5 https://pypi.io/packages/source/n/numpy/numpy-1.14.5.zip
1.14.4 https://pypi.io/packages/source/n/numpy/numpy-1.14.4.zip
1.14.3 https://pypi.io/packages/source/n/numpy/numpy-1.14.3.zip
1.14.2 https://pypi.io/packages/source/n/numpy/numpy-1.14.2.zip
1.14.1 https://pypi.io/packages/source/n/numpy/numpy-1.14.1.zip
1.14.0 https://pypi.io/packages/source/n/numpy/numpy-1.14.0.zip
1.13.3 https://pypi.io/packages/source/n/numpy/numpy-1.13.3.zip
1.13.1 https://pypi.io/packages/source/n/numpy/numpy-1.13.1.zip
1.13.0 https://pypi.io/packages/source/n/numpy/numpy-1.13.0.zip
1.12.1 https://pypi.io/packages/source/n/numpy/numpy-1.12.1.zip
1.12.0 https://pypi.io/packages/source/n/numpy/numpy-1.12.0.zip
1.11.3 https://pypi.io/packages/source/n/numpy/numpy-1.11.3.zip
1.11.2 https://pypi.io/packages/source/n/numpy/numpy-1.11.2.zip
1.11.1 https://pypi.io/packages/source/n/numpy/numpy-1.11.1.zip
1.11.0 https://pypi.io/packages/source/n/numpy/numpy-1.11.0.zip
1.10.4 https://pypi.io/packages/source/n/numpy/numpy-1.10.4.zip
1.9.3 https://pypi.io/packages/source/n/numpy/numpy-1.9.3.zip
1.9.2 https://pypi.io/packages/source/n/numpy/numpy-1.9.2.zip
1.9.1 https://pypi.io/packages/source/n/numpy/numpy-1.9.1.zip
Variants:
Name [Default] Allowed values Description
blas [on] True, False Build with BLAS support
lapack [on] True, False Build with LAPACK support
Installation Phases:
build install
Build Dependencies:
blas lapack py-setuptools python
Link Dependencies:
blas lapack python
Run Dependencies:
python
Virtual Packages:
None
spack version¶
spack version shows the available versions, for example
$ spack versions python
==> Safe versions (already checksummed):
3.7.4 3.7.0 3.6.5 3.6.1 3.5.1 3.3.6 2.7.15 2.7.11
3.7.3 3.6.8 3.6.4 3.6.0 3.5.0 3.2.6 2.7.14 2.7.10
3.7.2 3.6.7 3.6.3 3.5.7 3.4.10 3.1.5 2.7.13 2.7.9
3.7.1 3.6.6 3.6.2 3.5.2 3.4.3 2.7.16 2.7.12 2.7.8
==> Remote versions (not yet checksummed):
3.8.0b2 3.6.9 3.5.7rc1 3.5.0a2 3.4.0 3.1.2 2.7 2.4.3
3.8.0b1 3.6.8rc1 3.5.6rc1 3.5.0a1 3.3.7rc1 3.1.1 2.6.9 2.4.2
…
Installation of certain packages¶
for example:
$ spack install python@3.7.4
or to install py-numpy for Python 3.7.4:
$ spack install py-numpy ^python@3.7.4
Then the installation can be checked with
$ spack find --deps py-numpy
==> 1 installed package
-- darwin-mojave-x86_64 / clang@10.0.1-apple --------------------
py-numpy@1.16.4
^openblas@0.3.6
^python@3.7.4
^bzip2@1.0.8
^expat@2.2.5
^gdbm@1.18.1
^readline@7.0
^ncurses@6.1
^libffi@3.2.1
^openssl@1.1.1b
^zlib@1.2.11
^sqlite@3.28.0
^xz@5.2.4
Uninstall¶
$ spack uninstall py-numpy
or
$ spack uninstall --dependents py-numpy
Extensions and Python support¶
The Spack installation model assumes that each package lives in its own
installation prefix. Modules in interpreted languages such as Python are
typically installed in $prefix/lib/python-3.7/site-packages/, for example
/Users/veit/spack/opt/spack/darwin-mojave-x86_64/clang-10.0.1-apple/py-numpy-1.16.4-45sqnufha2yprpx6rxyelsokky65ucdy/lib/python3.7/site-packages/numpy.
However, packages installed in a different prefix can also be used. Such a
package is called an extension in Spack.
Suppose Python was installed with
$ spack find python
==> 1 installed package
-- darwin-mojave-x86_64 / clang@10.0.1-apple --------------------
python@3.7.4
so Extensions can be found with
$ spack extensions python
==> python@3.7.4%clang@10.0.1-apple+bz2+ctypes+dbm+lzma~nis~optimizations patches=210df3f28cde02a8135b58cc4168e70ab91dbf9097359d05938f1e2843875e57 +pic+pyexpat+pythoncmd+readline~shared+sqlite3+ssl~tix~tkinter~ucs4~uuid+zlib arch=darwin-mojave-x86_64/jqlxzxp
==> 623 extensions:
adios2 py-munch
antlr py-mx
…
==> 2 installed:
-- darwin-mojave-x86_64 / clang@10.0.1-apple --------------------
py-numpy@1.16.4 py-setuptools@41.0.1
==> None activated.
numpy can be added to the PYTHONPATH of the current shell with load:
$ spack load python
$ spack load py-numpy
$ python
Python 3.7.4 (default, Jul 28 2019, 20:00:06)
[Clang 10.0.1 (clang-1001.0.46.4)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>>
Often, however, certain packages should be permanently available to a Python
installation. Spack offers activate for this:
$ spack activate py-numpy
==> Activating extension py-numpy@1.16.4%clang@10.0.1-apple+blas+lapack arch=darwin-mojave-x86_64/45sqnuf for python@3.7.4%clang@10.0.1-apple+bz2+ctypes+dbm+lzma~nis~optimizations patches=210df3f28cde02a8135b58cc4168e70ab91dbf9097359d05938f1e2843875e57 +pic+pyexpat+pythoncmd+readline~shared+sqlite3+ssl~tix~tkinter~ucs4~uuid+zlib arch=darwin-mojave-x86_64/jqlxzxp