NumPy 1.26.0 Release Notes#
The NumPy 1.26.0 release is a continuation of the 1.25.x release cycle with the addition of Python 3.12.0 support. Python 3.12 dropped distutils, consequently supporting it required finding a replacement for the setup.py/distutils based build system NumPy was using. We have chosen to use the Meson build system instead, and this is the first NumPy release supporting it. This is also the first release that supports Cython 3.0 in addition to retaining 0.29.X compatibility. Supporting those two upgrades was a large project, over 100 files have been touched in this release. The changelog doesn’t capture the full extent of the work, special thanks to Ralf Gommers, Sayed Adel, Stéfan van der Walt, and Matti Picus who did much of the work in the main development branch.
The highlights of this release are:
Python 3.12.0 support.
Cython 3.0.0 compatibility.
Use of the Meson build system
Updated SIMD support
f2py fixes, meson and bind(x) support
Support for the updated Accelerate BLAS/LAPACK library
The Python versions supported in this release are 3.9-3.12.
Array API v2022.12 support in
numpy.array_apinow full supports the v2022.12 version of the array API standard. Note that this does not yet include the optional
fftextension in the standard.
Support for the updated Accelerate BLAS/LAPACK library#
Support for the updated Accelerate BLAS/LAPACK library, including ILP64 (64-bit integer) support, in macOS 13.3 has been added. This brings arm64 support, and significant performance improvements of up to 10x for commonly used linear algebra operations. When Accelerate is selected at build time, the 13.3+ version will automatically be used if available.
meson backend for
f2py in compile mode (i.e.
f2py -c) now accepts the
option. This is the default option for Python
3.12 on-wards. Older versions
will still default to
To support this in realistic use-cases, in compile mode
f2py takes a
--dep flag one or many times which maps to
dependency() calls in the
meson backend, and does nothing in the
There are no changes for users of
f2py only as a code generator, i.e.
bind(c) support for
Both functions and subroutines can be annotated with
handle both the correct type mapping, and preserve the unique label for other
bind(c, name = 'routine_name_other_than_fortran_routine') is not
honored by the
f2py bindings by design, since
bind(c) with the
is meant to guarantee only the same name in
Fortran, not in
iso_c_binding support for
Previously, users would have to define their own custom
f2cmap file to use
type mappings defined by the Fortran2003
iso_c_binding intrinsic module.
These type maps are now natively supported by
Build system changes#
In this release, NumPy has switched to Meson as the build system and
meson-python as the build backend. Installing NumPy or building a wheel can be
done with standard tools like
pypa/build. The following are
pip install numpyor (in a cloned repo)
pip install .
Building a wheel:
python -m build(preferred), or
pip wheel .
pip install -e . --no-build-isolation
Development builds through the custom CLI implemented with spin:
All the regular
pypa/build flags (e.g.,
--no-build-isolation) should work as expected.
NumPy-specific build customization#
Many of the NumPy-specific ways of customizing builds have changed.
NPY_* environment variables which control BLAS/LAPACK, SIMD, threading,
and other such options are no longer supported, nor is a
site.cfg file to
select BLAS and LAPACK. Instead, there are command-line flags that can be
passed to the build via
build’s config-settings interface. These
flags are all listed in the
meson_options.txt file in the root of the repo.
Detailed documented will be available before the final 1.26.0 release; for now
please see the SciPy “building from source” docs since most build
customization works in an almost identical way in SciPy as it does in NumPy.
While the runtime dependencies of NumPy have not changed, the build
dependencies have. Because we temporarily vendor Meson and meson-python,
there are several new dependencies - please see the
pyproject.toml for details.
This build system change is quite large. In case of unexpected issues, it is
still possible to use a
setup.py-based build as a temporary workaround (on
Python 3.9-3.11, not 3.12), by copying
pyproject.toml. However, please open an issue with details on the NumPy
issue tracker. We aim to phase out
setup.py builds as soon as possible, and
therefore would like to see all potential blockers surfaced early on in the
1.26.0 release cycle.
A total of 20 people contributed to this release. People with a “+” by their names contributed a patch for the first time.
Albert Steppi +
Bas van Beek
Filipe Laíns +
Liang Yan +
Marten van Kerkwijk
Melissa Weber Mendonça
Stefan van der Walt
Pull requests merged#
A total of 59 pull requests were merged for this release.
#24305: MAINT: Prepare 1.26.x branch for development
#24308: MAINT: Massive update of files from main for numpy 1.26
#24322: CI: fix wheel builds on the 1.26.x branch
#24326: BLD: update openblas to newer version
#24327: TYP: Trim down the
#24328: BUG: fix choose refcount leak
#24337: TST: fix running the test suite in builds without BLAS/LAPACK
#24338: BUG: random: Fix generation of nan by dirichlet.
#24340: MAINT: Dependabot updates from main
#24342: MAINT: Add back NPY_RUN_MYPY_IN_TESTSUITE=1
#24353: MAINT: Update
#24356: TST: fix distutils tests for deprecations in recent setuptools…
#24375: MAINT: Update cibuildwheel to version 2.15.0
#24381: MAINT: Fix codespaces setup.sh script
#24403: ENH: Vendor meson for multi-target build support
#24404: BLD: vendor meson-python to make the Windows builds with SIMD…
#24405: BLD, SIMD: The meson CPU dispatcher implementation
#24406: MAINT: Remove versioneer
#24409: REL: Prepare for the NumPy 1.26.0b1 release.
#24453: MAINT: Pin upper version of sphinx.
#24455: ENH: Add prefix to _ALIGN Macro
#24456: BUG: cleanup warnings [skip azp][skip circle][skip travis][skip…
#24460: MAINT: Upgrade to spin 0.5
asv devhas been removed, use
#24496: BUG: Fix meson build failure due to unchanged inplace auto-generated…
#24521: BUG: fix issue with git-version script, needs a shebang to run
#24522: BUG: Use a default assignment for git_hash [skip ci]
#24524: BUG: fix NPY_cast_info error handling in choose
#24526: BUG: Fix common block handling in f2py
#24541: CI,TYP: Bump mypy to 1.4.1
#24542: BUG: Fix assumed length f2py regression
#24544: MAINT: Harmonize fortranobject
#24545: TYP: add kind argument to numpy.isin type specification
#24561: BUG: fix comparisons between masked and unmasked structured arrays
#24590: CI: Exclude import libraries from list of DLLs on Cygwin.
#24591: BLD: fix
#24594: MAINT: Stop testing on ppc64le.
#24602: BLD: meson-cpu: fix SIMD support on platforms with no features
#24606: BUG: Change Cython
bindingdirective to “False”.
#24613: ENH: Adopt new macOS Accelerate BLAS/LAPACK Interfaces, including…
#24614: DOC: Update building docs to use Meson
#24615: TYP: Add the missing
#24616: TST: convert cython test from setup.py to meson
#24617: MAINT: Fixup
#24622: BUG, ENH: Fix
iso_c_bindingtype maps and fix
#24629: TYP: Allow
binary_reprto accept any object implementing…
#24630: TYP: Explicitly declare
#24638: MAINT: Bump actions/checkout from 3.6.0 to 4.0.0
#24648: MAINT: Refactor partial load Workaround for Clang
#24653: REL: Prepare for the NumPy 1.26.0rc1 release.
#24659: BLD: allow specifying the long double format to avoid the runtime…
#24665: BLD: fix bug in random.mtrand extension, don’t link libnpyrandom
#24675: BLD: build wheels for 32-bit Python on Windows, using MSVC
#24700: BLD: fix issue with compiler selection during cross compilation
#24701: BUG: Fix data stmt handling for complex values in f2py
#24707: TYP: Add annotations for the py3.12 buffer protocol
#24718: DOC: fix a few doc build issues on 1.26.x and update spin docs…