Merge remote-tracking branch 'upstream/main' into guilhermeleobas/arr…

…_ravel

.flake8

@@ -170,7 +170,6 @@ exclude =
     numba/tests/test_enums.py
     numba/tests/test_profiler.py
     numba/tests/test_numpyadapt.py
-    numba/tests/test_stencils.py
     numba/tests/cache_usecases.py
     numba/tests/true_div_usecase.py
     numba/tests/test_dataflow.py

.github/ISSUE_TEMPLATE/first_rc_checklist.md

@@ -17,10 +17,16 @@ labels: task
 * [ ] Merge change log changes.
     - [ ] "PR with changelog entries".
 * [ ] Create X.Y release branch.
+* [ ] Create PR against the release branch to make `numba/testing/main.py`
+      to refer to `origin/releaseX.Y` instead of `origin/main`.
 * [ ] Dependency version pinning on release branch
   * [ ] Pin llvmlite to `>=0.A.0rc1,<0.A+1.0`.
   * [ ] Pin NumPy if needed
   * [ ] Pin TBB if needed
+* [ ] Run the HEAD of the release branch through the build farm and confirm:
+  * [ ] Build farm CPU testing has passed.
+  * [ ] Build farm CUDA testing has passed.
+  * [ ] Build farm wheel testing has passed.
 * [ ] Annotated tag `X.Y.Zrc1` on release branch (no `v` prefix).
 * [ ] Build and upload conda packages on buildfarm (check "upload").
 * [ ] Build wheels and sdist on the buildfarm (check "upload").

.github/ISSUE_TEMPLATE/sub_rc_checklist.md

@@ -20,6 +20,10 @@ labels: task
 * [ ] Merge change log modifications and cherry-picks to X.Y release branch.
   * [ ] https://github.com/numba/numba/pull/XXXX
 * [ ] Review, merge and check execution of release notebook. (FINAL ONLY)
+* [ ] Run the HEAD of the release branch through the build farm and confirm:
+  * [ ] Build farm CPU testing has passed.
+  * [ ] Build farm CUDA testing has passed
+  * [ ] Build farm wheel testing has passed
 * [ ] Annotated tag X.Y.Z on release branch (no `v` prefix).
 * [ ] Build and upload conda packages on buildfarm (check `upload`).
 * [ ] Build wheels and sdist on the buildfarm (check "upload").

CHANGE_LOG

@@ -1,3 +1,43 @@
+Version 0.56.4 (3 November, 2022)
+---------------------------------
+
+This is a bugfix release to fix a regression in the CUDA target in relation to
+the ``.view()`` method on CUDA device arrays that is present when using NumPy
+version 1.23.0 or later.
+
+Pull-Requests:
+
+* PR `#8537 <https://github.com/numba/numba/pull/8537>`_: Make ol_compatible_view accessible on all targets (`gmarkall <https://github.com/gmarkall>`_)
+* PR `#8552 <https://github.com/numba/numba/pull/8552>`_: Update version support table for 0.56.4. (`stuartarchibald <https://github.com/stuartarchibald>`_)
+* PR `#8553 <https://github.com/numba/numba/pull/8553>`_: Update CHANGE_LOG for 0.56.4 (`stuartarchibald <https://github.com/stuartarchibald>`_)
+* PR `#8570 <https://github.com/numba/numba/pull/8570>`_: Release 0.56 branch: Fix overloads with ``target="generic"`` for CUDA (`gmarkall <https://github.com/gmarkall>`_)
+* PR `#8571 <https://github.com/numba/numba/pull/8571>`_: Additional update to CHANGE_LOG for 0.56.4 (`stuartarchibald <https://github.com/stuartarchibald>`_)
+
+Authors:
+
+* `gmarkall <https://github.com/gmarkall>`_
+* `stuartarchibald <https://github.com/stuartarchibald>`_
+
+Version 0.56.3 (13 October, 2022)
+---------------------------------
+
+This is a bugfix release to remove the version restriction applied to the
+``setuptools`` package and to fix a bug in the CUDA target in relation to
+copying zero length device arrays to zero length host arrays.
+
+Pull-Requests:
+
+* PR `#8475 <https://github.com/numba/numba/pull/8475>`_: Remove setuptools version pin (`gmarkall <https://github.com/gmarkall>`_)
+* PR `#8482 <https://github.com/numba/numba/pull/8482>`_: Fix #8477: Allow copies with different strides for 0-length data (`gmarkall <https://github.com/gmarkall>`_)
+* PR `#8486 <https://github.com/numba/numba/pull/8486>`_: Restrict the TBB development package to supported version in Azure. (`stuartarchibald <https://github.com/stuartarchibald>`_)
+* PR `#8503 <https://github.com/numba/numba/pull/8503>`_: Update version support table for 0.56.3 (`stuartarchibald <https://github.com/stuartarchibald>`_)
+* PR `#8504 <https://github.com/numba/numba/pull/8504>`_: Update CHANGE_LOG for 0.56.3 (`stuartarchibald <https://github.com/stuartarchibald>`_)
+
+Authors:
+
+* `gmarkall <https://github.com/gmarkall>`_
+* `stuartarchibald <https://github.com/stuartarchibald>`_
+
 Version 0.56.2 (1 September, 2022)
 ----------------------------------
 

buildscripts/gpuci/build.sh

@@ -24,6 +24,13 @@ else
   export NUMBA_CUDA_USE_NVIDIA_BINDING=0;
 fi;
 
+# Test with different NumPy versions with each toolkit (it's not worth testing
+# the Cartesian product of versions here, we just need to test with different
+# CUDA and NumPy versions).
+declare -A CTK_NUMPY_VMAP=( ["11.0"]="1.19" ["11.1"]="1.21" ["11.2"]="1.22" ["11.5"]="1.23")
+NUMPY_VER="${CTK_NUMPY_VMAP[$CUDA_TOOLKIT_VER]}"
+
+
 ################################################################################
 # SETUP - Check environment
 ################################################################################
@@ -40,7 +47,7 @@ gpuci_mamba_retry create -n numba_ci -y \
                   "python=${PYTHON_VER}" \
                   "cudatoolkit=${CUDA_TOOLKIT_VER}" \
                   "numba/label/dev::llvmlite" \
-                  "numpy=1.21" \
+                  "numpy=${NUMPY_VER}" \
                   "scipy" \
                   "cffi" \
                   "psutil" \

buildscripts/incremental/setup_conda_environment.cmd

@@ -5,7 +5,6 @@ set CONDA_CONFIG=cmd /C conda config
 %CONDA_CONFIG% --set remote_connect_timeout_secs 30.15
 %CONDA_CONFIG% --set remote_max_retries 10
 %CONDA_CONFIG% --set remote_read_timeout_secs 120.2
-%CONDA_CONFIG% --set restore_free_channel true
 %CONDA_CONFIG% --set show_channel_urls true
 cmd /C conda info
 %CONDA_CONFIG% --show
@@ -22,11 +21,9 @@ call deactivate
 @rem Display root environment (for debugging)
 conda list
 @rem Scipy, CFFI, jinja2 and IPython are optional dependencies, but exercised in the test suite
-conda create -n %CONDA_ENV% -q -y python=%PYTHON% numpy=%NUMPY% cffi pip jinja2 ipython gitpython pyyaml
+conda create -n %CONDA_ENV% -q -y python=%PYTHON% numpy=%NUMPY% cffi pip scipy jinja2 ipython gitpython pyyaml
 
 call activate %CONDA_ENV%
-@rem Scipy comes from conda-forge for NumPy 1.23
-if %NUMPY% == "1.23" (%CONDA_INSTALL% conda-forge::scipy) else (%CONDA_INSTALL% scipy)
 @rem Install latest llvmlite build
 %CONDA_INSTALL% -c numba/label/dev llvmlite=0.40
 @rem Install required backports for older Pythons

buildscripts/incremental/setup_conda_environment.sh

@@ -48,12 +48,7 @@ if [ "${VANILLA_INSTALL}" != "yes" ]; then
     # dependencies, but exercised in the test suite
     # pexpect is used to run the gdb tests.
     # ipykernel is used for testing ipython behaviours.
-    $CONDA_INSTALL ${EXTRA_CHANNELS} cffi jinja2 ipython ipykernel pygments pexpect
-    if [[ "$NUMPY" == "1.23" ]] ; then
-        $CONDA_INSTALL ${EXTRA_CHANNELS} conda-forge::scipy
-    else
-        $CONDA_INSTALL ${EXTRA_CHANNELS} scipy
-    fi
+    $CONDA_INSTALL ${EXTRA_CHANNELS} cffi jinja2 ipython ipykernel scipy pygments pexpect
 fi
 
 # Install the compiler toolchain and gdb (if available)
@@ -73,7 +68,7 @@ $CONDA_INSTALL -c numba/label/dev llvmlite=0.40
 if [ $PYTHON \< "3.9" ]; then $CONDA_INSTALL importlib_metadata; fi
 
 # Install dependencies for building the documentation
-if [ "$BUILD_DOC" == "yes" ]; then $CONDA_INSTALL sphinx=2.4.4 docutils=0.17 sphinx_rtd_theme pygments numpydoc; fi
+if [ "$BUILD_DOC" == "yes" ]; then $CONDA_INSTALL sphinx docutils sphinx_rtd_theme pygments numpydoc; fi
 if [ "$BUILD_DOC" == "yes" ]; then $PIP_INSTALL rstcheck; fi
 # Install dependencies for code coverage (codecov.io)
 if [ "$RUN_COVERAGE" == "yes" ]; then $PIP_INSTALL codecov; fi

docs/source/cuda-reference/kernel.rst

@@ -59,7 +59,9 @@ creating a specialized instance:
 .. autoclass:: numba.cuda.dispatcher.CUDADispatcher
    :members: inspect_asm, inspect_llvm, inspect_sass, inspect_types,
              get_regs_per_thread, specialize, specialized, extensions, forall,
-             get_shared_mem_per_block, get_const_mem_size
+             get_shared_mem_per_block, get_const_mem_size,
+             get_local_mem_per_thread
+
 
 
 Intrinsic Attributes and Functions

docs/source/extending/low-level.rst

@@ -59,13 +59,11 @@ and typing *operations* (or *functions*) on known value types.
    existing default (e.g. to treat ``float`` as ``numba.float32`` instead of
    ``numba.float64``).
 
-.. decorator:: as_numba_type.register
-
-   Register the decorated function as a type inference function used by
-   ``as_numba_type`` when trying to infer the Numba type of a Python type.
-   The decorated function is called with a single *py_type* argument
-   and returns either a corresponding Numba type, or None if it cannot infer
-   that *py_type*.
+   This function can also be used as a decorator. It registers the decorated
+   function as a type inference function used by ``as_numba_type`` when trying
+   to infer the Numba type of a Python type. The decorated function is called
+   with a single *py_type* argument and returns either a corresponding Numba
+   type, or ``None`` if it cannot infer that *py_type*.
 
 
 Lowering

docs/source/reference/numpysupported.rst

@@ -391,6 +391,8 @@ Reductions
 The following reduction functions are supported:
 
 * :func:`numpy.diff` (only the 2 first arguments)
+* :func:`numpy.amin` (only the first argument, also aliased as np.min)
+* :func:`numpy.amax` (only the first argument, also aliased as np.max)
 * :func:`numpy.median` (only the first argument)
 * :func:`numpy.nancumprod` (only the first argument)
 * :func:`numpy.nancumsum` (only the first argument)
@@ -479,6 +481,7 @@ The following top-level functions are supported:
 * :func:`numpy.histogram` (only the 3 first arguments)
 * :func:`numpy.hstack`
 * :func:`numpy.identity`
+* :func:`numpy.isclose`
 * :func:`numpy.kaiser`
 * :func:`numpy.iscomplex`
 * :func:`numpy.iscomplexobj`

docs/source/reference/pysupported.rst

@@ -1238,29 +1238,29 @@ Third-party modules
 Similarly to ctypes, Numba is able to call into `cffi`_-declared external
 functions, using the following C types and any derived pointer types:
 
-* :c:type:`char`
-* :c:type:`short`
-* :c:type:`int`
-* :c:type:`long`
-* :c:type:`long long`
-* :c:type:`unsigned char`
-* :c:type:`unsigned short`
-* :c:type:`unsigned int`
-* :c:type:`unsigned long`
-* :c:type:`unsigned long long`
-* :c:type:`int8_t`
-* :c:type:`uint8_t`
-* :c:type:`int16_t`
-* :c:type:`uint16_t`
-* :c:type:`int32_t`
-* :c:type:`uint32_t`
-* :c:type:`int64_t`
-* :c:type:`uint64_t`
-* :c:type:`float`
-* :c:type:`double`
-* :c:type:`ssize_t`
-* :c:type:`size_t`
-* :c:type:`void`
+* :c:expr:`char`
+* :c:expr:`short`
+* :c:expr:`int`
+* :c:expr:`long`
+* :c:expr:`long long`
+* :c:expr:`unsigned char`
+* :c:expr:`unsigned short`
+* :c:expr:`unsigned int`
+* :c:expr:`unsigned long`
+* :c:expr:`unsigned long long`
+* :c:expr:`int8_t`
+* :c:expr:`uint8_t`
+* :c:expr:`int16_t`
+* :c:expr:`uint16_t`
+* :c:expr:`int32_t`
+* :c:expr:`uint32_t`
+* :c:expr:`int64_t`
+* :c:expr:`uint64_t`
+* :c:expr:`float`
+* :c:expr:`double`
+* :c:expr:`ssize_t`
+* :c:expr:`size_t`
+* :c:expr:`void`
 
 The ``from_buffer()`` method of ``cffi.FFI`` and ``CompiledFFI`` objects is
 supported for passing Numpy arrays and other buffer-like objects.  Only

docs/source/user/installing.rst

@@ -261,6 +261,10 @@ information.
 +===========+==============+===========================+============================+==============================+===================+=============================+
 | 0.57.x    | TBC          | 3.8.x <= version < 3.12   | 1.19 <= version < 1.24     | 0.40.x                       | 11.x              | 2021.x                      |
 +-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
+| 0.56.4    | 2022-11-03   | 3.7.x <= version < 3.11   | 1.18 <= version < 1.24     | 0.39.x                       | 11.x              | 2021.x                      |
++-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
+| 0.56.3    | 2022-10-13   | 3.7.x <= version < 3.11   | 1.18 <= version < 1.24     | 0.39.x                       | 11.x              | 2021.x                      |
++-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
 | 0.56.2    | 2022-09-01   | 3.7.x <= version < 3.11   | 1.18 <= version < 1.24     | 0.39.x                       | 11.x              | 2021.x                      |
 +-----------+--------------+---------------------------+----------------------------+------------------------------+-------------------+-----------------------------+
 | 0.56.1    | NO RELEASE   |                           |                            |                              |                   |                             |

docs/source/user/jitclass.rst

@@ -99,7 +99,8 @@ First, using explicit Numba types and explicit construction.
 
 .. code-block:: python
 
-    from numba import jitclass, types, typed
+    from numba import types, typed
+    from numba.experimental import jitclass
 
     # key and value types
     kv_ty = (types.int64, types.unicode_type)
@@ -131,7 +132,8 @@ example:
 
 .. code-block:: python
 
-    from numba import jitclass, typed, typeof
+    from numba import typed, typeof
+    from numba.experimental import jitclass
 
     d = typed.Dict()
     d[1] = "apple"
@@ -158,7 +160,8 @@ instance of the type specified.
 
 .. code-block:: python
 
-    from numba import jitclass, types
+    from numba import types
+    from numba.experimental import jitclass
 
     dict_ty = types.DictType(types.int64, types.unicode_type)
 

docs/source/user/pycc.rst

@@ -25,7 +25,7 @@ Benefits
 
 .. seealso::
    Compiled extension modules are discussed in the
-   `Python packaging user guide <https://packaging.python.org/en/latest/extensions/>`_.
+   `Python packaging user guide <https://packaging.python.org/en/latest/guides/packaging-binary-extensions/>`_.
 
 
 Limitations

docs/source/user/stencil.rst

@@ -213,8 +213,8 @@ rather than relative indexing::
 ``StencilFunc``
 ===============
 
-The stencil decorator returns a callable object of type ``StencilFunc``.
-``StencilFunc`` objects contains a number of attributes but the only one of
+The stencil decorator returns a callable object of type ``StencilFunc``. A
+``StencilFunc`` object contains a number of attributes but the only one of
 potential interest to users is the ``neighborhood`` attribute.
 If the ``neighborhood`` option was passed to the stencil decorator then
 the provided neighborhood is stored in this attribute.  Else, upon

docs/source/user/troubleshoot.rst

@@ -797,7 +797,7 @@ the line in which the access violation occurred is printed.
 
 Continuing the example as a debugging session demonstration, first ``index``
 can be printed, and it is evidently 1e9. Printing ``c`` shows that it is a
-structure, so the type needs looking up and it can be seen that is it an
+structure, so the type needs looking up and it can be seen that it is an
 ``array(float64, 1d, C)`` type. Given the segfault came from an invalid access
 it would be informative to check the number of items in the array and compare
 that to the index requested. Inspecting the ``nitems`` member of the structure
@@ -874,12 +874,13 @@ breakpoint was hit, and after a ``continue`` was issued, it broke again at line
 Debugging in parallel regions
 -----------------------------
 
-The follow example is quite involved, it executes with ``gdb`` instrumentation
-from the outset as per the example above, but it also uses threads and makes use
-of the breakpoint functionality. Further, the last iteration of the parallel
-section calls the function ``work``, which is actually just a binding to
-``glibc``'s ``free(3)`` in this case, but could equally be some involved
-function that is presenting a segfault for unknown reasons.
+The following example is quite involved, it executes with ``gdb``
+instrumentation from the outset as per the example above, but it also uses
+threads and makes use of the breakpoint functionality. Further, the last
+iteration of the parallel section calls the function ``work``, which is
+actually just a binding to ``glibc``'s ``free(3)`` in this case, but could
+equally be some involved function that is presenting a segfault for unknown
+reasons.
 
 .. code-block:: python
   :linenos:

numba/_numba_common.h

@@ -10,11 +10,15 @@
  * but are not exposed outside of a shared library or executable.
  * Note this is default behaviour for global symbols under Windows.
  */
-#if (__has_attribute(visibility) || \
-     (defined(__GNUC__) && __GNUC__ >= 4))
-#define VISIBILITY_HIDDEN __attribute__ ((visibility("hidden")))
+#if defined(_MSC_VER)
+    #define VISIBILITY_HIDDEN
+    #define VISIBILITY_GLOBAL __declspec(dllexport)
+#elif (__has_attribute(visibility) || (defined(__GNUC__) && __GNUC__ >= 4))
+    #define VISIBILITY_HIDDEN __attribute__ ((visibility("hidden")))
+    #define VISIBILITY_GLOBAL __attribute__ ((visibility("default")))
 #else
-#define VISIBILITY_HIDDEN
+    #define VISIBILITY_HIDDEN
+    #define VISIBILITY_GLOBAL
 #endif
 
 /*

numba/cext/cext.h

@@ -9,6 +9,9 @@
 #define NUMBA_EXPORT_FUNC(_rettype) VISIBILITY_HIDDEN _rettype
 #define NUMBA_EXPORT_DATA(_vartype) VISIBILITY_HIDDEN _vartype
 
+/* Use to declare a symbol as exported (global). */
+#define NUMBA_GLOBAL_FUNC(_rettype) VISIBILITY_GLOBAL _rettype
+
 NUMBA_EXPORT_FUNC(Py_ssize_t)
 aligned_size(Py_ssize_t sz);