Merge remote-tracking branch 'numba/main' into cuda-dispatcher-base-2…

…0220204

.github/ISSUE_TEMPLATE/first_rc_checklist.md

@@ -22,6 +22,7 @@ labels: task
 * [ ] Build and upload conda packages on buildfarm (check "upload").
 * [ ] Build wheels (`$PYTHON_VERSIONS`) on the buildfarm.
 * [ ] Verify packages uploaded to Anaconda Cloud and move to `numba/label/main`.
+* [ ] Build sdist locally using `python setup.py sdist --owner=ci --group=numba` with umask `0022`.
 * [ ] Upload wheels and sdist to PyPI (upload from `ci_artifacts`).
 * [ ] Verify wheels for all platforms arrived on PyPi.
 * [ ] Initialize and verify ReadTheDocs build.

.github/ISSUE_TEMPLATE/sub_rc_checklist.md

@@ -19,6 +19,7 @@ labels: task
 * [ ] Verify packages uploaded to Anaconda Cloud and move to
   `numba/label/main`.
 * [ ] Build wheels (`$PYTHON_VERSIONS`) on the buildfarm.
+* [ ] Build sdist locally using `python setup.py sdist --owner=ci --group=numba` with umask `0022`.
 * [ ] Upload wheels and sdist to PyPI (upload from `ci_artifacts`).
 * [ ] Verify wheels for all platforms arrived on PyPi.
 * [ ] Verify ReadTheDocs build.

README.rst

@@ -13,11 +13,15 @@ Numba
 .. image:: https://zenodo.org/badge/3659275.svg
    :target: https://zenodo.org/badge/latestdoi/3659275
    :alt: Zenodo DOI
-   
+
 .. image:: https://img.shields.io/pypi/v/numba.svg
    :target: https://pypi.python.org/pypi/numba/
    :alt: PyPI
 
+.. image:: https://dev.azure.com/numba/numba/_apis/build/status/numba.numba?branchName=main
+    :target: https://dev.azure.com/numba/numba/_build/latest?definitionId=1?branchName=main
+    :alt: Azure Pipelines
+
 A Just-In-Time Compiler for Numerical Functions in Python
 #########################################################
 
@@ -31,54 +35,22 @@ parallelization of loops, generation of GPU-accelerated code, and creation of
 ufuncs and C callbacks.
 
 For more information about Numba, see the Numba homepage:
-https://numba.pydata.org
-
-Supported Platforms
-===================
-
-* Operating systems and CPUs:
-
-  - Linux: x86 (32-bit), x86_64, ppc64le (POWER8 and 9), ARMv7 (32-bit),
-    ARMv8 (64-bit).
-  - Windows: x86, x86_64.
-  - macOS: x86_64, (M1/Arm64, unofficial support only).
-  - \*BSD: (unofficial support only).
-
-* (Optional) Accelerators and GPUs:
-
-  * NVIDIA GPUs (Kepler architecture or later) via CUDA driver on Linux and
-    Windows.
+https://numba.pydata.org and the online documentation:
+https://numba.readthedocs.io/en/stable/index.html
 
-Dependencies
+Installation
 ============
 
-* Python versions: 3.7-3.10
-* llvmlite 0.39.*
-* NumPy >=1.18 (can build with 1.11 for ABI compatibility).
-
-Optionally:
-
-* SciPy >=1.0.0 (for ``numpy.linalg`` support).
+Please follow the instuctions:
 
-
-Installing
-==========
-
-The easiest way to install Numba and get updates is by using the Anaconda
-Distribution: https://www.anaconda.com/download
-
-::
-
-   $ conda install numba
-
-For more options, see the Installation Guide:
 https://numba.readthedocs.io/en/stable/user/installing.html
 
-Documentation
-=============
+Demo
+====
 
-https://numba.readthedocs.io/en/stable/index.html
+Please have a look and the demo notebooks via the mybinder service:
 
+https://mybinder.org/v2/gh/numba/numba-examples/master?filepath=notebooks
 
 Contact
 =======
@@ -87,11 +59,3 @@ Numba has a discourse forum for discussions:
 
 * https://numba.discourse.group
 
-
-
-Continuous Integration
-======================
-
-.. image:: https://dev.azure.com/numba/numba/_apis/build/status/numba.numba?branchName=main
-    :target: https://dev.azure.com/numba/numba/_build/latest?definitionId=1?branchName=main
-    :alt: Azure Pipelines

buildscripts/condarecipe.local/meta.yaml

@@ -27,9 +27,8 @@ requirements:
   build:
     - {{ compiler('c') }}      # [not (armv6l or armv7l or aarch64)]
     - {{ compiler('cxx') }}    # [not (armv6l or armv7l or aarch64)]
-    # both of these are needed on osx, llvm for the headers, Intel for the lib
+    # OpenMP headers from llvm needed for OSX.
     - llvm-openmp              # [osx]
-    - intel-openmp             # [osx]
   host:
     - python
     - numpy
@@ -71,7 +70,7 @@ test:
     - ipython                  # [not (armv6l or armv7l or aarch64)]
     - setuptools
     - tbb  >=2021              # [not (armv6l or armv7l or aarch64 or linux32 or ppc64le)]
-    - intel-openmp             # [osx]
+    - llvm-openmp              # [osx]
     # This is for driving gdb tests
     - pexpect                  # [linux64]
     # For testing ipython

buildscripts/incremental/setup_conda_environment.sh

@@ -81,8 +81,9 @@ if [[ $(uname) == Linux ]]; then
     fi
 elif  [[ $(uname) == Darwin ]]; then
     $CONDA_INSTALL clang_osx-64 clangxx_osx-64
-    # Install llvm-openmp and intel-openmp on OSX too
-    $CONDA_INSTALL llvm-openmp intel-openmp
+    # Install llvm-openmp on OSX for headers during build and runtime during
+    # testing
+    $CONDA_INSTALL llvm-openmp
 fi
 
 # `pip install` all the dependencies on Python 3.10

docs/source/cuda/cudapysupported.rst

@@ -65,6 +65,7 @@ The following built-in types support are inherited from CPU nopython mode.
 * bool
 * None
 * tuple
+* Enum, IntEnum
 
 See :ref:`nopython built-in types <pysupported-builtin-types>`.
 

docs/source/developer/contributing.rst

@@ -99,8 +99,7 @@ Then create an environment with the right dependencies::
 .. note::
    This installs an environment based on Python 3.8, but you can of course
    choose another version supported by Numba.  To test additional features,
-   you may also need to install ``tbb`` and/or ``llvm-openmp`` and
-   ``intel-openmp``.
+   you may also need to install ``tbb`` and/or ``llvm-openmp``.
 
 To activate the environment for the current shell session::
 

docs/source/user/installing.rst

@@ -167,9 +167,9 @@ otherwise build by default along with information on configuration options.
   * For Linux and Windows it is necessary to provide OpenMP C headers and
     runtime  libraries compatible with the compiler tool chain mentioned above,
     and for these to be accessible to the compiler via standard flags.
-  * For OSX the conda packages ``llvm-openmp`` and ``intel-openmp`` provide
-    suitable C headers and libraries. If the compilation requirements are not
-    met the OpenMP threading backend will not be compiled
+  * For OSX the conda package ``llvm-openmp`` provides suitable C headers and
+    libraries. If the compilation requirements are not met the OpenMP threading
+    backend will not be compiled.
 
 .. envvar:: NUMBA_DISABLE_TBB (default: not set)
 
@@ -211,8 +211,6 @@ vary with target operating system and hardware. The following lists them all
 
   * ``llvm-openmp`` (OSX) - provides headers for compiling OpenMP support into
     Numba's threading backend
-  * ``intel-openmp`` (OSX) - provides OpenMP library support for Numba's
-    threading backend.
   * ``tbb-devel`` - provides TBB headers/libraries for compiling TBB support
     into Numba's threading backend (version >= 2021 required).
   * ``importlib_metadata`` (for Python versions < 3.9)
@@ -226,8 +224,10 @@ vary with target operating system and hardware. The following lists them all
   * ``jinja2`` - for "pretty" type annotation output (HTML) via the ``numba``
     CLI
   * ``cffi`` - permits use of CFFI bindings in Numba compiled functions
-  * ``intel-openmp`` - (OSX) provides OpenMP library support for Numba's OpenMP
-    threading backend
+  * ``llvm-openmp`` - (OSX) provides OpenMP library support for Numba's OpenMP
+    threading backend.
+  * ``intel-openmp`` - (OSX) provides an alternative OpenMP library for use with
+    Numba's OpenMP threading backend.
   * ``ipython`` - if in use, caching will use IPython's cache
     directories/caching still works
   * ``pyyaml`` - permits the use of a ``.numba_config.yaml``

docs/source/user/threading-layer.rst

@@ -157,8 +157,9 @@ and requirements are as follows:
 |                      | Windows   | MS OpenMP libraries (very likely this will|
 |                      |           | already exist)                            |
 |                      |           |                                           |
-|                      | OSX       | The ``intel-openmp`` package (``$ conda   |
-|                      |           | install intel-openmp``)                   |
+|                      | OSX       | Either the ``intel-openmp`` package or the|
+|                      |           | ``llvm-openmp`` package                   |
+|                      |           | (``conda install`` the package as named). |
 +----------------------+-----------+-------------------------------------------+
 | ``workqueue``        | All       | None                                      |
 +----------------------+-----------+-------------------------------------------+

numba/cuda/cudaimpl.py

@@ -7,6 +7,7 @@
 
 from numba.core.imputils import Registry, lower_cast
 from numba.core.typing.npydecl import parse_dtype, signature
+from numba.core.datamodel import models
 from numba.core import types, cgutils
 from .cudadrv import nvvm
 from numba import cuda
@@ -987,7 +988,11 @@ def _generic_array(context, builder, shape, dtype, symbol_name, addrspace,
         raise ValueError("array length <= 0")
 
     # Check that we support the requested dtype
-    other_supported_type = isinstance(dtype, (types.Record, types.Boolean))
+    data_model = context.data_model_manager[dtype]
+    other_supported_type = (
+        isinstance(dtype, (types.Record, types.Boolean))
+        or isinstance(data_model, models.StructModel)
+    )
     if dtype not in types.number_domain and not other_supported_type:
         raise TypeError("unsupported type: %s" % dtype)
 

numba/cuda/dispatcher.py

@@ -384,6 +384,14 @@ def _prepare_args(self, ty, val, stream, retr, kernelargs):
             for t, v in zip(ty, val):
                 self._prepare_args(t, v, stream, retr, kernelargs)
 
+        elif isinstance(ty, types.EnumMember):
+            try:
+                self._prepare_args(
+                    ty.dtype, val.value, stream, retr, kernelargs
+                )
+            except NotImplementedError:
+                raise NotImplementedError(ty, val)
+
         else:
             raise NotImplementedError(ty, val)
 

numba/cuda/target.py

@@ -20,11 +20,13 @@
 class CUDATypingContext(typing.BaseContext):
     def load_additional_registries(self):
         from . import cudadecl, cudamath, libdevicedecl
+        from numba.core.typing import enumdecl
 
         self.install_registry(cudadecl.registry)
         self.install_registry(cudamath.registry)
         self.install_registry(cmathdecl.registry)
         self.install_registry(libdevicedecl.registry)
+        self.install_registry(enumdecl.registry)
 
     def resolve_value_type(self, val):
         # treat other dispatcher object as another device function
@@ -89,7 +91,7 @@ def load_additional_registries(self):
         # side effect of import needed for numba.cpython.*, the builtins
         # registry is updated at import time.
         from numba.cpython import numbers, tupleobj, slicing # noqa: F401
-        from numba.cpython import rangeobj, iterators # noqa: F401
+        from numba.cpython import rangeobj, iterators, enumimpl # noqa: F401
         from numba.cpython import unicode, charseq # noqa: F401
         from numba.cpython import cmathimpl
         from numba.np import arrayobj # noqa: F401

numba/cuda/tests/cudapy/extensions_usecases.py

@@ -0,0 +1,58 @@
+from numba import types
+from numba.core import config
+
+
+class TestStruct:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+
+
+class TestStructModelType(types.Type):
+    def __init__(self):
+        super().__init__(name="TestStructModelType")
+
+
+test_struct_model_type = TestStructModelType()
+
+
+if not config.ENABLE_CUDASIM:
+    from numba import int32
+    from numba.core.extending import (
+        models,
+        register_model,
+        make_attribute_wrapper,
+        typeof_impl,
+        type_callable
+    )
+    from numba.cuda.cudaimpl import lower
+    from numba.core import cgutils
+
+    @typeof_impl.register(TestStruct)
+    def typeof_teststruct(val, c):
+        return test_struct_model_type
+
+    @register_model(TestStructModelType)
+    class TestStructModel(models.StructModel):
+        def __init__(self, dmm, fe_type):
+            members = [("x", int32), ("y", int32)]
+            super().__init__(dmm, fe_type, members)
+
+    make_attribute_wrapper(TestStructModelType, 'x', 'x')
+    make_attribute_wrapper(TestStructModelType, 'y', 'y')
+
+    @type_callable(TestStruct)
+    def type_test_struct(context):
+        def typer(x, y):
+            if isinstance(x, types.Integer) and isinstance(y, types.Integer):
+                return test_struct_model_type
+        return typer
+
+    @lower(TestStruct, types.Integer, types.Integer)
+    def lower_test_type_ctor(context, builder, sig, args):
+        obj = cgutils.create_struct_proxy(
+            test_struct_model_type
+        )(context, builder)
+        obj.x = args[0]
+        obj.y = args[1]
+        return obj._getvalue()