Merge pull request #1132 from IntelPython/main

Merge into GOLD

Author: Diptorup Deb

.github/workflows/conda-package.yml

@@ -26,7 +26,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.9', '3.10', '3.11']
         os: [ubuntu-latest, windows-latest]
 
     runs-on: ${{ matrix.os }}
@@ -93,7 +93,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.9', '3.10', '3.11']
         os: [ubuntu-20.04, ubuntu-latest, windows-latest]
         experimental: [false]
 
@@ -157,8 +157,9 @@ jobs:
           echo "PACKAGE_VERSION=$PACKAGE_VERSION" >> $GITHUB_ENV
 
       # We want to make sure that all dependecies install automatically.
+      # intel::intel-opencl-rt is needed for set-intel-ocl-icd-registry.ps1
       - name: Install builded package
-        run: mamba install ${{ env.PACKAGE_NAME }}=${{ env.PACKAGE_VERSION }} opencl_rt pytest -c ${{ env.CHANNEL_PATH }}
+        run: mamba install ${{ env.PACKAGE_NAME }}=${{ env.PACKAGE_VERSION }} intel::intel-opencl-rt pytest -c ${{ env.CHANNEL_PATH }}
 
       - name: Setup OpenCL CPU device
         if: runner.os == 'Windows'
@@ -199,7 +200,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.9', '3.10', '3.11']
         os: [ubuntu-latest, windows-latest]
 
     runs-on: ${{ matrix.os }}

.github/workflows/docker.yml

@@ -22,7 +22,7 @@ on:
       oneapi_version:
         default: '2023.1.0.46401'
       python_versions:
-        default: '["3.8", "3.9", "3.10", "3.11"]'
+        default: '["3.9", "3.10", "3.11"]'
       buildkit_version:
         default: '0.11.6'
 

CHANGELOG.md

@@ -4,10 +4,26 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.21.3] - 2023-09-21
+
+### Fixed
+* Mangled kernel name generation (#1112)
+
+### Added
+* Python 3.11 support (#1123)
+* Security policy (#1117)
+* scikit-build to build native extensions (#1107, #1127)
+
+### Changed
+* The data model used by the DpnpNdArray type for kernel functions(#1118)
+
+### Removed
+* Support for Python 3.8 (#1113)
+
 ## [0.21.2] - 2023-08-07
 
 ### Fixed
-* Bugs (#1068, #774) in atomic addition caused due to imporper floating point atomic emulation. (#1103)
+* Bugs (#1068, #774) in atomic addition caused due to improper floating point atomic emulation. (#1103)
 
 ### Changed
 * Updated documentation and user guides (#1097, #879)

CMakeLists.txt

@@ -0,0 +1,34 @@
+#[=======================================================================[.rst:
+numba_dpex
+-----------
+
+A cmake file to compile the ``_dpexrt_python`` Python C extension for
+``numba_dpex``. You can build this component locally in-place by invoking these
+commands:
+
+.. code-block:: cmake
+    ~$ cmake .
+    ~$ cmake --build . --verbose
+
+Once compiled, the _dpexrt_python library will be in ``numba_dpex/core/runtime``
+folder.
+
+This ``CMakeLists.txt`` file will be used by ``setup.py``.
+#]=======================================================================]
+
+cmake_minimum_required(VERSION 3.21...3.27 FATAL_ERROR)
+
+message(STATUS "NUMBA_DPEX_VERSION=" "${NUMBA_DPEX_VERSION}")
+
+project(numba-dpex
+    DESCRIPTION "An extension for Numba to add data-parallel offload capability"
+    VERSION ${NUMBA_DPEX_VERSION}
+)
+
+if(IS_INSTALL)
+    install(DIRECTORY numba_dpex
+        DESTINATION ${CMAKE_INSTALL_PREFIX}
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+add_subdirectory(numba_dpex)

MANIFEST.in

@@ -1,9 +1,6 @@
 include MANIFEST.in
 include README.md setup.py LICENSE
 
-recursive-include numba_dpex *.cl
-recursive-include numba_dpex *.spir
-
 include versioneer.py
 include numba_dpex/_version.py
 

SECURITY.md

@@ -0,0 +1,12 @@
+# Security Policy
+
+## Report a Vulnerability
+
+Please report security issues or vulnerabilities to the [Intel® Security Center].
+
+For more information on how Intel® works to resolve security issues, see
+[Vulnerability Handling Guidelines].
+
+[Intel® Security Center]:https://www.intel.com/content/www/us/en/security-center/default.html
+
+[Vulnerability Handling Guidelines]:https://www.intel.com/content/www/us/en/security-center/vulnerability-handling-guidelines.html

build-environment.yml

@@ -17,7 +17,9 @@ requirements:
     host:
         - python
         - setuptools >=63.*
-        - numba 0.57*
+        - scikit-build >=0.15*
+        - cmake >=3.26*
+        - numba >=0.57*
         - dpctl >=0.14*
         - dpnp >=0.11*
         - dpcpp-llvm-spirv

conda-recipe/meta.yaml

@@ -37,7 +37,7 @@ Building from source
 --------------------
 
 ``numba-dpex`` can be built from source using either ``conda-build`` or
-``setuptools``.
+``setuptools`` (with ``scikit-build`` backend).
 
 Steps to build using ``conda-build``:
 
@@ -70,7 +70,7 @@ first step.
 
     # Create a conda environment that hass needed dependencies installed
     conda create -n numba-dpex-env                                             \
-        dpctl dpnp numba dpcpp-llvm-spirv llvmdev pytest           \
+        scikit-build cmake dpctl dpnp numba dpcpp-llvm-spirv llvmdev pytest    \
         -c intel -c conda-forge
     # Activate the environment
     conda activate numba-dpex-env

docs/source/getting_started.rst

@@ -16,6 +16,8 @@ dependencies:
   - mkl >=2021.3.0  # for dpnp
   - dpcpp-llvm-spirv
   - packaging
+  - scikit-build >=0.15*
+  - cmake >=3.26*
   - pytest
   - pip
   - pip:

environment.yml

@@ -20,3 +20,5 @@ dependencies:
   - pytest-cov
   - pytest-xdist
   - pexpect
+  - scikit-build>=0.15*
+  - cmake>=3.26*

environment/coverage.yml

@@ -11,6 +11,8 @@ dependencies:
   - gxx_linux-64
   - dpcpp_linux-64
   - numba=0.57
+  - scikit-build>=0.15*
+  - cmake>=3.26*
   - dpctl
   - dpnp
   - dpcpp-llvm-spirv

environment/docs.yml

@@ -0,0 +1,48 @@
+if(IS_INSTALL)
+    install(DIRECTORY core
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+add_subdirectory(core/runtime)
+
+if(IS_INSTALL)
+    install(DIRECTORY dpctl_iface
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY dpnp_iface
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY examples
+        DESTINATION numba_dpex)
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY numba_patches
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY ocl
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY tests
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()
+
+if(IS_INSTALL)
+    install(DIRECTORY utils
+        DESTINATION numba_dpex
+        FILES_MATCHING PATTERN "*.py")
+endif()

numba_dpex/CMakeLists.txt

@@ -35,6 +35,7 @@ def _optimize_final_module(self):
         pmb.opt_level = config.OPT
 
         pmb.disable_unit_at_a_time = False
+        pmb.inlining_threshold = 2
         pmb.disable_unroll_loops = True
         pmb.loop_vectorize = False
         pmb.slp_vectorize = False

numba_dpex/core/codegen.py

@@ -57,6 +57,30 @@ def __init__(self, dmm, fe_type):
         ]
         super(USMArrayModel, self).__init__(dmm, fe_type, members)
 
+    @property
+    def flattened_field_count(self):
+        """Return the number of fields in an instance of a USMArrayModel."""
+        flattened_member_count = 0
+        members = self._members
+        for member in members:
+            if isinstance(member, types.UniTuple):
+                flattened_member_count += member.count
+            elif isinstance(
+                member,
+                (
+                    types.scalars.Integer,
+                    types.misc.PyObject,
+                    types.misc.RawPointer,
+                    types.misc.CPointer,
+                    types.misc.MemInfoPointer,
+                ),
+            ):
+                flattened_member_count += 1
+            else:
+                raise UnreachableError
+
+        return flattened_member_count
+
 
 class DpnpNdArrayModel(StructModel):
     """Data model for the DpnpNdArray type.
@@ -138,35 +162,54 @@ def __init__(self, dmm, fe_type):
         super(SyclQueueModel, self).__init__(dmm, fe_type, members)
 
 
-def _init_data_model_manager():
+def _init_data_model_manager() -> datamodel.DataModelManager:
+    """Initializes a DpexKernelTarget-specific data model manager.
+
+    SPIRV kernel functions for certain types of devices require an explicit
+    address space qualifier for pointers. For OpenCL HD Graphics
+    devices, defining a kernel function (spir_kernel calling convention) with
+    pointer arguments that have no address space qualifier causes a run time
+    crash. For this reason, numba-dpex defines two separate data
+    models: USMArrayModel and DpnpNdArrayModel. When a dpnp.ndarray object is
+    passed as an argument to a ``numba_dpex.kernel`` decorated function it uses
+    the USMArrayModel and when passed to a ``numba_dpex.dpjit`` decorated
+    function it uses the DpnpNdArrayModel. The difference is due to the fact
+    that inside a ``dpjit`` decorated function a dpnp.ndarray object can be
+    passed to any other regular function.
+
+    Returns:
+        DataModelManager: A numba-dpex DpexKernelTarget-specific data model
+        manager
+    """
     dmm = datamodel.default_manager.copy()
     dmm.register(types.CPointer, GenericPointerModel)
     dmm.register(Array, USMArrayModel)
+
+    # Register the USMNdArray type to USMArrayModel in numba_dpex's data model
+    # manager. The dpex_data_model_manager is used by the DpexKernelTarget
+    dmm.register(USMNdArray, USMArrayModel)
+
+    # Register the DpnpNdArray type to USMArrayModel in numba_dpex's data model
+    # manager. The dpex_data_model_manager is used by the DpexKernelTarget
+    dmm.register(DpnpNdArray, USMArrayModel)
+
+    # Register the DpctlSyclQueue type to SyclQueueModel in numba_dpex's data
+    # model manager. The dpex_data_model_manager is used by the DpexKernelTarget
+    dmm.register(DpctlSyclQueue, SyclQueueModel)
+
     return dmm
 
 
 dpex_data_model_manager = _init_data_model_manager()
 
-# XXX A kernel function has the spir_kernel ABI and requires pointers to have an
-# address space attribute. For this reason, the UsmNdArray type uses dpex's
-# ArrayModel where the pointers are address space casted to have a SYCL-specific
-# address space value. The DpnpNdArray type can be used inside djit functions
-# as host function calls arguments, such as dpnp library calls. The DpnpNdArray
-# needs to use Numba's array model as its data model. Thus, from a Numba typing
-# perspective dpnp.ndarrays cannot be directly passed to a kernel. To get
-# around the limitation, the DpexKernelTypingContext does not resolve the type
-# of dpnp.array args to a kernel as DpnpNdArray type objects, but uses the
-# ``to_usm_ndarray`` utility function to convert them into a UsmNdArray type
-# object.
-
-# Register the USMNdArray type with the dpex ArrayModel
+
+# Register the USMNdArray type to USMArrayModel in numba's default data model
+# manager
 register_model(USMNdArray)(USMArrayModel)
-dpex_data_model_manager.register(USMNdArray, USMArrayModel)
 
-# Register the DpnpNdArray type with the Numba ArrayModel
+# Register the DpnpNdArray type to DpnpNdArrayModel in numba's default data
+# model manager
 register_model(DpnpNdArray)(DpnpNdArrayModel)
-dpex_data_model_manager.register(DpnpNdArray, DpnpNdArrayModel)
 
 # Register the DpctlSyclQueue type
 register_model(DpctlSyclQueue)(SyclQueueModel)
-dpex_data_model_manager.register(DpctlSyclQueue, SyclQueueModel)

numba_dpex/core/datamodel/models.py

@@ -136,6 +136,7 @@ def compile(
         kernel = cres.target_context.prepare_ocl_kernel(
             func, cres.signature.args
         )
+        cres.library._optimize_final_module()
         self._llvm_module = kernel.module.__str__()
         self._module_name = kernel.name
 

numba_dpex/core/kernel_interface/spirv_kernel.py

@@ -70,30 +70,6 @@ def _compile_kernel_parfor(
         func_ir, kernel_name
     )
 
-    # A cast from DpnpNdArray type to USMNdArray is needed for all arguments of
-    # DpnpNdArray type. Although, DpnpNdArray derives from USMNdArray the two
-    # types use different data models. USMNdArray uses the
-    # numba_dpex.core.datamodel.models.ArrayModel data model that defines all
-    # CPointer type members in the GLOBAL address space. The DpnpNdArray uses
-    # Numba's default ArrayModel that does not define pointers in any specific
-    # address space. For OpenCL HD Graphics devices, defining a kernel function
-    # (spir_kernel calling convention) with pointer arguments that have no
-    # address space qualifier causes a run time crash. By casting the argument
-    # type for parfor arguments from DpnpNdArray type to the USMNdArray type the
-    # generated kernel always has an address space qualifier, avoiding the issue
-    # on OpenCL HD graphics devices.
-
-    for i, argty in enumerate(argtypes):
-        if isinstance(argty, DpnpNdArray):
-            new_argty = USMNdArray(
-                ndim=argty.ndim,
-                layout=argty.layout,
-                dtype=argty.dtype,
-                usm_type=argty.usm_type,
-                queue=argty.queue,
-            )
-            argtypes[i] = new_argty
-
     # compile the kernel
     kernel.compile(
         args=argtypes,