Merge remote-tracking branch 'origin/master' into gold/2021

Author: oleksandr-pavlyk

.github/workflows/conda-package.yml

@@ -18,7 +18,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
     steps:
       - uses: actions/checkout@v3
         with:
@@ -63,7 +63,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
     env:
       conda-bld: C:\Miniconda\conda-bld\win-64\
     steps:
@@ -102,7 +102,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
         experimental: [false]
         runner: [ubuntu-20.04]
     continue-on-error: ${{ matrix.experimental }}
@@ -168,7 +168,7 @@ jobs:
         run: |
           . $CONDA/etc/profile.d/conda.sh
           conda activate test_dpctl
-          gdb --batch -ex r -ex 'info sharedlibrary' -ex 'set print elements 1000' -ex bt --args ${CONDA_PREFIX}/bin/python -m pytest -q -ra --disable-warnings --pyargs dpctl.tests.test_tensor_elementwise::test_cos_order -vv || true
+          gdb --batch -ex r -ex 'info sharedlibrary' -ex 'set print elements 1000' -ex bt --args ${CONDA_PREFIX}/bin/python -m pytest -q -ra --disable-warnings --pyargs dpctl.tests.elementwise.test_trigonometric::test_trig_order -vv || true
       - name: Run tests
         env:
           SYCL_QUEUE_THREAD_POOL_SIZE: 6
@@ -185,7 +185,7 @@ jobs:
         shell: cmd /C CALL {0}
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
         experimental: [false]
         runner: [windows-latest]
     continue-on-error: ${{ matrix.experimental }}
@@ -300,7 +300,7 @@ jobs:
     runs-on: ubuntu-20.04
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
     steps:
       - name: Download artifact
         uses: actions/download-artifact@v3
@@ -324,7 +324,7 @@ jobs:
     runs-on: windows-latest
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10']
+        python: ['3.8', '3.9', '3.10', '3.11']
     steps:
       - name: Download artifact
         uses: actions/download-artifact@v3

.github/workflows/generate-coverage.yaml

@@ -79,7 +79,7 @@ jobs:
       - name: Install dpctl dependencies
         shell: bash -l {0}
         run: |
-          pip install numpy cython"<3" setuptools pytest pytest-cov scikit-build cmake coverage[toml]
+          pip install numpy cython setuptools pytest pytest-cov scikit-build cmake coverage[toml]
 
       - name: Build dpctl with coverage
         shell: bash -l {0}

.github/workflows/generate-docs.yml

@@ -49,7 +49,7 @@ jobs:
         if: ${{ !github.event.pull_request || github.event.action != 'closed' }}
         shell: bash -l {0}
         run: |
-          pip install numpy cython"<3" setuptools scikit-build cmake sphinx sphinx_rtd_theme pydot graphviz sphinxcontrib-programoutput sphinxcontrib-googleanalytics
+          pip install numpy cython setuptools scikit-build cmake sphinx sphinx_rtd_theme pydot graphviz sphinxcontrib-programoutput sphinxcontrib-googleanalytics
       - name: Checkout repo
         uses: actions/checkout@v3
         with:

.github/workflows/os-llvm-sycl-build.yml

@@ -108,7 +108,7 @@ jobs:
       - name: Install dpctl dependencies
         shell: bash -l {0}
         run: |
-          pip install numpy cython"<3" setuptools pytest scikit-build cmake
+          pip install numpy cython setuptools pytest scikit-build cmake
 
       - name: Checkout repo
         uses: actions/checkout@v3

CHANGELOG.md

@@ -4,6 +4,100 @@ 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).
 
+## [dev]
+
+### Added
+### Changed
+
+* Removed `dpctl.tensor.numpy_usm_shared` obsolete class and associated tests which were being skipped
+
+### Fixed
+
+## [0.14.5] - 07/17/2023
+
+### Added
+
+* Added `dpctl.tensor.log2` and `dpctl.tensor.log10`: [#1267](https://github.com/IntelPython/dpctl/pull/1267)
+* Added `dpctl.tensor.negative`, `dpctl.tensor.positive`, `dpctl.tensor.square` [#1268](https://github.com/IntelPython/dpctl/pull/1268)
+* Added `dpctl.tensor.logical_not`, `dpctl.tensor.logical_and`, `dpctl.tensor.logical_or`, `dpctl.tensor.logical_xor` [#1270](https://github.com/IntelPython/dpctl/pull/1270)
+
+### Changed
+
+* `dpctl.tensor.astype` behavior for `newdtype=None` changes [#1261](https://github.com/IntelPython/dpctl/pull/1262)
+* `dpctl.tensor.usm_ndaray` constructor default value of `dtype` keyword argument changed to `None`: [#1265](https://github.com/IntelPython/dpctl/pull/1265)
+* Support for `out` arguments that overlap with inputs for unary elementwise functions[#1281](https://github.com/IntelPython/dpctl/pull/1281)
+* Copying from one array to another a no-op if both arrays view into the same memory [#1284](https://github.com/IntelPython/dpctl/pull/1284)
+
+## [0.14.4] - 06/14/2023
+
+### Added
+
+* Added `dpctl.tensor.less_equal`, `dpctl.tensor.greater`, `dpctl.tensor.greater_equal`: [#1239](https://github.com/IntelPython/dpctl/pull/1239)
+
+### Changed
+
+* Optimized in-place arithmetic operations for updating matrix with rows/columns via broadcasting: [#1244](https://github.com/IntelPython/dpctl/pull/1244)
+
+### Fixed
+
+* Fixed handling of 0d arrays in `dpctl.tensor.sum`: [#1238](https://github.com/IntelPython/dpctl/pull/1238)
+
+## [0.14.3] - 06/13/2023
+
+### Added
+
+* Added support of `axis=None` in `dpctl.tensor.concat` [#1125](https://github.com/IntelPython/dpctl/pull/1125)
+* Added caching for `dpctl.SyclDevice.filter_string` property [#1127](https://github.com/IntelPython/dpctl/pull/1127)
+* Added `dpctl.tensor.isdtype` from array API [#1133](https://github.com/IntelPython/dpctl/pull/1133)
+* Added `dpctl.tensor.unstack`, `dpctl.tensor.moveaxis`, `dpctl.tensor.swapaxes` [#1137](https://github.com/IntelPython/dpctl/pull/1137), [#1174](https://github.com/IntelPython/dpctl/pull/1174)
+* Allow for mutation of `dpctl.tensor.usm_ndarray.flags.writable` [#1141](https://github.com/IntelPython/dpctl/pull/1141)
+* Added `dpctl.tensor.where` from array API [#1147](https://github.com/IntelPython/dpctl/pull/1147)
+* Include libtensor headers in `dpctl` installation layout [#1185](https://github.com/IntelPython/dpctl/pull/1185)
+* Added new properties of `dpctl.tensor.usm_ndarray` object [#1199](https://github.com/IntelPython/dpctl/pull/1199)
+* Added a list of unary and binary elementwise functions from array API:
+   - [#1203](https://github.com/IntelPython/dpctl/pull/1203): `dpctl.tensor.add`, `dpctl.tensor.divide`, `dpctl.tensor.isnan`, `dpctl.tensor.isinf`, `dpctl.tensor.isfinite`, `dpctl.tensor.cos`, `dpctl.tensor.abs`, `dpctl.tensor.equal`
+   - [#1205](https://github.com/IntelPython/dpctl/pull/1205): `dpctl.tensor.sqrt`
+   - [#1209](https://github.com/IntelPython/dpctl/pull/1209): implements `out` keyword argument
+   - [#1211](https://github.com/IntelPython/dpctl/pull/1211): `dpctl.tensor.multiply`, `dpctl.tensor.subtract`
+   - [#1214](https://github.com/IntelPython/dpctl/pull/1214): `dpctl.tensor.not_equal`
+   - [#1216](https://github.com/IntelPython/dpctl/pull/1216): `dpctl.tensor.exp`, `dpctl.tensor.sin`
+   - [#1217](https://github.com/IntelPython/dpctl/pull/1217): `dpctl.tensor.real`, `dpctl.tensor.imag`, `dpctl.tensor.proj`
+   - [#1218](https://github.com/IntelPython/dpctl/pull/1218): `dpctl.tensor.log`, `dpctl.tensor.log1p`, `dpctl.tensor.expm1`
+   - [#1221](https://github.com/IntelPython/dpctl/pull/1221): `dpctl.tensor.floor_divide`
+   - [#1235](https://github.com/IntelPython/dpctl/pull/1235): `dpctl.tensor.less`
+   - [#1237](https://github.com/IntelPython/dpctl/pull/1237): in-place support for addition, multiplication and subtraction
+* Added `dpctl.tensor.all` and `dpctl.tensor.any` [#1204](https://github.com/IntelPython/dpctl/pull/1204)
+* Added `dpctl.tensor.sum` [#1210](https://github.com/IntelPython/dpctl/pull/1210)
+
+### Changed
+
+* Updated examples of native Python extensions built using `dpctl` [#1108](https://github.com/IntelPython/dpctl/pull/1108)
+* Used security flags to compile and link native extensions of `dpctl` [#1109](https://github.com/IntelPython/dpctl/pull/1109)
+* Changed types of `dpctl.tensor.finfo` and `dpctl.tensor.iinfo` output structure per array API spec [#1110](https://github.com/IntelPython/dpctl/pull/1110)
+* Consolidated multiple USM temporaries life-time management `host_task`s to improve test suite stability [#1111](https://github.com/IntelPython/dpctl/pull/1111)
+* MAINT: Improved cmake target dependency tracking [#1112](https://github.com/IntelPython/dpctl/pull/1112)
+* MAINT: Improved docstrings for existing `dpctl.tensor` functions [#1123](https://github.com/IntelPython/dpctl/pull/1123)
+* Changed default value of `mode` keyword in `dpctl.tensor.take` and `dpctl.take.put` from `clip` to `wrap` [#1132](https://github.com/IntelPython/dpctl/pull/1132)
+* Added support for (nested) sequence of `dpctl.tensor.usm_ndarray` objects in `dpctl.tensor.asarray` [#1139](https://github.com/IntelPython/dpctl/pull/1139)
+* Improved exception handling in `dpctl.tensor.usm_ndarray.__setitem__` special method [#1146](https://github.com/IntelPython/dpctl/pull/1146)
+* Simplified implementation of copy-and-cast kernels and removed special casing for 2D arrays to conserve binary size [#1165](https://github.com/IntelPython/dpctl/pull/1165)
+* Improved speed of `dpctl.tensor.usm_ndarray` printing functionality [#1187](https://github.com/IntelPython/dpctl/pull/1187)
+* Require DPC++ RT 2023.1 to build and run `dpctl` [#1195](ttps://github.com/IntelPython/dpctl/pull/1195)
+* Compile offloading native extensions with `-fno-sycl-id-queries-fit-in-int` fixing [gh-1184](https://github.com/IntelPython/dpctl/issues/1184), [#1200](https://github.com/IntelPython/dpctl/pull/1200)
+* Transition to conda-forge ecosystem [#1213](https://github.com/IntelPython/dpctl/pull/1213)
+
+
+
+### Fixed
+
+* Fix to add empty values check for `dpctl.tensor.place` [#1105](https://github.com/IntelPython/dpctl/pull/1105), [#1106](https://github.com/IntelPython/dpctl/pull/1106)
+* Fixed [gh-1089](https://github.com/IntelPython/dpctl/issues/1089) by improving `dpctl.tensor.asarray` handling of NumPy arrays viewing into host-accessible USM allocation objects.
+* MAINT: Fixed build break with newer GCC and SYCLOS [#1118](https://github.com/IntelPython/dpctl/pull/1118)
+* Fixed a bug in basic indexing of `dpctl.tensor.usm_ndarray` [#1136](https://github.com/IntelPython/dpctl/pull/1136)
+
+
+
+
 ## [0.14.2] - 03/07/2023
 
 ### Fixed

conda-recipe/build.sh

@@ -1,10 +1,7 @@
 #!/bin/bash
 
-# Workaround to Klocwork overwriting LD_LIBRARY_PATH that was modified
-# by DPC++ compiler conda packages. Will need to be added to DPC++ compiler
-# activation scripts.
-export LDFLAGS="$LDFLAGS -Wl,-rpath,$PREFIX/lib"
-export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$PREFIX/compiler/lib/intel64_lin:$PREFIX/compiler/lib:$PREFIX/lib"
+# This is necessary to help DPC++ find Intel libraries such as SVML, IRNG, etc in build prefix
+export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:${BUILD_PREFIX}/lib"
 
 # Intel LLVM must cooperate with compiler and sysroot from conda
 echo "--gcc-toolchain=${BUILD_PREFIX} --sysroot=${BUILD_PREFIX}/${HOST}/sysroot -target ${HOST}" > icpx_for_conda.cfg

conda-recipe/meta.yaml

@@ -13,23 +13,22 @@ build:
 requirements:
     build:
         - {{ compiler('cxx') }}
-        - {{ compiler('dpcpp') }} >=2023.1  # [not osx]
+        - {{ compiler('dpcpp') }} >=2023.2  # [not osx]
         - sysroot_linux-64 >=2.28  # [linux]
     host:
         - setuptools
         - cmake  >=3.21
         - ninja
         - git
-        - cython  <3
+        - cython
         - python
         - scikit-build
         - numpy
         - wheel
     run:
         - python
         - {{ pin_compatible('numpy', min_pin='x.x', max_pin='x') }}
-        - {{ pin_compatible('dpcpp-cpp-rt', min_pin='x.x', max_pin='x') }}  # [py<=39]
-        - dpcpp-cpp-rt >=2023.1  # [py>39]
+        - dpcpp-cpp-rt >=2023.2
         - level-zero  # [linux]
 
 test:

docs/Doxyfile.in

@@ -32,7 +32,7 @@ DOXYFILE_ENCODING      = UTF-8
 # title of most generated pages and in a few other places.
 # The default value is: My Project.
 
-PROJECT_NAME           = "Data-parallel Control (dpctl)"
+PROJECT_NAME           = "Data Parallel Control (dpctl)"
 
 # The PROJECT_NUMBER tag can be used to enter a project or revision number. This
 # could be handy for archiving the generated documentation or if some version

docs/conf.in

@@ -36,7 +36,7 @@ import extlinks_gen as urlgen
 
 # -- Project information -----------------------------------------------------
 
-project = "Data-parallel Control (dpctl)"
+project = "Data Parallel Control (dpctl)"
 copyright = "2020-21, Intel Corp."
 author = "Intel Corp."
 

docs/docfiles/intro.rst

@@ -1,7 +1,7 @@
-Welcome to the Data-parallel Control (dpctl) Documentation!
+Welcome to the Data Parallel Control (dpctl) Documentation!
 ===========================================================
 
-The data-parallel control (dpctl) library provides C and Python bindings for
+The Data Parallel Control (dpctl) package provides C and Python bindings for
 :sycl_spec_2020:`SYCL 2020 <>`. The SYCL 2020 features supported by dpctl are
 limited to those included by Intel(R) DPC++ compiler and specifically cover the
 SYCL runtime classes described in :sycl_runtime_classes:`Section 4.6 <>`

docs/generate_rst.py

@@ -170,7 +170,7 @@ def _is_class_property(o):
     Returns:
         [type]: [description]
     """
-    return isinstance(o, property) or (type(o) == _getset_descriptor)
+    return isinstance(o, property) or (type(o) is _getset_descriptor)
 
 
 def _is_class_method(o):

dpctl/_backend.pxd

@@ -95,7 +95,6 @@ cdef extern from "syclinterface/dpctl_sycl_enum_types.h":
         _usm_device_allocations             'usm_device_allocations',
         _usm_host_allocations               'usm_host_allocations',
         _usm_shared_allocations             'usm_shared_allocations',
-        _usm_restricted_shared_allocations  'usm_restricted_shared_allocations',
         _usm_system_allocations             'usm_system_allocations',
         _usm_atomic_host_allocations        'usm_atomic_host_allocations',
         _usm_atomic_shared_allocations      'usm_atomic_shared_allocations',

dpctl/_sycl_context.pyx

@@ -68,7 +68,7 @@ cdef class SyclContextCreationError(Exception):
     pass
 
 
-cdef void _context_capsule_deleter(object o):
+cdef void _context_capsule_deleter(object o) noexcept:
     cdef DPCTLSyclContextRef CRef = NULL
     if pycapsule.PyCapsule_IsValid(o, "SyclContextRef"):
         CRef = <DPCTLSyclContextRef> pycapsule.PyCapsule_GetPointer(

dpctl/_sycl_device.pyx

@@ -597,20 +597,6 @@ cdef class SyclDevice(_SyclDevice):
         cdef _aspect_type AT = _aspect_type._usm_shared_allocations
         return DPCTLDevice_HasAspect(self._device_ref, AT)
 
-    @property
-    def has_aspect_usm_restricted_shared_allocations(self):
-        """ Returns True if this device supports USM memory
-        allocated as restricted USM, False otherwise.
-
-        Returns:
-            bool: Indicates that the device supports USM memory allocated using
-            ``sycl::malloc_shared`` as restricted USM.
-
-        .. deprecated:: 0.14
-        """
-        cdef _aspect_type AT = _aspect_type._usm_restricted_shared_allocations
-        return DPCTLDevice_HasAspect(self._device_ref, AT)
-
     @property
     def has_aspect_usm_system_allocations(self):
         """ Returns True if system allocator may be used instead of SYCL USM

dpctl/_sycl_event.pyx

@@ -73,7 +73,7 @@ cdef api SyclEvent SyclEvent_Make(DPCTLSyclEventRef ERef):
     return SyclEvent._create(copied_ERef)
 
 
-cdef void _event_capsule_deleter(object o):
+cdef void _event_capsule_deleter(object o) noexcept:
     cdef DPCTLSyclEventRef ERef = NULL
     if pycapsule.PyCapsule_IsValid(o, "SyclEventRef"):
         ERef = <DPCTLSyclEventRef> pycapsule.PyCapsule_GetPointer(

dpctl/_sycl_queue.pyx

@@ -121,7 +121,7 @@ cdef class SyclAsynchronousError(Exception):
     """
 
 
-cdef void default_async_error_handler(int err) nogil except *:
+cdef void default_async_error_handler(int err) except * nogil:
     with gil:
         raise SyclAsynchronousError(err)
 
@@ -159,7 +159,7 @@ cdef int _parse_queue_properties(object prop) except *:
     return res
 
 
-cdef void _queue_capsule_deleter(object o):
+cdef void _queue_capsule_deleter(object o) noexcept:
     cdef DPCTLSyclQueueRef QRef = NULL
     if pycapsule.PyCapsule_IsValid(o, "SyclQueueRef"):
         QRef = <DPCTLSyclQueueRef> pycapsule.PyCapsule_GetPointer(

dpctl/apis/include/dpctl4pybind11.hpp

@@ -965,6 +965,15 @@ class usm_ndarray : public py::object
         return *(reinterpret_cast<sycl::queue *>(QRef));
     }
 
+    sycl::device get_device() const
+    {
+        PyUSMArrayObject *raw_ar = usm_array_ptr();
+
+        auto const &api = ::dpctl::detail::dpctl_capi::get();
+        DPCTLSyclQueueRef QRef = api.UsmNDArray_GetQueueRef_(raw_ar);
+        return reinterpret_cast<sycl::queue *>(QRef)->get_device();
+    }
+
     int get_typenum() const
     {
         PyUSMArrayObject *raw_ar = usm_array_ptr();

dpctl/apis/include/dpctl_capi.h

@@ -28,6 +28,7 @@
 // Ordering of includes is important here. dpctl_sycl_types defines types
 // used by dpctl's Python C-API headers.
 #include "syclinterface/dpctl_sycl_types.h"
+#define CYTHON_EXTERN_C extern "C"
 #include "../_sycl_device.h"
 #include "../_sycl_device_api.h"
 #include "../_sycl_context.h"

dpctl/tensor/CMakeLists.txt

@@ -54,6 +54,8 @@ if (WIN32)
   set(_clang_prefix "/clang:")
 endif()
 set_source_files_properties(
+  ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/full_ctor.cpp
+  ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/linear_sequences.cpp
   ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/elementwise_functions.cpp
   PROPERTIES COMPILE_OPTIONS "${_clang_prefix}-fno-fast-math")
 target_compile_options(${python_module_name} PRIVATE -fno-sycl-id-queries-fit-in-int)