Merge pull request #1396 from IntelPython/master

Merge 0.14.6dev5 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', '3.11']
+        python: ['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', '3.11']
+        python: ['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', '3.11']
+        python: ['3.9', '3.10', '3.11']
         experimental: [false]
         runner: [ubuntu-20.04]
     continue-on-error: ${{ matrix.experimental }}
@@ -185,7 +185,7 @@ jobs:
         shell: cmd /C CALL {0}
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10', '3.11']
+        python: ['3.9', '3.10', '3.11']
         experimental: [false]
         runner: [windows-latest]
     continue-on-error: ${{ matrix.experimental }}
@@ -300,7 +300,7 @@ jobs:
     runs-on: ubuntu-20.04
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10', '3.11']
+        python: ['3.9', '3.10', '3.11']
     steps:
       - name: Download artifact
         uses: actions/download-artifact@v3
@@ -324,7 +324,7 @@ jobs:
     runs-on: windows-latest
     strategy:
       matrix:
-        python: ['3.8', '3.9', '3.10', '3.11']
+        python: ['3.9', '3.10', '3.11']
     steps:
       - name: Download artifact
         uses: actions/download-artifact@v3
@@ -608,7 +608,11 @@ jobs:
             echo "Array API standard conformance tests failed to run for dpctl=$PACKAGE_VERSION."
             exit 1
           fi
+      - name: Output API summary
+        shell: bash -l {0}
+        run: echo "::notice ${{ env.MESSAGE }}"
       - name: Post result to PR
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork }}
         uses: mshick/add-pr-comment@v1
         with:
           message: |

.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 setuptools scikit-build cmake sphinx sphinx_rtd_theme pydot graphviz sphinxcontrib-programoutput sphinxcontrib-googleanalytics
+          pip install numpy cython setuptools scikit-build cmake sphinx"<7.2" sphinx_rtd_theme pydot graphviz sphinxcontrib-programoutput sphinxcontrib-googleanalytics
       - name: Checkout repo
         uses: actions/checkout@v3
         with:
@@ -76,7 +76,7 @@ jobs:
           mv ../cmake-install/docs/docs ~/docs
           git clean -dfx
       - name: Publish docs
-        if: ${{ github.ref == 'refs/heads/master' }}
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork && github.ref == 'refs/heads/master' }}
         shell: bash -l {0}
         run: |
           git remote add tokened_docs https://IntelPython:${{ secrets.GITHUB_TOKEN }}@github.com/IntelPython/dpctl.git
@@ -91,8 +91,15 @@ jobs:
           git config --global user.email 'github-actions[doc-deploy-bot]@users.noreply.github.com'
           git commit -m "Latest docs."
           git push tokened_docs gh-pages
+      - name: Save built docs as an artifact
+        if: ${{ github.event.pull_request && github.event.pull_request.head.repo.fork && github.event.action != 'closed'}}
+        uses: actions/upload-artifact@v3
+        with:
+          name: ${{ env.PACKAGE_NAME }} rendered documentation
+          path: ~/docs
+
       - name: Publish pull-request docs
-        if: ${{ github.event.pull_request && github.event.action != 'closed' }}
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork && github.event.action != 'closed' }}
         env:
           PR_NUM: ${{ github.event.number }}
         shell: bash -l {0}
@@ -111,7 +118,7 @@ jobs:
           git commit -m "Docs for pull request ${PR_NUM}"
           git push tokened_docs gh-pages
       - name: Unpublish pull-request docs
-        if: ${{ github.event.pull_request && github.event.action == 'closed' }}
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork && github.event.action == 'closed' }}
         env:
           PR_NUM: ${{ github.event.number }}
         shell: bash -l {0}
@@ -128,7 +135,7 @@ jobs:
           git commit -m "Removing docs for closed pull request ${PR_NUM}"
           git push tokened_docs gh-pages
       - name: Comment with URL to published pull-request docs
-        if: ${{ github.event.pull_request && github.event.action != 'closed' }}
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork && github.event.action != 'closed' }}
         env:
           PR_NUM: ${{ github.event.number }}
         uses: mshick/add-pr-comment@v1
@@ -138,7 +145,7 @@ jobs:
           repo-token: ${{ secrets.GITHUB_TOKEN }}
           repo-token-user-login: 'github-actions[bot]'
       - name: Comment with URL about removal of PR docs
-        if: ${{ github.event.pull_request && github.event.action == 'closed' }}
+        if: ${{ github.event.pull_request && !github.event.pull_request.head.repo.fork && github.event.action == 'closed' }}
         env:
           PR_NUM: ${{ github.event.number }}
         uses: mshick/add-pr-comment@v1

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

@@ -97,18 +97,18 @@ jobs:
       - name: Install system components
         shell: bash -l {0}
         run: |
-          sudo apt-get install ninja-build libtinfo5
+          sudo apt-get install libtinfo5
 
       - name: Setup Python
         uses: actions/setup-python@v4
         with:
-          python-version: '3.9'
+          python-version: '3.11'
           architecture: x64
 
       - name: Install dpctl dependencies
         shell: bash -l {0}
         run: |
-          pip install numpy cython setuptools pytest scikit-build cmake
+          pip install numpy cython setuptools pytest scikit-build cmake ninja
 
       - name: Checkout repo
         uses: actions/checkout@v3

cmake/FindDpctl.cmake

@@ -17,15 +17,8 @@
 #
 
 if(NOT Dpctl_FOUND)
-  set(_find_extra_args)
-  if(Dpctl_FIND_REQUIRED)
-    list(APPEND _find_extra_args REQUIRED)
-  endif()
-  if(Dpctl_FIND_QUIET)
-    list(APPEND _find_extra_args QUIET)
-  endif()
-  find_package(PythonInterp ${_find_extra_args})
-  find_package(PythonLibs ${_find_extra_args})
+  find_package(Python 3.9 REQUIRED
+    COMPONENTS Interpreter Development.Module)
 
   if(PYTHON_EXECUTABLE)
     execute_process(COMMAND "${PYTHON_EXECUTABLE}"

conda-recipe/run_test.bat

@@ -3,8 +3,8 @@
 "%PYTHON%" -c "import dpctl; print(dpctl.__version__)"
 if errorlevel 1 exit 1
 
-"%PYTHON%" -c "import dpctl; dpctl.lsplatform()"
+"%PYTHON%" -m dpctl -f
 if errorlevel 1 exit 1
 
-python -m pytest -q -p no:faulthandler -ra --disable-warnings --pyargs dpctl -vv
+python -m pytest -q -ra --disable-warnings --pyargs dpctl -vv
 if errorlevel 1 exit 1

conda-recipe/run_test.sh

@@ -3,5 +3,5 @@
 set -e
 
 ${PYTHON} -c "import dpctl; print(dpctl.__version__)"
-${PYTHON} -c "import dpctl; dpctl.lsplatform(verbosity=2)"
-${PYTHON} -m pytest -q -ra --disable-warnings -p no:faulthandler --cov dpctl --cov-report term-missing --pyargs dpctl -vv
+${PYTHON} -m dpctl -f
+${PYTHON} -m pytest -q -ra --disable-warnings --cov dpctl --cov-report term-missing --pyargs dpctl -vv

dpctl/_sycl_context.pyx

@@ -114,7 +114,7 @@ cdef class SyclContext(_SyclContext):
                 ctx = dpctl.SyclContext()
                 print(ctx.get_devices())
 
-        - Invoking the constuctor with a specific filter string that creates a
+        - Invoking the constructor with a specific filter string that creates a
           context for the device corresponding to the filter string.
 
         :Example:
@@ -127,7 +127,7 @@ cdef class SyclContext(_SyclContext):
                 d = ctx.get_devices()[0]
                 assert(d.is_gpu)
 
-        - Invoking the constuctor with a :class:`dpctl.SyclDevice` object
+        - Invoking the constructor with a :class:`dpctl.SyclDevice` object
           creates a context for that device.
 
         :Example:
@@ -141,7 +141,7 @@ cdef class SyclContext(_SyclContext):
                 d = ctx.get_devices()[0]
                 assert(d.is_gpu)
 
-        - Invoking the constuctor with a list of :class:`dpctl.SyclDevice`
+        - Invoking the constructor with a list of :class:`dpctl.SyclDevice`
           objects creates a common context for all the devices. This
           constructor call is especially useful when creation a context for
           multiple sub-devices.
@@ -159,7 +159,7 @@ cdef class SyclContext(_SyclContext):
                 ctx = dpctl.SyclContext(sub_devices)
                 assert(len(ctx.get_devices) == len(sub_devices))
 
-        - Invoking the constuctor with a named ``PyCapsule`` with name
+        - Invoking the constructor with a named ``PyCapsule`` with name
           **"SyclContextRef"** that carries a pointer to a ``sycl::context``
           object. The capsule will be renamed upon successful consumption
           to ensure one-time use. A new named capsule can be constructed by
@@ -430,7 +430,7 @@ cdef class SyclContext(_SyclContext):
             return num_devs
         else:
             raise ValueError(
-                "An error was encountered quering the number of devices "
+                "An error was encountered querying the number of devices "
                 "associated with this context"
             )
 

dpctl/_sycl_device.pyx

@@ -1163,7 +1163,7 @@ cdef class SyclDevice(_SyclDevice):
     def create_sub_devices(self, **kwargs):
         """create_sub_devices(partition=parition_spec)
         Creates a list of sub-devices by partitioning a root device based on the
-        provided partion specifier.
+        provided partition specifier.
 
         A partition specifier must be provided using a "partition"
         keyword argument. Possible values for the specifier are: an int, a

dpctl/_sycl_event.pyx

@@ -119,7 +119,7 @@ cdef class SyclEvent(_SyclEvent):
                 # Create a default SyclEvent
                 e = dpctl.SyclEvent()
 
-        - Invoking the constuctor with a named ``PyCapsule`` with name
+        - Invoking the constructor with a named ``PyCapsule`` with name
           **"SyclEventRef"** that carries a pointer to a ``sycl::event``
           object. The capsule will be renamed upon successful consumption
           to ensure one-time use. A new named capsule can be constructed by

dpctl/_sycl_queue.pyx

@@ -246,7 +246,7 @@ cdef class SyclQueue(_SyclQueue):
                 # create a queue for each sub-device using the common context
                 queues = [dpctl.SyclQueue(ctx, sub_d) for sub_d in sub_devices]
 
-        - Invoking the constuctor with a named ``PyCapsule`` with the name
+        - Invoking the constructor with a named ``PyCapsule`` with the name
           **"SyclQueueRef"** that carries a pointer to a ``sycl::queue``
           object. The capsule will be renamed upon successful consumption
           to ensure one-time use. A new named capsule can be constructed by

dpctl/_sycl_timer.py

@@ -37,14 +37,14 @@ class SyclTimer:
             q = dpctl.SyclQueue(property='enable_profiling')
 
             # create the timer
-            miliseconds_sc = 1e-3
-            timer = dpctl.SyclTimer(time_scale = miliseconds_sc)
+            milliseconds_sc = 1e-3
+            timer = dpctl.SyclTimer(time_scale = milliseconds_sc)
 
             # use the timer
             with timer(queue=q):
                 code_block
 
-            # retrieve elapsed times in miliseconds
+            # retrieve elapsed times in milliseconds
             sycl_dt, wall_dt = timer.dt
 
     Remark:

dpctl/apis/include/dpctl4pybind11.hpp

@@ -742,12 +742,12 @@ class usm_memory : public py::object
             return nullptr;
         }
 
-        auto convertor =
+        auto converter =
             ::dpctl::detail::dpctl_capi::get().as_usm_memory_pyobj();
 
         py::object res;
         try {
-            res = convertor(py::handle(o));
+            res = converter(py::handle(o));
         } catch (const py::error_already_set &e) {
             return nullptr;
         }

dpctl/memory/_memory.pyx

@@ -92,7 +92,7 @@ cdef void copy_via_host(void *dest_ptr, SyclQueue dest_queue,
                         void *src_ptr, SyclQueue src_queue, size_t nbytes):
     """
     Copies `nbytes` bytes from `src_ptr` USM memory to
-    `dest_ptr` USM memory using host as the intemediary.
+    `dest_ptr` USM memory using host as the intermediary.
 
     This is useful when `src_ptr` and `dest_ptr` are bound to incompatible
     SYCL contexts.

dpctl/tensor/CMakeLists.txt

@@ -37,6 +37,7 @@ pybind11_add_module(${python_module_name} MODULE
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/copy_and_cast_usm_to_usm.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/copy_numpy_ndarray_into_usm_ndarray.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/copy_for_reshape.cpp
+    ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/copy_for_roll.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/linear_sequences.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/integer_advanced_indexing.cpp
     ${CMAKE_CURRENT_SOURCE_DIR}/libtensor/source/boolean_advanced_indexing.cpp

dpctl/tensor/__init__.py

@@ -60,6 +60,7 @@
 from dpctl.tensor._device import Device
 from dpctl.tensor._dlpack import from_dlpack
 from dpctl.tensor._indexing_functions import extract, nonzero, place, put, take
+from dpctl.tensor._linear_algebra_functions import matrix_transpose
 from dpctl.tensor._manipulation_functions import (
     broadcast_arrays,
     broadcast_to,
@@ -199,6 +200,7 @@
     "tril",
     "triu",
     "where",
+    "matrix_transpose",
     "all",
     "any",
     "dtype",

dpctl/tensor/_copy_utils.py

@@ -56,7 +56,7 @@ def _copy_to_numpy(ary):
 
 def _copy_from_numpy(np_ary, usm_type="device", sycl_queue=None):
     "Copies numpy array `np_ary` into a new usm_ndarray"
-    # This may peform a copy to meet stated requirements
+    # This may perform a copy to meet stated requirements
     Xnp = np.require(np_ary, requirements=["A", "E"])
     alloc_q = normalize_queue_device(sycl_queue=sycl_queue, device=None)
     dt = Xnp.dtype
@@ -517,6 +517,11 @@ def copy(usm_ary, order="K"):
        - "K": match the layout of `usm_ary` as closely as possible.
 
     """
+    if len(order) == 0 or order[0] not in "KkAaCcFf":
+        raise ValueError(
+            "Unrecognized order keyword value, expecting 'K', 'A', 'F', or 'C'."
+        )
+    order = order[0].upper()
     if not isinstance(usm_ary, dpt.usm_ndarray):
         return TypeError(
             f"Expected object of type dpt.usm_ndarray, got {type(usm_ary)}"
@@ -585,11 +590,11 @@ def astype(usm_ary, newdtype, order="K", casting="unsafe", copy=True):
         return TypeError(
             f"Expected object of type dpt.usm_ndarray, got {type(usm_ary)}"
         )
-    if not isinstance(order, str) or order not in ["A", "C", "F", "K"]:
+    if len(order) == 0 or order[0] not in "KkAaCcFf":
         raise ValueError(
-            "Unrecognized value of the order keyword. "
-            "Recognized values are 'A', 'C', 'F', or 'K'"
+            "Unrecognized order keyword value, expecting 'K', 'A', 'F', or 'C'."
         )
+    order = order[0].upper()
     ary_dtype = usm_ary.dtype
     target_dtype = _get_dtype(newdtype, usm_ary.sycl_queue)
     if not dpt.can_cast(ary_dtype, target_dtype, casting=casting):

dpctl/tensor/_device.py

@@ -96,7 +96,7 @@ def sycl_context(self):
     @property
     def sycl_device(self):
         """
-        :class:`dpctl.SyclDevice` targed by this :class:`.Device`.
+        :class:`dpctl.SyclDevice` targeted by this :class:`.Device`.
         """
         return self.sycl_queue_.sycl_device