Merge remote-tracking branch 'origin/master' into revert-kw-workaround

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 }}
@@ -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

CHANGELOG.md

@@ -4,6 +4,97 @@ 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
+### 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

@@ -5,6 +5,9 @@
 # activation scripts.
 export LDFLAGS="$LDFLAGS -Wl,-rpath,$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 ${HOST}" > icpx_for_conda.cfg
 export ICPXCFG="$(pwd)/icpx_for_conda.cfg"

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 <>`

dpctl/tensor/__init__.py

@@ -94,12 +94,14 @@
 from ._elementwise_funcs import (
     abs,
     add,
+    ceil,
     conj,
     cos,
     divide,
     equal,
     exp,
     expm1,
+    floor,
     floor_divide,
     greater,
     greater_equal,
@@ -128,6 +130,7 @@
     sqrt,
     square,
     subtract,
+    trunc,
 )
 from ._reduction import sum
 
@@ -208,16 +211,21 @@
     "inf",
     "abs",
     "add",
+    "ceil",
     "conj",
     "cos",
+    "divide",
+    "equal",
     "exp",
     "expm1",
+    "floor",
+    "floor_divide",
     "greater",
     "greater_equal",
     "imag",
+    "isfinite",
     "isinf",
     "isnan",
-    "isfinite",
     "less",
     "less_equal",
     "log",
@@ -228,19 +236,17 @@
     "log1p",
     "log2",
     "log10",
+    "multiply",
     "negative",
+    "not_equal",
     "positive",
+    "pow",
     "proj",
     "real",
     "sin",
     "sqrt",
     "square",
-    "divide",
-    "multiply",
-    "pow",
     "subtract",
-    "equal",
-    "not_equal",
     "sum",
-    "floor_divide",
+    "trunc",
 ]

dpctl/tensor/_elementwise_common.py

@@ -58,7 +58,12 @@ def __call__(self, x, out=None, order="K"):
             x.dtype, self.result_type_resolver_fn_, x.sycl_device
         )
         if res_dt is None:
-            raise RuntimeError
+            raise TypeError(
+                f"function '{self.name_}' does not support input type "
+                f"({x.dtype}), "
+                "and the input could not be safely coerced to any "
+                "supported types according to the casting rule ''safe''."
+            )
 
         orig_out = out
         if out is not None:

dpctl/tensor/_elementwise_funcs.py

@@ -114,7 +114,30 @@
 # FIXME: implement B07
 
 # U09: ==== CEIL          (x)
-# FIXME: implement U09
+_ceil_docstring = """
+ceil(x, out=None, order='K')
+
+Returns the ceiling for each element `x_i` for input array `x`.
+The ceil of the scalar `x` is the smallest integer `i`, such that `i >= x`.
+
+Args:
+    x (usm_ndarray):
+        Input array, expected to have numeric data type.
+    out ({None, usm_ndarray}, optional):
+        Output array to populate.
+        Array have the correct shape and the expected data type.
+    order ("C","F","A","K", optional):
+        Memory layout of the newly output array, if parameter `out` is `None`.
+        Default: "K".
+Returns:
+    usm_narray:
+        An array containing the element-wise ceiling of input array.
+        The returned array has the same data type as `x`.
+"""
+
+ceil = UnaryElementwiseFunc(
+    "ceil", ti._ceil_result_type, ti._ceil, _ceil_docstring
+)
 
 # U10: ==== CONJ          (x)
 _conj_docstring = """
@@ -271,7 +294,30 @@
 )
 
 # U15: ==== FLOOR         (x)
-# FIXME: implement U15
+_floor_docstring = """
+floor(x, out=None, order='K')
+
+Returns the floor for each element `x_i` for input array `x`.
+The floor of the scalar `x` is the largest integer `i`, such that `i <= x`.
+
+Args:
+    x (usm_ndarray):
+        Input array, expected to have numeric data type.
+    out ({None, usm_ndarray}, optional):
+        Output array to populate.
+        Array have the correct shape and the expected data type.
+    order ("C","F","A","K", optional):
+        Memory layout of the newly output array, if parameter `out` is `None`.
+        Default: "K".
+Returns:
+    usm_narray:
+        An array containing the element-wise floor of input array.
+        The returned array has the same data type as `x`.
+"""
+
+floor = UnaryElementwiseFunc(
+    "floor", ti._floor_result_type, ti._floor, _floor_docstring
+)
 
 # B10: ==== FLOOR_DIVIDE  (x1, x2)
 _floor_divide_docstring_ = """
@@ -1031,4 +1077,29 @@
 # FIXME: implement U35
 
 # U36: ==== TRUNC       (x)
-# FIXME: implement U36
+_trunc_docstring = """
+trunc(x, out=None, order='K')
+
+Returns the truncated value for each element `x_i` for input array `x`.
+The truncated value of the scalar `x` is the nearest integer i which is
+ closer to zero than `x` is. In short, the fractional part of the
+ signed number `x` is discarded.
+
+Args:
+    x (usm_ndarray):
+        Input array, expected to have numeric data type.
+    out ({None, usm_ndarray}, optional):
+        Output array to populate.
+        Array have the correct shape and the expected data type.
+    order ("C","F","A","K", optional):
+        Memory layout of the newly output array, if parameter `out` is `None`.
+        Default: "K".
+Returns:
+    usm_narray:
+        An array containing the element-wise truncated value of input array.
+        The returned array has the same data type as `x`.
+"""
+
+trunc = UnaryElementwiseFunc(
+    "trunc", ti._trunc_result_type, ti._trunc, _trunc_docstring
+)

dpctl/tensor/_reduction.py

@@ -123,6 +123,10 @@ def sum(arr, axis=None, dtype=None, keepdims=False):
 
     res_usm_type = arr.usm_type
     if arr.size == 0:
+        if keepdims:
+            res_shape = res_shape + (1,) * red_nd
+            inv_perm = sorted(range(nd), key=lambda d: perm[d])
+            res_shape = tuple(res_shape[i] for i in inv_perm)
         return dpt.zeros(
             res_shape, dtype=res_dt, usm_type=res_usm_type, sycl_queue=q
         )