Dpnp

Added

* Added implementation of `dpnp.hamming` [2341](https://github.com/IntelPython/dpnp/pull/2341), [#2357](https://github.com/IntelPython/dpnp/pull/2357)
* Added implementation of `dpnp.hanning` [2358](https://github.com/IntelPython/dpnp/pull/2358)
* Added implementation of `dpnp.blackman` [2363](https://github.com/IntelPython/dpnp/pull/2363)
* Added implementation of `dpnp.bartlett` [2366](https://github.com/IntelPython/dpnp/pull/2366)

Changed

* Extended `dpnp.fft.fftfreq` and `dpnp.fft.rfftfreq` functions to support `dtype` keyword per Python Array API spec 2024.12 [2384](https://github.com/IntelPython/dpnp/pull/2384)
* Allowed input array of `uint64` dtype in `dpnp.bincount` [2361](https://github.com/IntelPython/dpnp/pull/2361)
* The vector norms `ord={None, 1, 2, inf}` and the matrix norms `ord={None, 1, 2, inf, "fro", "nuc"}` now consistently return zero for empty arrays, which are arrays with at least one axis of size zero. This change affects `dpnp.linalg.norm`, `dpnp.linalg.vector_norm`, and `dpnp.linalg.matrix_norm`. Previously, dpnp would either raise errors or return zero depending on the parameters provided [2371](https://github.com/IntelPython/dpnp/pull/2371)
* Improved performance of `dpnp.nansum`, `dpnp.nanprod`, `dpnp.nancumsum`, and `dpnp.nancumprod` by reusing `dpnp.nan_to_num` function in implementation of the functions [2339](https://github.com/IntelPython/dpnp/pull/2339)

Fixed

* Resolved an issue with an incorrect result returned due to missing dependency from the strided kernel on a copy event in `dpnp.erf` [2378](https://github.com/IntelPython/dpnp/pull/2378)

This release achieves 100% compliance with Python Array API specification (revision [2023.12](https://data-apis.org/array-api/2023.12/)).
The release provides enhanced compatibility with NumPy 2.2.3. Array manipulation, mathematical, logic, and statistics routines are complemented by a set of new functions.
Furthermore, a number of issues relating to running on NVIDIA GPUs have been resolved.

Added

* Added implementation of `dpnp.gcd` and `dpnp.lcm` functions [2091](https://github.com/IntelPython/dpnp/pull/2091)
* Added implementation of `dpnp.pad` function [2093](https://github.com/IntelPython/dpnp/pull/2093)
* Added implementation of `dpnp.linalg.svdvals` function [2094](https://github.com/IntelPython/dpnp/pull/2094)
* Added implementation of `dpnp.matrix_transpose` function and `dpnp.ndarray.mT` attribute [2095](https://github.com/IntelPython/dpnp/pull/2095)
* Exposed `cross`, `diagonal`, `matrix_norm`, `outer`, `tensordot`, `trace` and `vector_norm` functions as part of `dpnp.linalg` namespace [2099](https://github.com/IntelPython/dpnp/pull/2099)
* Added implementation of `dpnp.unstack` function [2106](https://github.com/IntelPython/dpnp/pull/2106)
* Added implementation of `dpnp.ldexp` function [2110](https://github.com/IntelPython/dpnp/pull/2110)
* Added implementation of `dpnp.vecdot` and `dpnp.linalg.vecdot` functions [2112](https://github.com/IntelPython/dpnp/pull/2112)
* Added implementation of `dpnp.i0` function [2118](https://github.com/IntelPython/dpnp/pull/2118)
* Added implementation of `dpnp.isfortran` function [2122](https://github.com/IntelPython/dpnp/pull/2122)
* Added implementation of `dpnp.spacing` function [2125](https://github.com/IntelPython/dpnp/pull/2125)
* Added implementation of `dpnp.sinc` function [2133](https://github.com/IntelPython/dpnp/pull/2133)
* Added implementation of `dpnp.corrcoef` function [2139](https://github.com/IntelPython/dpnp/pull/2139)
* Added implementation of `dpnp.delete` function [2142](https://github.com/IntelPython/dpnp/pull/2142)
* Added implementation of `dpnp.histogramdd` function [2143](https://github.com/IntelPython/dpnp/pull/2143)
* Added implementation of `dpnp.bincount` function [2145](https://github.com/IntelPython/dpnp/pull/2145)
* Added support of inplace matrix multiplication via the `=` operator [2147](https://github.com/IntelPython/dpnp/pull/2147)
* Added implementation of `dpnp.insert` function [2151](https://github.com/IntelPython/dpnp/pull/2151)
* Added implementation of `dpnp.broadcast_shapes` function [2153](https://github.com/IntelPython/dpnp/pull/2153)
* Added implementation of `dpnp.byte_bounds` function [2155](https://github.com/IntelPython/dpnp/pull/2155)
* Added implementation of `dpnp.ndindex` class [2157](https://github.com/IntelPython/dpnp/pull/2157)
* Added implementation of `dpnp.histogram2d` function [2262](https://github.com/IntelPython/dpnp/pull/2262)
* Added implementation of `dpnp.binary_repr` function [2168](https://github.com/IntelPython/dpnp/pull/2168)
* Added implementation of `dpnp.apply_along_axis` function [2169](https://github.com/IntelPython/dpnp/pull/2169)
* Added implementation of `dpnp.cumulative_sum` and `dpnp.cumulative_prod` functions [2171](https://github.com/IntelPython/dpnp/pull/2171)
* Added implementation of `dpnp.apply_over_axes` function [2174](https://github.com/IntelPython/dpnp/pull/2174)
* Added implementation of `dpnp.compress` function and `dpnp_array.compress` method [2177](https://github.com/IntelPython/dpnp/pull/2177)
* Added implementation of `dpnp.correlate` function [2180](https://github.com/IntelPython/dpnp/pull/2180), [#2203](https://github.com/IntelPython/dpnp/pull/2203)
* Added implementation of `dpnp.nanmedian` function [2191](https://github.com/IntelPython/dpnp/pull/2191)
* Added implementation of `dpnp.ndarray.__iter__` method [2206](https://github.com/IntelPython/dpnp/pull/2206)
* Added implementation of `dpnp.iterable` function [2208](https://github.com/IntelPython/dpnp/pull/2208)
* Added missing aliases on integer data types [2230](https://github.com/IntelPython/dpnp/pull/2230)
* Enabled validation of dpnp conda/wheel packages with Python 3.13 (limited support) [2249](https://github.com/IntelPython/dpnp/pull/2249)
* Added implementation of `dpnp.ndarray.__array_namespace__` method [2252](https://github.com/IntelPython/dpnp/pull/2252)
* Added implementation of `dpnp.ndarray.__usm_ndarray__` protocol [2261](https://github.com/IntelPython/dpnp/pull/2261)
* Added implementation of `dpnp.isdtype` function [2274](https://github.com/IntelPython/dpnp/pull/2274)
* Added implementation of Python Array API Inspection namespace [2275](https://github.com/IntelPython/dpnp/pull/2275)
* Added implementation of `dpnp.matvec` and `dpnp.vecmat` functions [2288](https://github.com/IntelPython/dpnp/pull/2288)
* Added implementation of `dpnp.unique_all`, `dpnp.unique_counts`, `dpnp.unique_inverse` and `dpnp.unique_values` functions [2320](https://github.com/IntelPython/dpnp/pull/2320)

Changed

* Improved performance of `dpnp.histogram` function by implementing a dedicated kernel [2027](https://github.com/IntelPython/dpnp/pull/2027)
* Improved performance of `dpnp.ndarray.fill` method by leveraging on dpctl extension exposing `fill` kernel [2055](https://github.com/IntelPython/dpnp/pull/2055)
* Extended `dpnp.ndarray.reshape` method and `dpnp.reshape` function to support `shape` and `newshape` keywords [2080](https://github.com/IntelPython/dpnp/pull/2080)
* Extended support of `order` keyword in like-functions [2088](https://github.com/IntelPython/dpnp/pull/2088)
* Updated `dpnp.einsum` function to comply with NEP-50 [2120](https://github.com/IntelPython/dpnp/pull/2120)
* Extended `dpnp.linalg.pinv` and `dpnp.linalg.matrix_rank` functions to support `rtol` keyword [2124](https://github.com/IntelPython/dpnp/pull/2124)
* Extended `dpnp.array` function to support `ndmin` keyword [2135](https://github.com/IntelPython/dpnp/pull/2135)
* Leveraged `dpctl.tensor` implementation for `dpnp.put_along_axis` function [2134](https://github.com/IntelPython/dpnp/pull/2134)
* Corrected `dpnp.ndarray.item` implemented to return a python scalar instead of zero-dimensional array [2138](https://github.com/IntelPython/dpnp/pull/2138)
* Bumped NumPy, CuPy and Python versions used for building docs [2158](https://github.com/IntelPython/dpnp/pull/2158)
* Extended `dpnp.sort` and `dpnp.argsort` functions to support `kind="mergesort"` and `kind="radixsort"` values [2159](https://github.com/IntelPython/dpnp/pull/2159)
* Revised and updated information in `README.md` document [2166](https://github.com/IntelPython/dpnp/pull/2166)
* Permitted `"same_kind"` casting for elementwise inplace operators [2170](https://github.com/IntelPython/dpnp/pull/2170)
* Bumped oneMKL version up to `0.6` and added new `--onemkl-interfaces-dir` option to build script [2193](https://github.com/IntelPython/dpnp/pull/2193)
* Updated implementation of `dpnp.linalg.solve` function to align with `numpy >= 2.0` and Python array API [2198](https://github.com/IntelPython/dpnp/pull/2198)
* Improved performance of `dpnp.choose` function by implementing a dedicated kernel [2201](https://github.com/IntelPython/dpnp/pull/2201)
* Aligned with the functional changes introduced by NumPy 2.2 [2226](https://github.com/IntelPython/dpnp/pull/2226)
* Improved performance of `dpnp.nan_to_num` function by implementing a dedicated kernel [2228](https://github.com/IntelPython/dpnp/pull/2228)
* Enabled Intel MKL backends when building from the source with `--onemkl-interfaces` option [2229](https://github.com/IntelPython/dpnp/pull/2229)
* Extended `intersphinx_mapping` with a link to CuPy documentation to make cupy functions clickable from the rendered pages [2232](https://github.com/IntelPython/dpnp/pull/2232)
* Improved performance of `dpnp.nanmedian` function when `axis` is not `None` passed [2240](https://github.com/IntelPython/dpnp/pull/2240)
* Aligned `dpnp.trim_zeros` with NumPy 2.2 and added support of a multi-dimensional input array [2241](https://github.com/IntelPython/dpnp/pull/2241)
* Disallowed implicit conversion of `dpnp.ndarray` to `numpy.ndarray` [2260](https://github.com/IntelPython/dpnp/pull/2260)
* Extended `dpnp.ndarray.to_device` method to support `stream` keyword [2263](https://github.com/IntelPython/dpnp/pull/2263)
* Extended `dpnp.sort` and `dpnp.argsort` functions and `dpnp.ndarray.sort` and `dpnp.ndarray.argsort` methods to support `descending` keyword [2269](https://github.com/IntelPython/dpnp/pull/2269)
* Extended `dpnp.std` and `dpnp.var` functions and `dpnp.ndarray.std` and `dpnp.ndarray.var` methods to support `mean` keyword [2271](https://github.com/IntelPython/dpnp/pull/2271)
* Aligned `qr`, `eig`, `eigh`, `svd` and `slogdet` functions from `dpnp.linalg` to return namedtuple per Python array API [2276](https://github.com/IntelPython/dpnp/pull/2276)
* Extended `dpnp.std` and `dpnp.var` functions and `dpnp.ndarray.std` and `dpnp.ndarray.var` methods to support `correction` keyword [2300](https://github.com/IntelPython/dpnp/pull/2300)
* Extended `dpnp.cov` function to support all keyword arguments [2303](https://github.com/IntelPython/dpnp/pull/2303)
* Disallowed `minlength=None` value passed into `dpnp.bincount` function [2310](https://github.com/IntelPython/dpnp/pull/2310)
* Added build support with `oneMath` (new name of `oneMKL` interface) [2313](https://github.com/IntelPython/dpnp/pull/2313)
* Aligned the signature of `dpnp.astype` function with Python array API [2318](https://github.com/IntelPython/dpnp/pull/2318)

Fixed

* Migrated to experimental extension of DPC++ compiler with `group_load/group_store` per deprecation build warning [2123](https://github.com/IntelPython/dpnp/pull/2123)
* Fixed `DeprecationWarning` appearing during running the tests [2156](https://github.com/IntelPython/dpnp/pull/2156)
* Modified installation path for the tests to enable tests run by `pytest --pyargs dpnp.tests` command [2116](https://github.com/IntelPython/dpnp/pull/2116)
* Resolved issues with `dpnp.linalg.svd` implementation invoked on NVIDIA GPU [2212](https://github.com/IntelPython/dpnp/pull/2212)
* Fixed compilation warnings when building from the source with `--target=cuda` option [2225](https://github.com/IntelPython/dpnp/pull/2225)
* Passed `striped` property to `group_load/group_store` functions replicating legacy behavior [2238](https://github.com/IntelPython/dpnp/pull/2238)
* Added a workaround to prevent a race condition in `dpnp.linalg.qr` when running on NVIDIA GPU [2265](https://github.com/IntelPython/dpnp/pull/2265)
* Resolved `IndexError` exception raised by `dpnp.matmul` [2278](https://github.com/IntelPython/dpnp/pull/2278)
* Declared missing required dependencies on numpy and dpctl packages from dpnp wheel package [2283](https://github.com/IntelPython/dpnp/pull/2283)
* Resolved an issue with wrong result shape returned by `dpnp.vecdot` [2294](https://github.com/IntelPython/dpnp/pull/2294)
* Resolved an issue with wrong result returned by `dpnp.tensordot` for integer data types [2296](https://github.com/IntelPython/dpnp/pull/2296)
* Resolved `ValueError` exception raised by `dpnp.linalg.qr` with non-contiguous input array [2314](https://github.com/IntelPython/dpnp/pull/2314)
* Resolved an issue with wrong result returned by `dpnp.fft.fftn` and `dpnp.fft.rfftn` when running on NVIDIA GPU [2332](https://github.com/IntelPython/dpnp/pull/2332)
* Added a workaround to prevent a memory corruption in `dpnp.correlate` [2333](https://github.com/IntelPython/dpnp/pull/2333)

Fixed

* Bumped min version of DPC++ compiler required to support experimental SYCL properties [2231](https://github.com/IntelPython/dpnp/pull/2231)

Fixed

* Enabled `dpnp` in virtual environment on Windows platform [2242](https://github.com/IntelPython/dpnp/pull/2242)

This is a bug-fix release.

Changed

* Changed to use `Miniforge` installer in GitHub actions [2057](https://github.com/IntelPython/dpnp/pull/2057)
* Updated `README.md` to reflect current installation requirements and available options [2166](https://github.com/IntelPython/dpnp/pull/2166)
* Corrected the list of owners and code maintainers [2185](https://github.com/IntelPython/dpnp/pull/2185)
* Bumped the version of `oneMKL` interface used in dpnp build by default to align it with `2025.0` oneAPI release [2193](https://github.com/IntelPython/dpnp/pull/2193)

Fixed

* Resolved an issue with Compute Follows Data inconsistency in `dpnp.extract` function [2172](https://github.com/IntelPython/dpnp/pull/2172)
* Resolved an import error when using `dpnp` in virtual environment on Linux [2199](https://github.com/IntelPython/dpnp/pull/2199)
* Fixed incorrect result produced by `dpnp.fft.fft` function when input array has negative strides [2202](https://github.com/IntelPython/dpnp/pull/2202)
* Fixed an issue with `numpy.ndarray` input processing in the `dpnp.from_dlpack` function and updated the documentation [2209](https://github.com/IntelPython/dpnp/pull/2209)
* Resolved a compilation error when building with DPC++ 2025.1 compiler [2211](https://github.com/IntelPython/dpnp/pull/2211)

This release reaches an important milestone by making offloading fully asynchronous. Calls to `dpnp` submit tasks for execution to DPC++ runtime and return without waiting for execution of these tasks to finish. The sequential semantics a user comes to expect from execution of Python script is preserved though.
In addition, this release completes implementation of `dpnp.fft` module and adds several new array manipulation, indexing and elementwise routines. Moreover, it adds support to build `dpnp` for Nvidia GPUs.

Added

* Added implementation of `dpnp.gradient` function [1859](https://github.com/IntelPython/dpnp/pull/1859)
* Added implementation of `dpnp.sort_complex` function [1864](https://github.com/IntelPython/dpnp/pull/1864)
* Added implementation of `dpnp.fft.fft` and `dpnp.fft.ifft` functions [1879](https://github.com/IntelPython/dpnp/pull/1879)
* Added implementation of `dpnp.isneginf` and `dpnp.isposinf` functions [1888](https://github.com/IntelPython/dpnp/pull/1888)
* Added implementation of `dpnp.fft.fftfreq` and `dpnp.fft.rfftfreq` functions [1898](https://github.com/IntelPython/dpnp/pull/1898)
* Added implementation of `dpnp.fft.fftshift` and `dpnp.fft.ifftshift` functions [1900](https://github.com/IntelPython/dpnp/pull/1900)
* Added implementation of `dpnp.isreal`, `dpnp.isrealobj`, `dpnp.iscomplex`, and `dpnp.iscomplexobj` functions [1916](https://github.com/IntelPython/dpnp/pull/1916)
* Added support to build `dpnp` for Nvidia GPU [1926](https://github.com/IntelPython/dpnp/pull/1926)
* Added implementation of `dpnp.fft.rfft` and `dpnp.fft.irfft` functions [1928](https://github.com/IntelPython/dpnp/pull/1928)
* Added implementation of `dpnp.nextafter` function [1938](https://github.com/IntelPython/dpnp/pull/1938)
* Added implementation of `dpnp.trim_zero` function [1941](https://github.com/IntelPython/dpnp/pull/1941)
* Added implementation of `dpnp.fft.hfft` and `dpnp.fft.ihfft` functions [1954](https://github.com/IntelPython/dpnp/pull/1954)
* Added implementation of `dpnp.logaddexp2` function [1955](https://github.com/IntelPython/dpnp/pull/1955)
* Added implementation of `dpnp.flatnonzero` function [1956](https://github.com/IntelPython/dpnp/pull/1956)
* Added implementation of `dpnp.float_power` function [1957](https://github.com/IntelPython/dpnp/pull/1957)
* Added implementation of `dpnp.fft.fft2`, `dpnp.fft.ifft2`, `dpnp.fft.fftn`, and `dpnp.fft.ifftn` functions [1961](https://github.com/IntelPython/dpnp/pull/1961)
* Added implementation of `dpnp.array_equal` and `dpnp.array_equiv` functions [1965](https://github.com/IntelPython/dpnp/pull/1965)
* Added implementation of `dpnp.nan_to_num` function [1966](https://github.com/IntelPython/dpnp/pull/1966)
* Added implementation of `dpnp.fix` function [1971](https://github.com/IntelPython/dpnp/pull/1971)
* Added implementation of `dpnp.fft.rfft2`, `dpnp.fft.irfft2`, `dpnp.fft.rfftn`, and `dpnp.fft.irfftn` functions [1982](https://github.com/IntelPython/dpnp/pull/1982)
* Added implementation of `dpnp.argwhere` function [2000](https://github.com/IntelPython/dpnp/pull/2000)
* Added implementation of `dpnp.real_if_close` function [2002](https://github.com/IntelPython/dpnp/pull/2002)
* Added implementation of `dpnp.ndim` and `dpnp.size` functions [2014](https://github.com/IntelPython/dpnp/pull/2014)
* Added implementation of `dpnp.append` and `dpnp.asarray_chkfinite` functions [2015](https://github.com/IntelPython/dpnp/pull/2015)
* Added implementation of `dpnp.array_split`, `dpnp.split`, `dpnp.hsplit`, `dpnp.vsplit`, and `dpnp.dsplit` functions [2017](https://github.com/IntelPython/dpnp/pull/2017)
* Added runtime dependency on `intel-gpu-ocl-icd-system` package [2023](https://github.com/IntelPython/dpnp/pull/2023)
* Added implementation of `dpnp.ravel_multi_index` and `dpnp.unravel_index` functions [2022](https://github.com/IntelPython/dpnp/pull/2022)
* Added implementation of `dpnp.resize` and `dpnp.rot90` functions [2030](https://github.com/IntelPython/dpnp/pull/2030)
* Added implementation of `dpnp.require` function [2036](https://github.com/IntelPython/dpnp/pull/2036)

Changed

* Extended pre-commit pylint check to `dpnp.fft` module [1860](https://github.com/IntelPython/dpnp/pull/1860)
* Reworked `vm` vector math backend to reuse `dpctl.tensor` functions around unary and binary functions [1868](https://github.com/IntelPython/dpnp/pull/1868)
* Extended `dpnp.ndarray.astype` method to support `device` keyword argument [1870](https://github.com/IntelPython/dpnp/pull/1870)
* Improved performance of `dpnp.linalg.solve` by implementing a dedicated kernel for its batch implementation [1877](https://github.com/IntelPython/dpnp/pull/1877)
* Extended `dpnp.fabs` to support `order` and `out` keyword arguments by writing a dedicated kernel for it [1878](https://github.com/IntelPython/dpnp/pull/1878)
* Extended `dpnp.linalg` module to support `usm_ndarray` as input [1880](https://github.com/IntelPython/dpnp/pull/1880)
* Reworked `dpnp.mod` implementation to be an alias for `dpnp.remainder` [1882](https://github.com/IntelPython/dpnp/pull/1882)
* Removed the legacy implementation of linear algebra functions from the backend [1887](https://github.com/IntelPython/dpnp/pull/1887)
* Removed the legacy implementation of elementwise functions from the backend [1890](https://github.com/IntelPython/dpnp/pull/1890)
* Extended `dpnp.all` and `dpnp.any` to support `out` keyword argument [1893](https://github.com/IntelPython/dpnp/pull/1893)
* Reworked `dpnp.repeat` to add a explicit type check of input array [1894](https://github.com/IntelPython/dpnp/pull/1894)
* Improved performance of different functions by adopting asynchronous implementation of `dpctl` [1897](https://github.com/IntelPython/dpnp/pull/1897)
* Extended `dpnp.fmax` and `dpnp.fmin` to support `order` and `out` keyword arguments by writing dedicated kernels for them [1905](https://github.com/IntelPython/dpnp/pull/1905)
* Removed the legacy implementation of array creation and manipulation functions from the backend [1903](https://github.com/IntelPython/dpnp/pull/1903)
* Extended `dpnp.extract` implementation to align with NumPy [1906](https://github.com/IntelPython/dpnp/pull/1906)
* Reworked backend implementation to align with non-backward compatible changes in DPC++ 2025.0 [1907](https://github.com/IntelPython/dpnp/pull/1907)
* Removed the legacy implementation of indexing functions from the backend [1908](https://github.com/IntelPython/dpnp/pull/1908)
* Extended `dpnp.take` implementation to align with NumPy [1909](https://github.com/IntelPython/dpnp/pull/1909)
* Extended `dpnp.place` implementation to align with NumPy [1912](https://github.com/IntelPython/dpnp/pull/1912)
* Reworked the implementation of indexing functions to avoid unnecessary casting to `dpnp_array` when input is `usm_ndarray` [1913](https://github.com/IntelPython/dpnp/pull/1913)
* Reduced code duplication in the implementation of sorting functions [1914](https://github.com/IntelPython/dpnp/pull/1914)
* Removed the obsolete dparray interface [1915](https://github.com/IntelPython/dpnp/pull/1915)
* Improved performance of `dpnp.linalg` module for BLAS routines by adopting asynchronous implementation of `dpctl` [1919](https://github.com/IntelPython/dpnp/pull/1919)
* Relocated `dpnp.einsum` utility functions to a separate file [1920](https://github.com/IntelPython/dpnp/pull/1920)
* Improved performance of `dpnp.linalg` module for LAPACK routines by adopting asynchronous implementation of `dpctl` [1922](https://github.com/IntelPython/dpnp/pull/1922)
* Reworked `dpnp.matmul` to allow larger batch size to be used [1927](https://github.com/IntelPython/dpnp/pull/1927)
* Removed data synchronization where it is not needed [1930](https://github.com/IntelPython/dpnp/pull/1930)
* Leveraged `dpctl.tensor` implementation for `dpnp.where` to support scalar as input [1932](https://github.com/IntelPython/dpnp/pull/1932)
* Improved performance of `dpnp.linalg.eigh` by implementing a dedicated kernel for its batch implementation [1936](https://github.com/IntelPython/dpnp/pull/1936)
* Reworked `dpnp.isclose` and `dpnp.allclose` to comply with compute follows data approach [1937](https://github.com/IntelPython/dpnp/pull/1937)
* Extended `dpnp.deg2rad` and `dpnp.radians` to support `order` and `out` keyword arguments by writing dedicated kernels for them [1943](https://github.com/IntelPython/dpnp/pull/1943)
* `dpnp` uses pybind11 2.13.1 [1944](https://github.com/IntelPython/dpnp/pull/1944)
* Extended `dpnp.degrees` and `dpnp.rad2deg` to support `order` and `out` keyword arguments by writing dedicated kernels for them [1949](https://github.com/IntelPython/dpnp/pull/1949)
* Extended `dpnp.unwrap` to support all keyword arguments provided by NumPy [1950](https://github.com/IntelPython/dpnp/pull/1950)
* Leveraged `dpctl.tensor` implementation for `dpnp.count_nonzero` function [1962](https://github.com/IntelPython/dpnp/pull/1962)
* Leveraged `dpctl.tensor` implementation for `dpnp.diff` function [1963](https://github.com/IntelPython/dpnp/pull/1963)
* Leveraged `dpctl.tensor` implementation for `dpnp.take_along_axis` function [1969](https://github.com/IntelPython/dpnp/pull/1969)
* Reworked `dpnp.ediff1d` implementation through existing functions instead of a separate kernel [1970](https://github.com/IntelPython/dpnp/pull/1970)
* Reworked `dpnp.unique` implementation through existing functions when `axis` is given otherwise through leveraging `dpctl.tensor` implementation [1972](https://github.com/IntelPython/dpnp/pull/1972)
* Improved performance of `dpnp.linalg.svd` by implementing a dedicated kernel for its batch implementation [1936](https://github.com/IntelPython/dpnp/pull/1936)
* Leveraged `dpctl.tensor` implementation for `shape.setter` method [1975](https://github.com/IntelPython/dpnp/pull/1975)
* Extended `dpnp.ndarray.copy` to support compute follow data keyword arguments [1976](https://github.com/IntelPython/dpnp/pull/1976)
* Reworked `dpnp.select` implementation through existing functions instead of a separate kernel [1977](https://github.com/IntelPython/dpnp/pull/1977)
* Leveraged `dpctl.tensor` implementation for `dpnp.from_dlpack` and `dpnp.ndarray.__dlpack__` functions [1980](https://github.com/IntelPython/dpnp/pull/1980)
* Reworked `dpnp.linalg` module backend implementation for BLAS rouitnes to work with OneMKL interfaces [1981](https://github.com/IntelPython/dpnp/pull/1981)
* Reworked `dpnp.ediff1d` implementation to reduce code duplication [1983](https://github.com/IntelPython/dpnp/pull/1983)
* `dpnp` can be used with any NumPy from 1.23 to 2.0 [1985](https://github.com/IntelPython/dpnp/pull/1985)
* Reworked `dpnp.unique` implementation to properly handle NaNs values [1972](https://github.com/IntelPython/dpnp/pull/1972)
* Removed `dpnp.issubcdtype` per NumPy 2.0 recommendation [1996](https://github.com/IntelPython/dpnp/pull/1996)
* Reworked `dpnp.unique` implementation to align with NumPy 2.0 [1999](https://github.com/IntelPython/dpnp/pull/1999)
* Reworked `dpnp.linalg.solve` backend implementation to work with OneMKL Interfaces [2001](https://github.com/IntelPython/dpnp/pull/2001)
* Reworked `dpnp.trapezoid` implementation through existing functions instead of falling back on NumPy [2003](https://github.com/IntelPython/dpnp/pull/2003)
* Added `copy` keyword to `dpnp.array` to align with NumPy 2.0 [2006](https://github.com/IntelPython/dpnp/pull/2006)
* Extended `dpnp.heaviside` to support `order` and `out` keyword arguments by writing dedicated kernel for it [2008](https://github.com/IntelPython/dpnp/pull/2008)
* `dpnp` uses pybind11 2.13.5 [2010](https://github.com/IntelPython/dpnp/pull/2010)
* Added `COMPILER_VERSION_2025_OR_LATER` flag to be able to run `dpnp.fft` module with both 2024.2 and 2025.0 versions of the compiler [2025](https://github.com/IntelPython/dpnp/pull/2025)
* Cleaned up an implementation of `dpnp.gradient` by removing obsolete TODO which is not going to be done [2032](https://github.com/IntelPython/dpnp/pull/2032)
* Updated `Array Manipulation Routines` page in documentation to add missing functions and to remove duplicate entries [2033](https://github.com/IntelPython/dpnp/pull/2033)
* `dpnp` uses pybind11 2.13.6 [2041](https://github.com/IntelPython/dpnp/pull/2041)
* Updated `dpnp.fft` backend to depend on `INTEL_MKL_VERSION` flag to ensures that the appropriate code segment is executed based on the version of OneMKL [2035](https://github.com/IntelPython/dpnp/pull/2035)
* Use `dpctl::tensor::alloc_utils::sycl_free_noexcept` instead of `sycl::free` in `host_task` tasks associated with life-time management of temporary USM allocations [2058](https://github.com/IntelPython/dpnp/pull/2058)
* Improved implementation of `dpnp.kron` to avoid unnecessary copy for non-contiguous arrays [2059](https://github.com/IntelPython/dpnp/pull/2059)
* Updated the test suit for `dpnp.fft` module [2071](https://github.com/IntelPython/dpnp/pull/2071)
* Reworked `dpnp.clip` implementation to align with Python Array API 2023.12 specification [2048](https://github.com/IntelPython/dpnp/pull/2048)
* Skipped outdated tests for `dpnp.linalg.solve` due to compatibility issues with NumPy 2.0 [2074](https://github.com/IntelPython/dpnp/pull/2074)
* Updated installation instructions [2098](https://github.com/IntelPython/dpnp/pull/2098)

Fixed

* Resolved an issue with `dpnp.matmul` when an f_contiguous `out` keyword is passed to the the function [1872](https://github.com/IntelPython/dpnp/pull/1872)
* Resolved a possible race condition in `dpnp.inv` [1940](https://github.com/IntelPython/dpnp/pull/1940)
* Resolved an issue with failing tests for `dpnp.append` when running on a device without fp64 support [2034](https://github.com/IntelPython/dpnp/pull/2034)
* Resolved an issue with input array of `usm_ndarray` passed into `dpnp.ix_` [2047](https://github.com/IntelPython/dpnp/pull/2047)
* Added a workaround to prevent crash in tests on Windows in internal CI/CD (when running on either Lunar Lake or Arrow Lake) [2062](https://github.com/IntelPython/dpnp/pull/2062)
* Fixed a crash in `dpnp.choose` caused by missing control of releasing temporary allocated device memory [2063](https://github.com/IntelPython/dpnp/pull/2063)
* Resolved compilation warning and error while building in debug mode [2066](https://github.com/IntelPython/dpnp/pull/2066)
* Fixed an issue with asynchronous execution in `dpnp.fft` module [2067](https://github.com/IntelPython/dpnp/pull/2067)