Merge origin into update_isclose

Author: vlad-perevezentsev

CHANGELOG.md

@@ -34,6 +34,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 * Replaced `ci` section in `.pre-commit-config.yaml` with a new GitHub workflow with scheduled run to autoupdate the `pre-commit` configuration [#2542](https://github.com/IntelPython/dpnp/pull/2542)
 * FFT module is updated to perform in-place FFT in intermediate steps of ND FFT [#2543](https://github.com/IntelPython/dpnp/pull/2543)
 * Reused dpctl tensor include to enable experimental SYCL namespace for complex types [#2546](https://github.com/IntelPython/dpnp/pull/2546)
+* Changed Windows-specific logic in dpnp initialization [#2553](https://github.com/IntelPython/dpnp/pull/2553)
 * Improved performance of `dpnp.isclose` function by implementing a dedicated kernel for scalar `rtol` and `atol` arguments [#2540](https://github.com/IntelPython/dpnp/pull/2540)
 
 ### Deprecated

doc/reference/linalg.rst

@@ -43,6 +43,7 @@ Decompositions
    dpnp.linalg.cholesky
    dpnp.linalg.outer
    dpnp.linalg.qr
+   dpnp.linalg.lu_factor
    dpnp.linalg.svd
    dpnp.linalg.svdvals
 

dpnp/__init__.py

@@ -40,17 +40,13 @@
 # where to search for DLLs towards both DPNP backend and DPCTL Sycl interface,
 # otherwise DPNP import will be failing. This is because the libraries
 # are not installed under any of default paths where Python is searching.
-from platform import system
-
-if system() == "Windows":  # pragma: no cover
-    if hasattr(os, "add_dll_directory"):
-        os.add_dll_directory(mypath)
-        os.add_dll_directory(dpctlpath)
 
+if sys.platform == "win32":  # pragma: no cover
+    os.add_dll_directory(mypath)
+    os.add_dll_directory(dpctlpath)
     os.environ["PATH"] = os.pathsep.join(
         [os.getenv("PATH", ""), mypath, dpctlpath]
     )
-
     # For virtual environments on Windows, add folder with DPC++ libraries
     # to the DLL search path
     if sys.base_exec_prefix != sys.exec_prefix and os.path.isfile(

dpnp/linalg/dpnp_iface_linalg.py

@@ -914,38 +914,39 @@ def lu_factor(a, overwrite_a=False, check_finite=True):
     where `P` is a permutation matrix, `L` is lower triangular with unit
     diagonal elements, and `U` is upper triangular.
 
+    For full documentation refer to :obj:`scipy.linalg.lu_factor`.
+
     Parameters
     ----------
-    a : (M, N) {dpnp.ndarray, usm_ndarray}
+    a : (..., M, N) {dpnp.ndarray, usm_ndarray}
         Input array to decompose.
     overwrite_a : {None, bool}, optional
-        Whether to overwrite data in `a` (may increase performance)
+        Whether to overwrite data in `a` (may increase performance).
+
         Default: ``False``.
     check_finite : {None, bool}, optional
         Whether to check that the input matrix contains only finite numbers.
         Disabling may give a performance gain, but may result in problems
         (crashes, non-termination) if the inputs do contain infinities or NaNs.
 
+        Default: ``True``.
+
     Returns
     -------
-    lu :(M, N) dpnp.ndarray
-        Matrix containing U in its upper triangle, and L in its lower triangle.
-        The unit diagonal elements of L are not stored.
-    piv (K, ): dpnp.ndarray
-        Pivot indices representing the permutation matrix P:
+    lu : (..., M, N) dpnp.ndarray
+        Matrix containing `U` in its upper triangle,
+        and `L` in its lower triangle.
+        The unit diagonal elements of `L` are not stored.
+    piv : (..., K) dpnp.ndarray
+        Pivot indices representing the permutation matrix `P`:
         row i of matrix was interchanged with row piv[i].
-        ``K = min(M, N)``.
+        Where ``K = min(M, N)``.
 
     Warning
     -------
     This function synchronizes in order to validate array elements
     when ``check_finite=True``.
 
-    Limitations
-    -----------
-    Only two-dimensional input matrices are supported.
-    Otherwise, the function raises ``NotImplementedError`` exception.
-
     Examples
     --------
     >>> import dpnp as np

dpnp/linalg/dpnp_utils_linalg.py

@@ -481,7 +481,7 @@ def _batched_lu_factor_scipy(a, res_type):  # pylint: disable=too-many-locals
         if any(dev_info_h):
             diag_nums = ", ".join(str(v) for v in dev_info_h if v > 0)
             warn(
-                f"Diagonal number {diag_nums} are exactly zero. "
+                f"Diagonal numbers {diag_nums} are exactly zero. "
                 "Singular matrix.",
                 RuntimeWarning,
                 stacklevel=2,
@@ -2463,17 +2463,18 @@ def dpnp_lu_factor(a, overwrite_a=False, check_finite=True):
     # - not writeable
     if not overwrite_a or _is_copy_required(a, res_type):
         a_h = dpnp.empty_like(a, order="F", dtype=res_type)
-        ht_ev, copy_ev = ti._copy_usm_ndarray_into_usm_ndarray(
+        ht_ev, dep_ev = ti._copy_usm_ndarray_into_usm_ndarray(
             src=a_usm_arr,
             dst=a_h.get_array(),
             sycl_queue=a_sycl_queue,
             depends=_manager.submitted_events,
         )
-        _manager.add_event_pair(ht_ev, copy_ev)
+        _manager.add_event_pair(ht_ev, dep_ev)
+        dep_ev = [dep_ev]
     else:
         # input is suitable for in-place modification
         a_h = a
-        copy_ev = None
+        dep_ev = _manager.submitted_events
 
     m, n = a.shape
 
@@ -2493,14 +2494,14 @@ def dpnp_lu_factor(a, overwrite_a=False, check_finite=True):
         a_h.get_array(),
         ipiv_h.get_array(),
         dev_info_h,
-        depends=[copy_ev] if copy_ev is not None else [],
+        depends=dep_ev,
     )
     _manager.add_event_pair(ht_ev, getrf_ev)
 
     if any(dev_info_h):
         diag_nums = ", ".join(str(v) for v in dev_info_h if v > 0)
         warn(
-            f"Diagonal number {diag_nums} are exactly zero. Singular matrix.",
+            f"Diagonal number {diag_nums} is exactly zero. Singular matrix.",
             RuntimeWarning,
             stacklevel=2,
         )

dpnp/tests/test_linalg.py

@@ -1911,7 +1911,7 @@ def test_lu_factor(self, shape, order, dtype):
         A_cast = a_dp.astype(LU.dtype, copy=False)
         PA = self._apply_pivots_rows(A_cast, piv)
 
-        assert_allclose(LU, PA, rtol=1e-6, atol=1e-6)
+        assert dpnp.allclose(LU, PA, rtol=1e-6, atol=1e-6)
 
     @pytest.mark.parametrize("dtype", get_float_complex_dtypes())
     def test_overwrite_inplace(self, dtype):
@@ -1928,7 +1928,7 @@ def test_overwrite_inplace(self, dtype):
         PA = self._apply_pivots_rows(a_dp_orig, piv)
         LU = L @ U
 
-        assert_allclose(LU, PA, rtol=1e-6, atol=1e-6)
+        assert dpnp.allclose(LU, PA, rtol=1e-6, atol=1e-6)
 
     @pytest.mark.parametrize("dtype", get_float_complex_dtypes())
     def test_overwrite_copy(self, dtype):
@@ -1945,7 +1945,7 @@ def test_overwrite_copy(self, dtype):
         PA = self._apply_pivots_rows(a_dp_orig, piv)
         LU = L @ U
 
-        assert_allclose(LU, PA, rtol=1e-6, atol=1e-6)
+        assert dpnp.allclose(LU, PA, rtol=1e-6, atol=1e-6)
 
     def test_overwrite_copy_special(self):
         # F-contig but dtype != res_type
@@ -1972,7 +1972,7 @@ def test_overwrite_copy_special(self):
                 a_orig.astype(L.dtype, copy=False), piv
             )
             LU = L @ U
-            assert_allclose(LU, PA, rtol=1e-6, atol=1e-6)
+            assert dpnp.allclose(LU, PA, rtol=1e-6, atol=1e-6)
 
     @pytest.mark.parametrize("shape", [(0, 0), (0, 2), (2, 0)])
     def test_empty_inputs(self, shape):
@@ -2003,7 +2003,7 @@ def test_strided(self, sl):
         PA = self._apply_pivots_rows(a_dp, piv)
         LU = L @ U
 
-        assert_allclose(LU, PA, rtol=1e-6, atol=1e-6)
+        assert dpnp.allclose(LU, PA, rtol=1e-6, atol=1e-6)
 
     def test_singular_matrix(self):
         a_dp = dpnp.array([[1.0, 2.0], [2.0, 4.0]])
@@ -2070,7 +2070,7 @@ def test_lu_factor_batched(self, shape, order, dtype):
             L, U = self._split_lu(lu_3d[i], m, n)
             A_cast = a_3d[i].astype(L.dtype, copy=False)
             PA = self._apply_pivots_rows(A_cast, piv_2d[i])
-            assert_allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
+            assert dpnp.allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
 
     @pytest.mark.parametrize("dtype", get_float_complex_dtypes())
     @pytest.mark.parametrize("order", ["C", "F"])
@@ -2082,7 +2082,7 @@ def test_overwrite(self, dtype, order):
         )
 
         assert lu is not a_dp
-        assert_allclose(a_dp, a_dp_orig)
+        assert dpnp.allclose(a_dp, a_dp_orig)
 
         m = n = 2
         lu_3d = lu.reshape((-1, m, n))
@@ -2092,7 +2092,7 @@ def test_overwrite(self, dtype, order):
             L, U = self._split_lu(lu_3d[i], m, n)
             A_cast = a_3d[i].astype(L.dtype, copy=False)
             PA = self._apply_pivots_rows(A_cast, piv_2d[i])
-            assert_allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
+            assert dpnp.allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
 
     @pytest.mark.parametrize(
         "shape", [(0, 2, 2), (2, 0, 2), (2, 2, 0), (0, 0, 0)]
@@ -2119,7 +2119,7 @@ def test_strided(self):
             PA = self._apply_pivots_rows(
                 a_stride[i].astype(L.dtype, copy=False), piv[i]
             )
-            assert_allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
+            assert dpnp.allclose(L @ U, PA, rtol=1e-6, atol=1e-6)
 
     def test_singular_matrix(self):
         a = dpnp.zeros((3, 2, 2), dtype=dpnp.default_float_type())