…rage` (#1654)

* implement dpnp.nanmean, dpnp.nanstd, dpnp.nansum and dpnp.average

* change size huge test case

* fix TestArithmetic test

* fix test nanmean

* address comments

* using assert_dtype_allclose

* add rtol to comparison

* Update tests/test_mathematical.py

---------

Co-authored-by: Anton <[email protected]>

* Add a new impl of dpnp.linalg._lu_factor

* Get dev_info_array after calling getrf

* Add an extra dev_info return to _lu_factor

* qwe

* Add a logic for a.ndim > 2 in _lu_factor

* Add an implementation of dpnp.linalg.slogdet

* Add a new test_norms.py file in cupy tests

* Expand test scope in public CI

* A small update _lu_factor func

* Remove w/a for dpnp.count_nonzero in slogdet

* getrf returns pair of events and uses dpctl.utils.keep_args_alive

* Update dpnp.linalg.det using slogdet

* Add new cupy tests for dpnp.linalg.det

* Add ipiv_vecs and dev_info_vecs in _lu_factor for the batch case

* Skip test_det on CPU due to bug in MKL

* Small update of cupy tests in test_norms.py

* Add support of complex dtype for dpnp.diagonal and update test_diagonal

* lu_factor func returns the result of LU decomposition as c-contiguous and add explanatory comments

* Add getrf_batch MKL extension

* Update docstring for slogdet

* Add more tests

* Remove accidentally added file

* Modify sign parameter calculation

* Remove the old backend implementation of dpnp_det

* qwe

* Keep lexographical order

* Add dpnp_slogdet to dpnp_utils_linalg

* Move _lu_factor above

* A minor update

* A minor changes for _lu_factor

* Remove trash files

* Use getrf_batch only on CPU

* Update tests for dpnp.linalg.slogdet

* Address remarks

* Add _real_type func

* Add test_det in test_usm_type

* Add more checks in getrf and getf_batch functions

* Improve error handler in getrf_impl

* Improve error handler in getrf_batch_impl

* dev_info is allocated as zeros

* Remove skipif for singular tests

* Implement _lu_factor logic with dev_info as a python list

* Update getrf_rf error handler with mkl_lapack::batch_error

* Remove passing n parameter to _getrf

* Add a new test_slogdet_singular_matrix_3D test

* Update tests for dpnp.linalg.det

* Use is_exception_caught flag in getrf and getrf_batch error handler

* Update gesv error handler

* Reshape results after calling getrf_batch

* Add a new dpnp.linalg.det impl and refresh dpnp_utils_linalg

* Remove Limitations from dpnp_det and dpnp_slogdet docstings

* Address remarks

* Remove det_dtype variable and use the abs val of diag for det

* Expand cupy tests for dpnp.linalg.det()

* Update TestDet and TestSlogdet

* Add new blas extension and update matmul impl

* Add support for N-D array

add N-dimension

* support more special cases + add new tests

* fix random behavior on cpu

* correct dtypes + support more keywords

* add strided support

* check input arrays

* address comments - first round

* address comments - second round

* address comments - third round

* fix pre-commit

* improve test coverage

* address comments

* update _gemm_res_dtype func

* fix a test for result_type

* fix minor issues

* skip tests for matmul

---------

Co-authored-by: Vahid Tavanashad <[email protected]>
Co-authored-by: vtavana <[email protected]>
Co-authored-by: Anton <[email protected]>

* update copyright date for 2024

* update doc/conf.py

* update dpnp/backend/cmake/Modules/TBBConfig.cmake

* Update dpnp.indices function

* Clean up dpnp_indices

* address comments

* address comments

* cancel changes for test_indices_list3

* Added description of all parameters

* Implement dpnp.mgrid and dpnp.ogrid function (#1633)

* Implement dpnp.mgrid and dpnp.ogrid function

* address comments

* address comments

* Skip dtype check for Iris Xe

* Update test_ranges.py

* Added description of all parameters

* Resolve pre-commit issue

---------

Co-authored-by: Anton <[email protected]>
Co-authored-by: Anton Volkov <[email protected]>

* update out keyword elementwise

* improve test coverage

* address comments

…tl` (#1662)

* update tests

update more tests

rename varibale

update test hypot

* including both type of erros in tests

* address comments

* Add a new impl of dpnp.linalg.cholesky

* Add cupy tests for dpnp.linalg.cholesky

* Add a batch impl of dpnp.linalg.cholesky

* Remove an old impl of dpnp_cholesky

* Remove DPNP_FN_CHOLESKY_EXT in dpnp_iface_fptr

* Remove dpnp_cholesky_ext_c

* Add a new _dpnp_cholesky_batch func

* Update test_cholesky in test_sycl_queue

* Expand test scope in public CI

* Add more tests for dpnp.linalg.cholesky

* Remove TODOs in cholesky() and update docstings

* Use _common_type in dpnp_cholesky

* Update dpnp_cholesky and dpnp_cholesky_batch

* Keep the lexicographic order

* Remove passing n parameter to _potrf

* Add additional checks to potrf and potrf_batch

* Extend potrf error handler

* Extend potrf_batch error handler

* Update tests for dpnp.linalg.cholesky

* Update license year

* Update cholesky docstrings

* Add support upper paramenetr for potrf

* Add support upper paramenetr for potrf_batch and update dpnp_cholesky

* Add tests for upper parameter of dpnp.linalg.cholesky

* Address remarks

* Fix validation check

---------

Co-authored-by: vtavana <[email protected]>

Co-authored-by: vtavana <[email protected]>

* implement dpnp.reciprocal and dpnp.angle

* fix docstring

* skip umath test for reciprocal

* address comments

* add a note for reciprocal

* Update dpnp/dpnp_iface_trigonometric.py

Co-authored-by: Anton <[email protected]>

* fix docstring

* update retunrs for angle and reciprocal

update out keyword implementation for angle and reciprocal to make it
consistent with implementation of other functions

* Updated TestReciprocal to be compliant with the latest available dpctl in internal CI

---------

Co-authored-by: Anton <[email protected]>
Co-authored-by: Anton Volkov <[email protected]>

* implement sort and argsort

* add more tests

* update for zero dimensional arrays

* address comments

* fix typo

* Allow for newer version of DPC++ compiler

* Stepped dpctl versions in all yaml files

* Impl dpnp.linalg.inv for 2d array

* Remove an old impl of dpnp_inv

* Add batch implementation of dpnp.linalg.inv func

* Add cupy tests for dpnp.linalg.inf

* Add dpnp tests for dpnp.linalg.inv

* Add check_lapack_dev_info func

* Add dev_info size check for getri_batch and getrf_batch

* Add size check dev_info and error_matrices_ids

* Remove dpnp_inv_ext_c

* Rename check_lapack_dev_info to _check_lapack_dev_info

* Skip test_inv in TestInvInvalid

---------

Co-authored-by: Anton <[email protected]>