Update dpnp.linalg.svd() function #1604
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
View rendered docs @ https://intelpython.github.io/dpnp/index.html |
vlad-perevezentsev
force-pushed
the
impl_svd
branch
from
c6667e1 to
9d5afde
Compare
antonwolfy
reviewed
Jan 23, 2024
|
|
||
| if compute_uv: | ||
| out_u = dpnp.array(u_matrices, order="F") | ||
| out_vt = dpnp.array(vt_matrices, order="F") |
There was a problem hiding this comment.
Need to return C-contiguous array to match the output of numpy.linalg.svd
| ht_lapack_ev.wait() | ||
| a_ht_copy_ev.wait() | ||
|
|
||
| if compute_uv: |
There was a problem hiding this comment.
Need to return C-contiguous array to match the output of numpy.linalg.svd
antonwolfy
approved these changes
Jan 24, 2024
github-actions bot
added a commit
to antonwolfy/dpnp
that referenced
this pull request
Feb 2, 2024
* Draft commit of dpnp.linalg.svd impl * Pass empty arrays if compute_uv=False * Add logic for the input array n < m * Add a new cupy test_decomposition * Rename gesvd input parameters * Correspondence of passed parameters to gesvd signature * Correct initialization of result variables in dpnp_svd * Update test_decomposition * Add implementation of _dpnp_svd_batch * Add test_decomposition to the scope of public CI * Improve error handling for mkl_lapack::gesvd function * Declate detail variable * Use a_usm_type and a_sycl_queue variables * Add additional checks for gesvd function * Remove old dpnp_svd backend * Refresh test_svd in test_linalg * Add detailed comments for gesvd arguments * gesvd returns pair of events and uses dpctl.utils.keep_args_alive * Keep a lexicographical order * Update docstrings for svd * Add test_svd to test_usm_type * Add a new impl to get s_type * Add a description for _stacked_identity * Simplify dpnp_svd_batch * Update tests for dpnp.linalg.svd * Add hermitian argument support * Add test_svd_hermitian * Update svd docstrings * Tune tolerance * Update test_svd_errors * Update _common_type and _common_inexact_type * Remove passing n and m parameteres to _gesvd * Simplify results return logic for dpnp_svd_batch * Update condition and random files in cupy/testing to use fix_random and repeat decorators * Rename cupy/testing/condition.py to .../_condition.py * Use self._tol in TestSvd * Update gesvd error handler * dpnp_svd works with F contiguous arrays * Add additional checks for output arrays * Impl parallel calculation in dpnp_svd_batch * Skip using @_condition.repeat in cupy tests * Add additional checks for output arrays * Update docstrings for svd * Use dpctl.SyclEvent.wait_for in dpnp_svd_batch * Add TODO : matching the order of returned arrays * Skip cupy tests on windows * Rename condition to _condition * Set setUpClass to skip cupy tests on cpu c7770fd
6 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR update
dpnp.linalg.svd()function to calculate the Singular Value Decomposition usingoneapi::mkl::lapack::gesvdThe implementation is written as a pybind11 extension above required LAPACK functions.
The changes are related to support of all types and support of function arguments as
full_matricesandcompute_uv