Obtain a proper result type on device without fp64 #1429
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
|
@antonwolfy please remove |
For the backend I'd like to keep that, since it might affect other dependent component. |
antonwolfy
added a commit
that referenced
this pull request
Jun 15, 2023
* Add a copy for strided arrays in dpnp.dot() * update versions * Rework transpose methods to use dpctl.tensor functions (#1389) * Rework transpose methods to call dpctl.tensor functions * Applied review comments & added more tests * Obtain a proper result type on device without fp64 (#1429) * Obtain a proper result type on device without fp64 * remove excess defenitions * Reuse dpctl.tensor.reshape (#1391) * Rework transpose methods to call dpctl.tensor functions * Reuse dpctl.tensor.reshape * added dpnp.shape() and unmuted more tests * fixed compiling issue & unmuted reshaper tests with order param * Resolve merge issues * resolve type mismatch on Win * Use MCG59 engine on GPU device (#1423) * Use MCG59 engine on GPU device * Fix issue for Windows * reduce precision in tests * move w/a before import dpctl * Reuse add(), multiply() and subtract() from dpctl (#1430) * Reuse add(), multiply() and subtract() from dpctl * add in-place support * Reuse dpctl.tensor.sum for dpnp.sum (#1426) * Reuse dpctl.tensor.sun for dpnp.sum * Update tests for dpnp.sum * Fix remarks * Update tests/third_party/cupy/testing/helper.py --------- Co-authored-by: Anton <100830759+antonwolfy@users.noreply.github.com> * Workaround to Klocwork (#1433) * Add inplace support of divide (#1434) * Add dpnp.result_type() support (#1435) * Add dpnp.result_type() support * Update dpnp/dpnp_iface_manipulation.py Co-authored-by: Natalia Polina <natalia.polina@intel.com> --------- Co-authored-by: Natalia Polina <natalia.polina@intel.com> * Implementation of dpnp.mean() (#1431) * Reuse dpctl.tensor.sun for dpnp.sum * Update tests for dpnp.sum * Fix remarks * Implementation of dpnp.mean * Update logic for dpnp.mean function * add normalize_axis_tuple * Additional tests for dpnp.mean * Fix minor remarks * Add inplace support of divide * Use inplace divide only for dpnp.inexact types * Update tests for dpnp.mean * Skip test_sample.py::TestRandint2::test_bound_float1 * Remove unused import * Update dtype check * Update dpnp/dpnp_iface_statistics.py * Return deleted skips --------- Co-authored-by: Anton Volkov <antonwolfy@gmail.com> Co-authored-by: Anton <100830759+antonwolfy@users.noreply.github.com> * New implementation of dpnp.outer (#1436) * Add a new implementation of dpnp.outer * Update dpnp.outer implementation --------- Co-authored-by: Anton <100830759+antonwolfy@users.noreply.github.com> * Set dpnp dependency on dpctl>=0.14.3 (#1437) * Implement dpnp.cov() though existing dpnp methods (#1396) * Implement dpnp.cov() though existing dpnp methods * Applied review comments * Clean up the code to get rid of todo * use dpnp.mean() * Added ChangeLog.md (#1439) * Added ChangeLog.md * exclude PR for comprasion ops * Update version to 0.12.0 (#1440) --------- Co-authored-by: vlad-perevezentsev <vladislav.perevezentsev@intel.com> Co-authored-by: Natalia Polina <natalia.polina@intel.com>
antonwolfy
added a commit
that referenced
this pull request
Jun 15, 2023
* Obtain a proper result type on device without fp64 * remove excess defenitions
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.
The PR fixes an issue how result type is calculated internally in cython layout on a device without fp64 support.
The issue comes from #1418.