Revert gh 2122 fix for gh 2121 #2158
Open
Conversation
This file contains hidden or 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
sycloid
requested review from
ndgrigorian,
antonwolfy and
vlad-perevezentsev
as code owners
|
Can one of the admins verify this patch? |
Also make Caster object static constexpr in generic copy kernel as it is done in contig kernel.
Since convert_impl input argument is a reference, implement NumPy's interpretation of bool (underlying byte has any bits set in it) to override C++'s interpretation (underlying byte has the first bit set). To allow such an intepretation to work correctly bool arguments must be passed by reference, rather than by value. Passing by value creates a copy where C++ masks higher bits out.
sycloid
force-pushed
the
revert-gh-2122-fix-for-gh-2121
branch
from
c3a147f to
cc0756f
Compare
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Outdated suggestions cannot be applied.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 reverts changes from gh-2122 and fixes gh-2121 in
dpctl::tensor::type_utils::convert_implinstead.Doing so improves performance of conversion from boolean arrays by avoiding temporary allocation and the cost of launching a kernel implementing
not_equaloperation.The test for gh-2121 exists in
test_usm_ndarray_ctor.py::test_astype_gh_2121.A comment #2121 (comment) was left in #2121 to observe that C++ behaves differently than NumPy and that this was the cause for behavior reported in #2121.