Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix issues for numpy 2.1.0 #1804

Merged
merged 11 commits into from
Aug 21, 2024
Merged

Fix issues for numpy 2.1.0 #1804

merged 11 commits into from
Aug 21, 2024

Conversation

oleksandr-pavlyk
Copy link
Collaborator

Fixed test_full_gh_1230 to work with NumPy 2.1.0.

Fixed dpt.from_dlpack by handling ValueError possibly raised by ndarray.__dlpack__ when its dl_device
keyword argument is unsupported.

  • Have you provided a meaningful PR description?
  • Have you added a test, reproducer or referred to an issue with a reproducer?
  • Have you tested your changes locally for CPU and GPU devices?
  • Have you made sure that new changes do not introduce compiler warnings?
  • Have you checked performance impact of proposed changes?
  • If this PR is a work in progress, are you opening the PR as a draft?

@github-actions GitHub Actions
Copy link

github-actions bot commented Aug 19, 2024 •
edited
Loading

Deleted rendered PR docs from intelpython.github.com/dpctl, latest should be updated shortly. 🤞

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_329 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@coveralls
Copy link
Collaborator

coveralls commented Aug 19, 2024 •
edited
Loading

Coverage Status

coverage: 87.901% (-0.009%) from 87.91%
when pulling 5e6271d on fix-issues-for-numpy-2.1.0
into 1e52918 on master.

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_331 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@ndgrigorian
Copy link
Collaborator

@oleksandr-pavlyk
Unfortunately, it seems that changing _to_scalar has exposed some other, subtle overflows in other tests.

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_333 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_334 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_336 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_338 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_337 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@oleksandr-pavlyk
Correct test_full_gh_1230 for NumPy 2.1.0
6e81050 
NumPy 2.1.0 now raises exception when fill_value is not in
bounds for the array data type.

Rework the logic of the test to accomodate that. The test
continue to work with NumPy 2.0.1
@oleksandr-pavlyk
Adapt from_dlpack to NumPy 2.1.0 behavior
845382d 
NumPy 2.1.0 supports DLPack 1.0 now, and raises ValueError
if __dlpack__ method gets unsupport dl_device value.

Function from_dlpack now handles ValueError, the same way
as it handles BufferError
@oleksandr-pavlyk
Convert object for fill_value using asarray for requested type
1c17c34 
This relies on NumPy to raise OverflowError if the object can not
be represented by requested data type.
@oleksandr-pavlyk
Check that full(shape, fill_val, dtype=dt) handles out of bound integ…
abfee14 
…ral fill values

It should raise OverflowError, like NumPy 2 does.
@oleksandr-pavlyk
test_full_gh_1230 to be have differently dependending on NumPy version
a53ca06 
For 2.0.0 and later, it is to raise OverflowError, otherwise it is expected
to wrap up the out-of-bounds integral fill-in values given as Python scalars.
@oleksandr-pavlyk
Fix test_create_program_from_source to not use out-of-bounds fill-value
f441f01 
Restrict the input size to allow math the test relies on to work without
using out-of-bounds Python scalar inputs to tensor.arange
@oleksandr-pavlyk
Fix test_matmul_simple to avoid use of out-of-bounds Python scalars
032ceca 
Change the test so that input matrices that get multiplied only have
blocks of ones no larger than the max integer for the type, rest is
populated with zeros. This change applies to integral types only.
@oleksandr-pavlyk
Pin setuptools to <72.2 in hope to fix test_examples_linux workflow
2be4989
@oleksandr-pavlyk
Use Python 3.12 to run OS LLVM workflow
7c461f1
@oleksandr-pavlyk
Update components in collect coverage
04f8a4c 
Use Python 3.12, install recent NumPy (instead of 1.25),
update gtest from 1.13 to 1.15.2
@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_339 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

@github-actions GitHub Actions
Copy link

Array API standard conformance tests for dpctl=0.18.0dev0=py310hdf72452_340 ran successfully.
Passed: 894
Failed: 1
Skipped: 119

ndgrigorian
ndgrigorian approved these changes Aug 21, 2024
View reviewed changes
Copy link
Collaborator

@ndgrigorian ndgrigorian left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tests are back to passing, LGTM!

@oleksandr-pavlyk oleksandr-pavlyk merged commit 4297fef into master Aug 21, 2024
46 of 53 checks passed
@oleksandr-pavlyk oleksandr-pavlyk deleted the fix-issues-for-numpy-2.1.0 branch August 21, 2024 00:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@ndgrigorian ndgrigorian ndgrigorian approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@oleksandr-pavlyk @coveralls @ndgrigorian