Disable GitHub workflows to be run with python 3.13#2322
Merged
antonwolfy merged 4 commits intomasterfrom Feb 18, 2025
Merged
Conversation
Contributor
|
View rendered docs @ https://intelpython.github.io/dpnp/index.html |
Collaborator
Contributor
|
Array API standard conformance tests for dpnp=0.17.0dev6=py312he4f9c94_22 ran successfully. |
…tCopytoFromScalar to pass for all dtypes
antonwolfy
requested review from
AlexanderKalistratov,
ndgrigorian,
vlad-perevezentsev and
vtavana
as code owners
Contributor
Author
|
The night workflow passed here: https://github.com/IntelPython/dpnp/actions/runs/13398569974 |
github-actions bot
added a commit
that referenced
this pull request
Feb 18, 2025
There is currently no way to install dpnp with any numpy version > 2.0, because - there is no such version available on `https://software.repos.intel.com/python/conda/` channel - numpy from `conda-forge` channel depends on BLAS libraries which has a runtime dependency `mkl < 2025.0` (since BLAS packages with build number 29) There is draft PR ([gh-128](conda-forge/blas-feedstock#128)) to build BLAS with MKL version 2025.0, but it is blocked due to MKL issue ([gh-83](conda-forge/intel_repack-feedstock#83)). Meanwhile there was conda-forge patch ([gh-959](conda-forge/conda-forge-repodata-patches-feedstock#959)) merged recently to add `mkl < 2025.0` constraint to already published BLAS packaged on `conda-forge` with older build number. DPNP package can't fulfill the requirement `mkl < 2025.0`, because it would mandate dpnp build pinning with DPC++ compiler 2024.2. While there were non-backward compatible changes between 2024.2 and 2025.x versions, that is dpnp built with 2024.2 compiler can not be installed with any DPC++ RT 2025.x version, but it is required since targeted by the upcoming release. Thus the PR proposes to temporary disable or workaround all scenario where it is need to tests with numpy from `conda-forge` channel, including tests workflows for python 3.13 (per PR trigger and nightly) and workflow for testing of oneMath interfaces with the latest numpy. The changes has to be rolled back once the above issues are resolved. b09533e
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
There is currently no way to install dpnp with any numpy version > 2.0, because
https://software.repos.intel.com/python/conda/channelconda-forgechannel depends on BLAS libraries which has a runtime dependencymkl < 2025.0(since BLAS packages with build number 29)There is draft PR (gh-128) to build BLAS with MKL version 2025.0, but it is blocked due to MKL issue (gh-83).
Meanwhile there was conda-forge patch (gh-959) merged recently to add
mkl < 2025.0constraint to already published BLAS packaged onconda-forgewith older build number.DPNP package can't fulfill the requirement
mkl < 2025.0, because it would mandate dpnp build pinning with DPC++ compiler 2024.2. While there were non-backward compatible changes between 2024.2 and 2025.x versions, that is dpnp built with 2024.2 compiler can not be installed with any DPC++ RT 2025.x version, but it is required since targeted by the upcoming release.Thus the PR proposes to temporary disable or workaround all scenario where it is need to tests with numpy from
conda-forgechannel, including tests workflows for python 3.13 (per PR trigger and nightly) and workflow for testing of oneMath interfaces with the latest numpy.The changes has to be rolled back once the above issues are resolved.