This repository was archived by the owner on Jan 25, 2023. It is now read-only.
Update/numba 0.52.0 #175
Closed
Update/numba 0.52.0 #175
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
The f4 variant was failing because the input value (0.5425) was not exactly representable using a float32, and the value being used for the test was 0.542500019, which is not halfway between 0.542 and 0.543. The value is changed to 0.3425 to enable the test to pass. A variant testing float64 is added, which can succeed with the value 0.5425 as input when rounding to three digits.
As title. Adds lots of options for helping with CFG views.
This also required changing the lowering for isfinite, ldexp, and frexp to use the libdevice functions.
The lgamma implementation in Python differs to that in GNU libc to the point that the results can differ relatively by as much as ~4.22e-14. This causes a fail with cudasim, which is rectified by reducing the rtol down to 1e-13. This is still far tighter than the rtol originally used in these tests, which was 1e-5.
Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com>
This also fixes violations in `numba.cuda.kernels` and `numba.cuda.tests.cudasim`.
These tests are passing as of LLVM 10, except for the Unicode array one which was actually caused by numba#4563 anyways. Since that is the only remaining failure, go ahead and change the skip flag to point to numba#4563.
Adds support for 1.19. * Skips a tests which has problems on 1.19 (awaiting upstream patch for stability: numpy/numpy#17457) * Updates CI.
…_binops Support operators inlining in InlineOverloads
Extend Cudasim to support most of the memory functionality.
Remove Python 2 compatibility from numba.core.utils
Make find_callname only lookup functions that are likely part of NumPy.
Remove compatibility mode
Show channel URLs
Fix array analysis regression in 0.52 RC2 for tuple of 1D arrays
Fix ASCII flag in Unicode slicing (0.52.0rc2 regression)
Fix numba#6444: pruner issues with reference stealing functions
Fix asfarray kwarg default handling.
…_again Fix refprune on obfuscated refs and stabilize optimisation WRT wrappers.
Add skip on PPC64LE for tests causing SIGABRT in LLVM.
Update changelog for 0.52.0rc3
Preparing for 0.52.0rc3
fix abstract base class import
Restrict maximum version of python
Update changelog for 0.52.0rc3
Finalizing 0.52rc3
Fix requirements.txt pinning
setup.py: fix py version guard
Update changelog for 0.52.0 final
Preparing 0.52.0 final
Update llvmlite dependency
|
Hi @reazulhoque. I think we should do simpler: apply patches we did to |
|
@PokhodenkoSA I think that is a good suggestion, but we have a lot of patches and there are some commits that seems duplicate: 42c2477 and 7e503d2. I am not sure how this is supposed to be dealt with. Please advice. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
19 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.
This PR updates patched branch to Numba 0.52.0.