Minutes_2022_09_06

Numba Meeting: 2022-09-06

Attendees: FPOC (last week):

NOTE: All communication is subject to the Numba Code of Conduct.

Please refer to this calendar for the next meeting date.

Release 0.56.2
- CUDA Caching fixes
- NumPy 1.23 support
- SciPy w/ NumPy 1.23 needs SciPy from conda-forge at present
deferred discussion https://github.com/numba/numba/pull/8396
- Defer again for next week
Awkward Arrays CUDA Numba functions - Graham Markall
Science Advances publication (https://www.science.org/doi/10.1126/sciadv.abm7935)
Changing the way we get LAPACK functions

(numba: 8407 llvmlite: 877)

#8408 - TypeError!
#8417 - Error flag NameError: global name 'length_of_iterator' is not defined with inline function

#8416 - ImportError: cannot import name 'set_parallel_chunksize' from 'numba.np.ufunc'

#8409 - fix compilation error on mac
- Is it related to use of system (non-conda) compilers?
- C++ compiler version?
- Error messages when building with #8389?
#8410 - Remove old warning about imports after SVML init
#8415 - Refactor numba.np.arraymath methods from lower_builtins to overloads
#8418 - Fixes ravel failure on 1d arrays (#5229)
#8419 - increment version support table entry for 0.56.* to 0.56.2
#8420 - Refactor np.sum to use @overload decorator and support for axis > 3

llvmlite: