This release updates mkl_umath to be aligned with both numpy-1.26.x and numpy-2.x.x.
Added
- The definition of
signfunction for complex floating point data types is updated to match numpy-2.x.x gh-65 ldexpfunction is updated to allowint64explicitly similar to numpy-2.x.x behavior gh-73
Changed
- Migrated from
setup.pytopyproject tomlgh-63 - Changed to dynamic linking and added interface and threading layers gh-72
Fixed
- Fixed a bug for
mkl_umath.is_patchedfunction gh-66