0.21.0 release

Issues

• Set minimum supported NumPy version to 2.0.0

  Unread
  IntelPython/dpnp#2615·antonwolfy opened Oct 8, 2025 · Updated Apr 22, 2026·Draft
  2

  

• Feature sparse linalg solvers

  Unread
  IntelPython/dpnp#2841·abagusetty opened Apr 9, 2026 · Updated Oct 2, 2026
  160

• Integrate dpnp.tensor docs to dpnp

  Unread
  IntelPython/dpnp#2946·vlad-perevezentsev opened Jun 11, 2026 · Updated Jun 30, 2026
  2

  

• Speed up in-place vector-to-C-contiguous-matrix broadcast on CPU

  Unread
  IntelPython/dpnp#2981·vchamarthi opened Jun 30, 2026 · Updated Sep 30, 2026
  1

• task: coverity

  Unread
  IntelPython/dpnp#3070·jharlow-intel opened Sep 15, 2026 · Updated Sep 25, 2026
  4

  

• Fix dpnp.median/dpnp.nanmedian result shape for empty and size-1 kept dimensions

  Unread
  IntelPython/dpnp#3081·antonwolfy opened Oct 1, 2026 · Updated Oct 3, 2026
  4

  

• Support 'nuc' norm in dpnp.linalg.cond

  Unread
  IntelPython/dpnp#3084·antonwolfy opened Oct 2, 2026 · Updated Oct 3, 2026
  3