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

  

• 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

  

• Add third-party test for cumsum with axis=None and out

  Unread
  IntelPython/dpnp#3088·antonwolfy opened Oct 7, 2026 · Updated Oct 7, 2026
  3

  

• Fix kernel launches over more than INT_MAX work-items with DPC++ 2026.2

  Unread
  IntelPython/dpnp#3089·antonwolfy opened Oct 7, 2026 · Updated Oct 7, 2026
  3

  

• Use computation dtype in dpnp.linalg.pinv for default tolerance and empty input

  Unread
  IntelPython/dpnp#3090·antonwolfy opened Oct 7, 2026 · Updated Oct 7, 2026
  3