ReleaseNVIDIANVIDIApublished Dec 18, 2025seen Jun 26

NVIDIA/cccl python-0.4.3

NVIDIA/cccl

Open original ↗

Captured source

source ↗
GH
GitHub/github.com/NVIDIA/cccl
NVIDIA/cccl python-0.4.3
Source ↗
published Dec 18, 2025seen Jun 26captured Aug 30http 200method plain

CCCL Python Libraries (v0.4.3)

Repository: NVIDIA/cccl

Tag: python-0.4.3

Published: 2025-12-18T14:54:50Z

Prerelease: no

Release notes: These are the release notes for the cuda-cccl Python package version 0.4.3, dated December 18th, 2025. The previous release was v0.4.2.

cuda.cccl is in "experimental" status, meaning that its API and feature set can change quite rapidly.

Installation

Please refer to the install instructions here

Features

-

Improvements and bug fixes

  • Add missing OpKind docs entries (https://github.com/NVIDIA/cccl/pull/6910)
  • Unify operator handling in cuda.compute (https://github.com/NVIDIA/cccl/pull/6938)
  • [cuda.compute] Refactor code for creating void* wrappers (https://github.com/NVIDIA/cccl/pull/6941)
  • Remove need for hardcoded LevelT for histogram in c.parallel and cuda.compute (https://github.com/NVIDIA/cccl/pull/6915)
  • c.parallel: reuse CUB agent policies for histogram (https://github.com/NVIDIA/cccl/pull/6974)
  • [cuda.compute]: fix alignment not being set properly for gpu_struct types (https://github.com/NVIDIA/cccl/pull/6995)