"The result is fat binaries, nearly gigabyte-sized wheels, and install pages that read like puzzle books. A coalition from NVIDIA, Astral, and QuantSight has been working on Wheel Next: A set of PEPs that let packages declare what hardware they need and let installers like uv pick the right build automatically."
"Jonathan Dekhtiar has been one of the driving forces behind the Wheel Next initiative and the associated PEPs for over a year, focusing on improving NVIDIA's CUDA and Python offering."
"Ralf Gommers, co-CEO of Quansight, has been maintaining foundational libraries like NumPy and SciPy since 2010 and created the PyPackaging Native guide to document core problems in native Python packaging."
"Charlie Marsh, founder of Astral, is the creator of uv, which aims to simplify the installation of Python packages by automatically selecting the right builds based on hardware requirements."
The Wheel Next initiative addresses limitations in Python package installations, particularly for packages with compiled code. Current installations often default to older CPU features, leading to large binaries and complex installation processes. The initiative, supported by NVIDIA, Astral, and QuantSight, proposes a set of PEPs that allow packages to specify hardware requirements. This enables installers to automatically select the appropriate builds, streamlining the installation process. The goal is to enhance support for modern optimizations like AVX2 and GPU configurations, simplifying user experience.
Read at Talkpython
Unable to calculate read time
Collection
[
|
...
]