@charliermarsh: It turns out that I forgot to enable hardware-accelerated SHA-256 for ARM. So uv will get much faster in the next relea…
Summary
uv will get faster in the next release after enabling hardware-accelerated SHA-256 for ARM, improving Trio install wall-time by up to 10%.
View Cached Full Text
Cached at: 07/30/26, 11:54 AM
It turns out that I forgot to enable hardware-accelerated SHA-256 for ARM. So uv will get much faster in the next release.
With it, wall-time for (e.g.) Trio installs improved by up to 10%. https://t.co/L6r4t44JH5
Similar Articles
@charliermarsh: We found a few optimizations that let us reduce the size of uv.lock files by "compressing" the environment markers. Som…
uv's lockfile size reduced by 40-50% through compression of environment markers.
@charliermarsh: Inspired by some code in Bun, I moved uv’s ZIP extraction into a single blocking task per archive (with synchronous fil…
uv's ZIP extraction has been optimized by moving it to a single blocking task with synchronous filesystem operations, inspired by code from Bun, leading to faster installation of packages like Jupyter from PyPI.
@charliermarsh: We found an optimization in uv's resolver that sped up Jupyter by 13%, and Boto3 by 2.4x (since it requires so much bac…
An optimization in uv's resolver improves performance for Jupyter by 13% and Boto3 by 2.4x, detailed in a write-up.
@charliermarsh: We now ship code-signed binaries for uv, for both macOS and Windows! This applies to our standalone installers and the …
uv now ships code-signed binaries for macOS and Windows, available in standalone installers and PyPI wheels, enhancing security and distribution.
Speeding up gearhash on ARM64 (2× faster)
The article describes optimizations to the `gearhash` Rust crate, adding a NEON backend that doubles performance on ARM64 architectures, improving content-defined chunking for applications like Hugging Face's Xet client.