Skip to content

chore(release): 0.4.0 #43

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
asher merged 1 commit into from
Aug 23, 2026
Merged

chore(release): 0.4.0 #43

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,17 @@ adhere to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]

## [0.4.0]

### Changed
- Requires mlx 0.32.1 (was 0.31.2). This is the breaking change behind
the major-minor bump: environments pinned to mlx 0.31.x must stay on
mlx-kquant 0.3.x.
- The arena keeps a standalone Metal residency set. mlx 0.32 made its
queue-attached residency set private and budgeted by set_wired_limit;
the arena's GPU mappings now stay resident without consuming that
budget, so streaming-mode servers can run with the wired limit at zero.
- Builds against nanobind 2.13.0 (was 2.12.0).
- Quantized matmul at 4 to 8 rows is faster on machines with no NAX tile:
an 8-row split-K tile (one simdgroup along M, four along N) replaces the
16-row tile there, so a speculative verify of 8 rows pays for 8 rows of
Expand Down
2 changes: 1 addition & 1 deletion mlx_kquant/_version.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1 @@
__version__ = "0.3.13"
__version__ = "0.4.0"