Skip to content
Navigation Menu
Sign in
Products
Platform
GitHub
Repository hosting, code review, and collaboration
Copilot
AI coding agent for tasks, projects, and automated work
Build & Ship
Actions
CI/CD and automation
Code Quality
Quality analysis, fixes, and enforcement
Codespaces
Cloud development environments
Secure
Code Security
Vulnerability detection and remediation
Secret Protection
Secret scanning and leak prevention
All Features
Developers
Learn
Documentation
Training
Certifications
Build
Build with Copilot
Customize Copilot
Integrations
Explore
Trending repos
GitHub Sponsors
Developer articles
Changelog
Changelog
Your source for the latest features, improvements, and fixes across GitHub products, all in one place.
Businesses
By Stage
Enterprises
Team
Startups
Nonprofits
By Industry
Healthcare
Financial services
Manufacturing
Government
See all industries
Evaluate GitHub
Why GitHub
Customer Stories
Business insights
Trust Center
Add-ons
GitHub Advanced Security
Enterprise-grade code security
Copilot for business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Resources
Learn & Connect
Documentation
Resource center
Blog
Events & webinars
Community Forum
Customer Stories
Integrations
Support
Customer support
Training
GitHub Status
Company
About
Partners
Changelog
GitHub Universe
GitHub Universe
The event for developers, by developers — talks, workshops, and what’s next for GitHub.
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
BackendStack21
mlx-runner
Repository navigation
Code
Issues
Pull requests
Actions
Projects
Security and quality
Insights
More
items
Performance: decode overlap, GQA-shared split attention, short-prompt GEMM tile (1.0.3)
- #3
#3
Merged
jkyberneees
merged 5 commits into
main
BackendStack21/mlx-runner:main
from
perf/decode-overlap-attn-split-short-prefill
BackendStack21/mlx-runner:perf/decode-overlap-attn-split-short-prefill
Copy head branch name to clipboard
Oct 31, 2026
Conversation
Commits
5
(5)
Checks
Files changed
Merged
Performance: decode overlap, GQA-shared split attention, short-prompt GEMM tile (1.0.3)
#3
jkyberneees
merged 5 commits into
main
BackendStack21/mlx-runner:main
from
perf/decode-overlap-attn-split-short-prefill
BackendStack21/mlx-runner:perf/decode-overlap-attn-split-short-prefill
Copy head branch name to clipboard
Commits
Commits on Oct 31, 2026
Decode: encode the next step's command buffer while the GPU runs the current one
Show description for cab4e28
jkyberneees
and
claude
committed
cab4e28
View commit details
Copy full SHA for cab4e28
Browse repository at this point
Short prompts (5-16 tokens) use an 8/16-token GEMM tile instead of padding to 32
Show description for d951187
jkyberneees
and
claude
committed
d951187
View commit details
Copy full SHA for d951187
Browse repository at this point
Decode attention: prefetch 4 keys per simdgroup; long-context split kernel shares K/V across GQA heads
Show description for b9e43cd
jkyberneees
and
claude
committed
b9e43cd
View commit details
Copy full SHA for b9e43cd
Browse repository at this point
Decode attention: keep the K/V prefetch off (AU = 1); measured no faster at short context
Show description for f72e812
jkyberneees
and
claude
committed
f72e812
View commit details
Copy full SHA for f72e812
Browse repository at this point
1.0.3: README performance table, CHANGELOG
Show description for 38e984d
jkyberneees
and
claude
committed
38e984d
View commit details
Copy full SHA for 38e984d
Browse repository at this point
You can’t perform that action at this time.