Skip to content
Navigation Menu
Sign in
Appearance settings
Platform
AI CODE CREATION
GitHub Copilot
Write better code with AI
GitHub Copilot app
Direct agents from issue to merge
MCP Registry
Integrate external tools
DEVELOPER WORKFLOWS
Actions
Automate any workflow
Codespaces
Instant dev environments
Issues
Plan and track work
Code Review
Manage code changes
Code Quality
Enforce quality at merge
APPLICATION SECURITY
GitHub Advanced Security
Find and fix vulnerabilities
Code security
Secure your code as you build
Secret protection
Stop leaks before they start
EXPLORE
Why GitHub
Documentation
Blog
Changelog
Marketplace
View all features
Solutions
BY COMPANY SIZE
Enterprises
Small and medium teams
Startups
Nonprofits
BY USE CASE
App Modernization
DevSecOps
DevOps
CI/CD
View all use cases
BY INDUSTRY
Healthcare
Financial services
Manufacturing
Government
View all industries
View all solutions
Resources
EXPLORE BY TOPIC
AI
Software Development
DevOps
Security
View all topics
EXPLORE BY TYPE
Customer stories
Events & webinars
Ebooks & reports
Business insights
GitHub Skills
SUPPORT & SERVICES
Documentation
Customer support
Community forum
Trust center
Partners
View all resources
Open Source
COMMUNITY
GitHub Sponsors
Fund open source developers
PROGRAMS
Security Lab
Maintainer Community
GitHub Stars
Archive Program
REPOSITORIES
Topics
Trending
Collections
Enterprise
ENTERPRISE SOLUTIONS
Enterprise platform
AI-powered developer platform
AVAILABLE ADD-ONS
GitHub Advanced Security
Enterprise-grade security features
Copilot for Business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
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 }}
thecodacus
/
llama.cpp
Public
forked from
ggml-org/llama.cpp
Notifications
You must be signed in to change notification settings
Fork
73
Star
271
Code
Pull requests
4
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Pull requests
Actions
Projects
Security and quality
Insights
feat: TurboQuant KV-cache quantization — all backends
- #1
#1
Merged
thecodacus
merged 14 commits into
perf
thecodacus/llama.cpp:perf
from
fable5/turboquant
thecodacus/llama.cpp:fable5/turboquant
Copy head branch name to clipboard
Jul 18, 2026
Conversation
Commits
14
(14)
Checks
Files changed
Merged
feat: TurboQuant KV-cache quantization — all backends
#1
thecodacus
merged 14 commits into
perf
thecodacus/llama.cpp:perf
from
fable5/turboquant
thecodacus/llama.cpp:fable5/turboquant
Copy head branch name to clipboard
Commits
Commits on Jul 17, 2026
feat: TurboQuant KV-cache quantization — core types + CPU backend
Show description for a579e99
thecodacus
committed
a579e99
View commit details
Copy full SHA for a579e99
Browse repository at this point
feat: turbo KV cache — graph wiring and CLI types
Show description for 7fcd91d
thecodacus
committed
7fcd91d
View commit details
Copy full SHA for 7fcd91d
Browse repository at this point
feat: turbo KV cache — Metal backend (WHT kernel + FA dispatch)
Show description for f386b46
thecodacus
committed
f386b46
View commit details
Copy full SHA for f386b46
Browse repository at this point
feat: turbo KV cache — CUDA backend (set-rows quantize, fattn vec dispatch, dequant)
Show description for 42b6620
thecodacus
committed
42b6620
View commit details
Copy full SHA for 42b6620
Browse repository at this point
feat: turbo KV cache — Vulkan backend (set-rows quantize, turbo3 FA dispatch, WHT op)
Show description for fb945e8
thecodacus
committed
fb945e8
View commit details
Copy full SHA for fb945e8
Browse repository at this point
fix: extern initialization warning and C linkage for wht group size
Show description for 5d1cb9d
thecodacus
committed
5d1cb9d
View commit details
Copy full SHA for 5d1cb9d
Browse repository at this point
feat: turbo2/turbo4 FA dequant for Vulkan
Show description for 0a28733
thecodacus
committed
0a28733
View commit details
Copy full SHA for 0a28733
Browse repository at this point
Commits on Jul 18, 2026
fix: bump RPC proto patch version for GGML_OP_TURBO_WHT
thecodacus
committed
a60c2c5
View commit details
Copy full SHA for a60c2c5
Browse repository at this point
fix: export WHT group size via GGML_API for Windows shared builds
thecodacus
committed
df835fd
View commit details
Copy full SHA for df835fd
Browse repository at this point
fix: HIP/MUSA portability for turbo kernels
Show description for d13f848
thecodacus
committed
d13f848
View commit details
Copy full SHA for d13f848
Browse repository at this point
fix: drop SYCL turbo fattn-vec instances (no turbo dequant support in SYCL fattn)
thecodacus
committed
6ee6b7b
View commit details
Copy full SHA for 6ee6b7b
Browse repository at this point
fix: size vec_dot buffer by vec_dot_type and handle turbo/tq types in quantize test
Show description for 4a870bb
thecodacus
committed
4a870bb
View commit details
Copy full SHA for 4a870bb
Browse repository at this point
fix: add turbo fattn-vec instances to HIP and MUSA source lists
thecodacus
committed
38f24f2
View commit details
Copy full SHA for 38f24f2
Browse repository at this point
fix: add turbo fattn vec kernels to HIP VGPR ignore list
Show description for 12a0202
thecodacus
committed
12a0202
View commit details
Copy full SHA for 12a0202
Browse repository at this point
You can’t perform that action at this time.