Skip to content

Navigation Menu

Appearance settings

View all features
- 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
- 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
- COMMUNITY
  GitHub SponsorsFund open source developers
- PROGRAMS
  Security Lab
  Maintainer Community
  Accelerator
  GitHub Stars
  Archive Program
- REPOSITORIES
  Topics
  Trending
  Collections
- ENTERPRISE SOLUTIONS
  Enterprise platformAI-powered developer platform
- AVAILABLE ADD-ONS
  GitHub Advanced SecurityEnterprise-grade security features
  Copilot for BusinessEnterprise-grade AI features
  Premium SupportEnterprise-grade 24/7 support
Pricing

Search code, repositories, users, issues, pull requests...

Search

Clear

Search syntax tips

Provide feedback

We read every piece of feedback, and take your input very seriously.

Include my email address so I can be contacted

Saved searches

Use saved searches to filter your results more quickly

Name

Query

To see all available qualifiers, see our documentation.

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

ggml-org / llama.cpp Public

Notifications You must be signed in to change notification settings
Fork 15.6k
Star 98.6k

Code
Issues 486
Pull requests 793
Discussions
Actions
Projects
Wiki
Security 12
Insights

Additional navigation options

Code
Issues
Pull requests
Discussions
Actions
Projects
Wiki
Security
Insights

Vulkan Scalar Flash Attention Refactor#19625

Merged

0cc4m merged 46 commits intomasterggml-org/llama.cpp:masterfrom

0cc4m/vulkan-fa-scalar-optggml-org/llama.cpp:0cc4m/vulkan-fa-scalar-optCopy head branch name to clipboard

Feb 24, 2026

Conversation Commits46 (46)Checks Files changed

Merged

Vulkan Scalar Flash Attention Refactor#19625
0cc4m merged 46 commits intomasterfrom
0cc4m/vulkan-fa-scalar-opt

Commits

Commits on Feb 23, 2026

vulkan: allow using fp16 in coopmat1 flash attention shader
0cc4m
committed
split rows inside of subgroups for faster synchronization
0cc4m
committed
use row_split when Br >= 4, change reductions to use shared memory if row_split == 1
0cc4m
committed
use f32 scalar FA if f16 is not supported by device
0cc4m
committed
fix amd workgroup size issue
0cc4m
committed
optimize masksh use
0cc4m
committed
add medium rows FA shader Br size
0cc4m
committed
fixes
0cc4m
committed
add padding to mask shmem buffer
0cc4m
committed
cache q values into registers for KQ
0cc4m
committed
fuse lf accumulation, pf and v accumulation into a loop
0cc4m
committed
stage K loads through shmem
0cc4m
committed
stage V loads through shmem
0cc4m
committed
only stage through shmem on Nvidia
0cc4m
committed
default to Bc 32
0cc4m
committed
also stage V through shmem when this is done for K
0cc4m
committed
dynamic subgroups for intel
0cc4m
committed
use vectorized stores
0cc4m
committed
use float_type for dequantize4 functions
0cc4m
committed
use smaller scalar rows size for smaller rows count
0cc4m
committed
relax flash attention split_k condition to allow non-gqa use
0cc4m
committed
use minimal subgroup size on Intel
0cc4m
committed
fix shmem support function
0cc4m
committed
fix rebase issues
0cc4m
committed
fixes
0cc4m
committed
Bc 4 for scalar FA is not a valid configuration
0cc4m
committed
Use wave32 on AMD RDNA for scalar FA
0cc4m
committed
add Intel shader core count lookup-table
0cc4m
committed
fix regressions
0cc4m
committed
device tuning
0cc4m
committed
tmpsh size fix
0cc4m
committed
fix editorconfig
0cc4m
committed
refactor fa tuning logic into a single place
0cc4m
committed
fix gqa opt logic
0cc4m
committed
fix block_rows with small n_rows
0cc4m
committed
amd tuning
0cc4m
committed
fix hsk=72/80 issue
0cc4m
committed
tuning
0cc4m
committed
allow condition skipping for column check
0cc4m
committed
use float16 for Of if available
0cc4m
committed
address feedback
0cc4m
committed
fix bad RDNA performance on head size <= 128 by limiting occupancy
0cc4m
committed
allow printing pipeline stats
0cc4m
committed
cleanup and fixes
0cc4m
committed
limit occupancy for GCN for small batch FA with large HSK
0cc4m
committed
disable f16 FA for GCN AMD GPUs on the proprietary driver
0cc4m
committed

Footer

© 2026 GitHub, Inc.

Footer navigation

Terms
Privacy
Security
Status
Community
Docs
Contact

You can’t perform that action at this time.