llama.cpp
26edf081
- test-backend-ops : replace the FA vec tune mode with a bounded (Q,NE) slice
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
5 days ago
test-backend-ops : replace the FA vec tune mode with a bounded (Q,NE) slice
References
#26570 - metal : per-device tuned (Q, NE) for flash-attn vec (replay)
#26620 - tools : split Metal FA-vec tuning into a standalone tuner
Author
forforever73
Committer
forforever73
Parents
35c08cd0
Loading