Gandalf
Named by 2 kept items in this workspace.
The items that name it 2
SSD KV cache offloading system for LLM serving with SSD lifetime awareness. It prioritises admitting full-attention KV cache into SSD over cache from other attention types, which extends SSD lifetime without sacrificing cache hit rate.
Strongest baseline verifier compared against Sauron.
In every workspace you can read 5
AI
All technologies in this workspace.