Inference | Hybrid prefix caching.#3225

Open

lmcafee-nvidia wants to merge 14 commits intoNVIDIA:mainfrom

lmcafee-nvidia:prefix-caching-mamba

Commits on Feb 9, 2026

Commits on Feb 16, 2026

Adding unit tests
oyilmaz-nvidia
committed

Commits on Mar 6, 2026

Extend test_full_match_adds_single_token for multi-request concurrent writes

lmcafee-nvidia
and
claude
committed