Model techniques map
Techniquesinference & servingKV cache management

specific method · filed under inference & serving

KDA-aware prefix-cache management

Manages KDA recurrent state and MLA KV together in a paged cache to support prefix reuse.

Also called joint KDA–MLA prefix cache management, KDA-aware prefix cache, cache-aware prefix cache.

source
1
model
1
lab adopt it
1
strongest
core

How sources treat it

One count per evidence span, weakest treatment to strongest.

core 2

Documented in

Evidence

2 spans quoted from the sources, strongest treatment first.

We therefore design a KDA-aware prefix cache that manages the two cache types jointly—from a unified paged layout to fine-grained prefix reuse and consistency under concurrent scheduling

coreinference servingin Kimi K3Moonshot AI

a KDA-aware prefix cache packs the fixed-size recurrent state into the same paged pool as the MLA KV cache and keeps long prefixes reusable across requests.

coreinference servingin Kimi K3Moonshot AI

Filed alongside

Other methods under inference & serving :: KV cache management.