# R23: MPCache: MPC-Friendly KV Cache Eviction for Efficient Private Large Language Model Inference

MPCache evaluates MPC-friendly static and dynamic KV-cache eviction in a three-party SecretFlow/SPU setting; its artifact focuses state selection and operator costs. PLLM status: PLLM reimplementation.

[View canonical HTML](https://pllm.run/research/papers/r23-mpcache/)

Document ID: `pllm.docs.research.papers.r23-mpcache`  
Release: `0.1.0`

**Status:** `PLLM reimplementation`
**Year:** 2025
**Authors:** Wenxuan Zeng, Ye Dong, Jinjin Zhou, Junming Ma, Jin Tan, Runsheng Wang, Meng Li
**External source:** [MPCache: MPC-Friendly KV Cache Eviction for Efficient Private Large Language Model Inference](https://arxiv.org/abs/2501.06807v2)

## Source contribution

MPCache evaluates MPC-friendly static and dynamic KV-cache eviction in a three-party SecretFlow/SPU setting; its artifact focuses state selection and operator costs.

## PLLM perspective

Reproduce cached generation and use its explicit KV state as a comparison oracle for PLLM decoder state without inferring complete protected boundary coverage.

No MPCache reproduction has been executed. Current PLLM work implements a deterministic structural DecoderPlan adaptation for dense Qwen2/Qwen3 fixed-shape caches; protected execution, fidelity, assurance, and matched benchmarks remain unchecked.

PLLM implements only a structural `DecoderPlan` adaptation. It does not reproduce MPCache's three-party protected runtime. Fidelity, protected execution, assurance, and matched benchmark status remain unchecked.

The source's boundary remains binding: Three online semi-honest parties and incomplete protected-boundary evidence are incompatible with the one-evaluator complete-model target.

## PLLM modules and documentation

Target module identifiers: `protocol.mpcache`, `state.secret_shared_kv`, `gate.mpcache_attention`. These identifiers describe scoped work unless this page explicitly links an implementation source. Dependencies: None.

- [Research backlog](/research/backlog/)
- [Reimplementation backlog](/research/backlog/)
- [Clean-room workflow](/research/clean-room/)
- [PLLM method status](/research/methods/)
- [Decoder plan usage](/sdk/plans/)
- [MPCache structural adaptation source](https://github.com/blairhudson/pllm/blob/main/crates/pllm-models/src/cache.rs)

## Backlog

This source is assigned to reimplementation priority 5, **Protected KV-cache policy**. Its source-registry order is 23. Follow current gates and unfinished work in the [research backlog](/research/backlog/).