KV-cache compression
Mooncake
Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving
Superseded baseline#58 of 234 most-superseded · first seen Jun 24, 2024
Superseded — cited as a baseline and beaten by newer methods
2 papers critique it · 1 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites Mooncake as a baseline.
Mooncake uses reactive placement---blocks are placed based on current utilization---while our Bayesian predictor proactively positions blocks based on predicted reuse probability, achieving 70--84% hit rates vs. Mooncake's reported 65--80%
“their approaches are restricted to sharing in the layer or text segment within adjacent layers or the same LLM, limiting the broader applicability”
Beaten on benchmarks
Head-to-head results where a newer method reports beating Mooncake. Values are copied from the source paper's tables — verify against the cited paper.
FlowKV beats Mooncake
0.0044 vs 0.3010
KV Cache transfer latency (seconds) · [Single Machine]
FlowKV: A Disaggregated Inference Framework with Low-Latency KV Cache Transfer and Load-Aware Scheduling
What to use instead
Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.
- May 22, 2026
- Apr 28, 2026
- Predictive Multi-Tier Memory ManagementPredictive Multi-Tier Memory Management for KV Cache in Large-Scale GPU InferenceApr 19, 2026
- TableCacheTableCache: Primary Foreign Key Guided KV Cache Precomputation for Low Latency Text-to-SQLJan 13, 2026
- Jan 5, 2026
- SemShareKVSemShareKV: Efficient KVCache Sharing for Semantically Similar Prompts via Token-Level LSH MatchingSep 29, 2025