# IceCache: Memory-efficient KV-cache Management for Long-Sequence LLMs

arXiv: https://arxiv.org/abs/2604.10539 · HTML: https://mlsanity.com/paper/2604.10539/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in KV cache optimization, one of 12 directions in Efficient LLM inference and serving.
- Builds on 25 of its 27 references that are in the map.
- First posted on arXiv 12 Apr 2026.

## Placed in

- Efficient LLM inference and serving / [KV cache optimization](https://mlsanity.com/inference/kv-cache-optimization/)

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Reformer: The Efficient Transformer](https://mlsanity.com/paper/2001.04451/) (2020)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models](https://mlsanity.com/paper/2306.14048/) (2023)
- [Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2406.10774/) (2024)
- [Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention](https://mlsanity.com/paper/2502.11089/) (2025)
- [FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving](https://mlsanity.com/paper/2501.01005/) (2025)
- [Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Time](https://mlsanity.com/paper/2310.17157/) (2023)
- [InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management](https://mlsanity.com/paper/2406.19707/) (2024)
- [Landmark Attention: Random-Access Infinite Context Length for Transformers](https://mlsanity.com/paper/2305.16300/) (2023)
- [MagicPIG: LSH Sampling for Efficient LLM Generation](https://mlsanity.com/paper/2410.16179/) (2024)
- [PQCache: Product Quantization-based KVCache for Long Context LLM Inference](https://mlsanity.com/paper/2407.12820/) (2024)
- [Memory-efficient Transformers via Top-k Attention](https://mlsanity.com/paper/2106.06899/) (2021)
- [SCBench: A KV Cache-Centric Analysis of Long-Context Methods](https://mlsanity.com/paper/2412.10319/) (2024)
- [LongGenBench: Benchmarking Long-Form Generation in Long Context LLMs](https://mlsanity.com/paper/2409.02076/) (2024)
- [IceFormer: Accelerated Inference with Long-Sequence Transformers on CPUs](https://mlsanity.com/paper/2405.02842/) (2024)
- [Fast k-Nearest Neighbour Search via Dynamic Continuous Indexing](https://mlsanity.com/paper/1512.00442/) (2015)
- [Fast k-Nearest Neighbour Search via Prioritized DCI](https://mlsanity.com/paper/1703.00440/) (2017)

## Built on by (most cited first, 5 of 5)

- [KARA: Efficient Reasoning LLM Serving via Sliding-Window KV Cache Compression](https://mlsanity.com/paper/2607.01237/) (2026)
- [H2MT: Semantic Hierarchy-Aware Hierarchical Memory Transformer](https://mlsanity.com/paper/2605.24930/) (2026)
- [AVQ-Attention: Adaptive Vector-Quantized Attention](https://mlsanity.com/paper/2607.12789/) (2026)
- [SPLASH: Co-Designing Sparse Attention with High-Bandwidth Flash for Efficient Long-Context Inference](https://mlsanity.com/paper/2609.23816/) (2026)
- [Accelerating Cross-Model LLM Inference via Attention Map Reuse](https://mlsanity.com/paper/2610.14688/) (2026)

## Abstract (arXiv)

Key-Value (KV) cache plays a crucial role in accelerating inference in large language models (LLMs) by storing intermediate attention states and avoiding redundant computation during autoregressive generation. However, its memory footprint scales linearly with sequence length, often leading to severe memory bottlenecks on resource-constrained hardware. Prior work has explored offloading KV cache to the CPU while retaining only a subset on the GPU, but these approaches often rely on imprecise token selection and suffer performance degradation in long-generation tasks such as chain-of-thought reasoning. In this paper, we propose a novel KV cache management strategy, IceCache, which integrates semantic token clustering with PagedAttention. By organizing semantically related tokens into contiguous memory regions managed by a hierarchical, dynamically updatable data structure, our method enables more efficient token selection and better utilization of memory bandwidth during CPU-GPU transfers. Experimental results on LongBench show that, with a 256-token budget, IceCache maintains 99% of the original accuracy achieved by the full KV cache model. Moreover, compared to other offloading-based methods, IceCache attains competitive or even superior latency and accuracy while using only 25% of the KV cache token budget, demonstrating its effectiveness in long-sequence scenarios. The code is available on our project website at https://yuzhenmao.github.io/IceCache/.