paper-with-me

Papers

Rethinking Key-Value Cache Compression Techniques for Large Language Model Serving

2025-03-31 · Wei Gao, Xinyu Zhou, Peng Sun, Tianwei Zhang, Yonggang Wen

Key-Value cache (\texttt{KV} \texttt{cache}) compression has emerged as a promising technique to optimize Large Language Model (LLM) serving. It primarily decreases the memory consumption of \texttt{KV} \texttt{cache} to reduce the computation cost. Despite the development of many compression algorithms, their applications in production environments are still not prevalent. In this paper, we revisit mainstream \texttt{KV} \texttt{cache} compression solutions from a practical perspective. Our contributions are three-fold. First, we comprehensively review existing algorithmic designs and benchmark studies for \texttt{KV} \texttt{cache} compression and identify missing pieces in their performance measurement, which could hinder their adoption in practice. Second, we empirically evaluate representative \texttt{KV} \texttt{cache} compression methods to uncover two key issues that affect the computational efficiency: (1) while compressing \texttt{KV} \texttt{cache} can reduce memory consumption, current implementations (e.g., FlashAttention, PagedAttention) do not optimize for production-level LLM serving, resulting in suboptimal throughput performance; (2) compressing \texttt{KV} \texttt{cache} may lead to longer outputs, resulting in increased end-to-end latency. We further investigate the accuracy performance of individual samples rather than the overall performance, revealing the intrinsic limitations in \texttt{KV} \texttt{cache} compression when handling specific LLM tasks. Third, we provide tools to shed light on future \texttt{KV} \texttt{cache} compression studies and facilitate their practical deployment in production. They are open-sourced in \href{https://github.com/LLMkvsys/rethink-kv-compression}{https://github.com/LLMkvsys/rethink-kv-compression}.

📄 PDF Abstract BibTeX arXiv:2503.24000

Code (1)

llmkvsys/rethink-kv-compression 공식 구현 pytorch

Tasks

Computational EfficiencyLanguage ModelingLanguage ModellingLarge Language Model

Similar Papers 제목 키워드 기반

Rethinking LoRA Memory Through the Lens of KV Cache Compression

2026-06-04 · Chunsheng Zuo, Liaoyaqi Wang, William Jurayj, William Fleshman 외 arxiv

Parametric retrieval augmentation encodes document information into lightweight, document-specific modules such as LoRA adapters, reducing the need to include all evidence as input context. However, it remains unclear ho…

Question AnsweringAnswer Generation

SVDq: 1.25-bit and 410x Key Cache Compression for LLM Attention

2025-02-21 · Hong Yankun, Li Xing, Zhen Hui-Ling, Yu Xianzhi 외

For the efficient inference of Large Language Models (LLMs), the effective compression of key-value (KV) cache is essential. Three main types of KV cache compression techniques, namely sparsity, channel compression, and …

Quantization

Key, Value, Compress: A Systematic Exploration of KV Cache Compression Techniques

2025-03-14 · Neusha Javidnia, Bita Darvish Rouhani, Farinaz Koushanfar

Large language models (LLMs) have demonstrated exceptional capabilities in generating text, images, and video content. However, as context length grows, the computational cost of attention increases quadratically with th…

Cache Me If You Must: Adaptive Key-Value Quantization for Large Language Models

2025-01-31 · Alina Shutova, Vladimir Malinovskii, Vage Egiazarian, Denis Kuznedelev 외

Efficient real-world deployments of large language models (LLMs) rely on Key-Value (KV) caching for processing and generating long outputs, reducing the need for repetitive computation. For large contexts, Key-Value cach…

GPUQuantization

Unifying KV Cache Compression for Large Language Models with LeanKV

2024-12-04 · Yanqi Zhang, Yuwei Hu, Runyuan Zhao, John C. S. Lui 외

Large language models (LLMs) exhibit exceptional performance but incur significant serving costs due to their substantial memory requirements, with the key-value (KV) cache being a primary bottleneck. Existing KV cache c…

GPUQuantization