中文
相关论文

相关论文: Adaptive KV Cache Reuse for Fast Long-Context LLM …

200 篇论文

The growing complexity of LLM usage today, e.g., multi-round conversation and retrieval-augmented generation (RAG), makes contextual states (i.e., KV cache) reusable across user requests. Given the capacity constraints of GPU memory, only a…

分布式、并行与集群计算 · 计算机科学 2024-10-08 Shiwei Gao , Youmin Chen , Jiwu Shu

Across large language model (LLM) applications, we observe an emerging trend for reusing KV caches to save the prefill delays of processing repeated input texts in different LLM inputs. This has led to a broad design space, including…

网络与互联网体系结构 · 计算机科学 2025-03-20 Hanchen Li , Yuhan Liu , Yihua Cheng , Kuntai Du , Junchen Jiang

Interacting with humans through multi-turn conversations is a fundamental feature of large language models (LLMs). However, existing LLM serving engines executing multi-turn conversations are inefficient due to the need to repeatedly…

计算与语言 · 计算机科学 2024-07-02 Bin Gao , Zhuomin He , Puru Sharma , Qingxuan Kang , Djordje Jevdjic , Junbo Deng , Xingkun Yang , Zhou Yu , Pengfei Zuo

Multi-modal Large Language Models (MLLMs) serving systems commonly employ KV-cache compression to reduce memory footprint. However, existing compression methods introduce significant processing overhead and queuing delays, particularly in…

多媒体 · 计算机科学 2025-03-12 Jianian Zhu , Hang Wu , Haojie Wang , Yinghui Li , Biao Hou , Ruixuan Li , Jidong Zhai

Large language model (LLM) applications often reuse previously processed context, such as chat history and documents, which introduces significant redundant computation. Existing LLM serving systems address such redundant computation by…

Prefix KV caching has become a key mechanism in LLM serving: it reduces time to first token (TTFT) by avoiding redundant computation across requests that share a prefix (i.e., the system prompt). However, the accumulated KV cache is often…

分布式、并行与集群计算 · 计算机科学 2026-05-25 Yu Zhu , Aditya Dhakal , Yunming Xiao , Dejan Milojicic , Gustavo Alonso

Large Language Models (LLMs) rely on Key-Value (KV) caching to accelerate inference, and many serving systems further share the KV cache across users' requests to reduce redundant computation. While widely adopted, unrestricted cross-user…

密码学与安全 · 计算机科学 2026-05-25 Guanlong Wu , Zhaohan li , Yao Zhang , Zheng Zhang , Jianyu Niu , Ye Wu , Yinqian Zhang

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…

机器学习 · 计算机科学 2026-04-14 Yuzhen Mao , Qitong Wang , Martin Ester , Ke Li

Cache fusion accelerates generation process of LLMs equipped with RAG through KV caching and selective token recomputation, thereby reducing computational costs and improving efficiency. However, existing methods primarily rely on local…

数据库 · 计算机科学 2026-04-13 Jianxin Yan , Zeheng Qian , Wangze Ni , Zhitao Shen , Zhiping Wang , Haoyang Li , Jia Zhu , Lei Chen , Kui Ren

KV cache has traditionally been stored in GPU memory to accelerate the decoding phase of large language model (LLM) inference. However, it is increasingly necessary to move KV caches outside GPU devices, to enable cache reuse across…

How to efficiently serve Large Language Models (LLMs) has become a pressing issue because of their huge computational cost in their autoregressive generation process. To mitigate computational costs, LLMs often employ the KV Cache technique…

计算与语言 · 计算机科学 2024-07-23 Zheng Wang , Boxiao Jin , Zhongzhi Yu , Minjia Zhang

Recently, large vision-language models (LVLMs) have rapidly gained popularity for their strong generation and reasoning capabilities given diverse multimodal inputs. However, these models incur significant computational and memory overhead…

计算机视觉与模式识别 · 计算机科学 2025-10-20 Ao Wang , Hui Chen , Jiaxin Li , Jianchao Tan , Kefeng Zhang , Xunliang Cai , Zijia Lin , Jungong Han , Guiguang Ding

In Large Language Model (LLM) serving, the KV-cache (KVC) bottleneck causes high tail Time-to-First-Token (TTFT) and Time-Between-Tokens (TBT), impairing user experience, particularly in time-sensitive applications. However, satisfying both…

计算与语言 · 计算机科学 2025-03-26 Haiying Shen , Tanmoy Sen , Masahiro Tanaka

Whether attention key value (KV) states computed for one prompt for a small LLM can be reused to accelerate inference on a new similar prompt, giving an increase to the space to its context memory using an approach called token recycling.…

机器学习 · 计算机科学 2025-12-16 Prashant Pandey

The expanding context windows in large language models (LLMs) have greatly enhanced their capabilities in various applications, but they also introduce significant challenges in maintaining low latency, particularly in Time to First Token…

分布式、并行与集群计算 · 计算机科学 2024-10-10 Yi Xiong , Hao Wu , Changxu Shao , Ziqing Wang , Rui Zhang , Yuhong Guo , Junping Zhao , Ke Zhang , Zhenxuan Pan

Large Language Models (LLMs) are increasingly deployed in scenarios demanding ultra-long context reasoning, such as agentic workflows and deep research understanding. However, long-context inference is constrained by the KV cache, a…

硬件体系结构 · 计算机科学 2026-03-11 Jianlong Lei , Shashikant Ilager

Large Language Models (LLMs) are increasingly deployed in long-context tasks such as reasoning, code generation, and multi-turn dialogue. However, inference over extended contexts is bottlenecked by the Key-Value (KV) cache, whose memory…

计算与语言 · 计算机科学 2026-05-21 Seonghwan Choi , Beomseok Kang , Dongwon Jo , Jae-Joon Kim

Key-value (KV) caching is critical for efficient inference in large language models (LLMs), yet its memory footprint scales linearly with context length, resulting in a severe scalability bottleneck. Existing approaches largely treat KV…

计算与语言 · 计算机科学 2026-04-23 Gradwell Dzikanyanga , Weihao Yang , Hao Huang , Donglei Wu , Shihao Wang , Wen Xia , Sanjeeb K C

Large Language Models (LLMs) excel across a variety of language tasks yet are constrained by limited input lengths and high computational costs. Existing approaches\textemdash such as relative positional encodings (e.g., RoPE, ALiBi) and…

计算与语言 · 计算机科学 2025-02-18 Kun-Hui Lee , Eunhwan Park , Donghoon Han , Seung-Hoon Na

Large Language Models (LLMs), epitomized by ChatGPT's release in late 2022, have revolutionized various industries with their advanced language comprehension. However, their efficiency is challenged by the Transformer architecture's…

计算与语言 · 计算机科学 2024-11-21 Luohe Shi , Hongyi Zhang , Yao Yao , Zuchao Li , Hai Zhao