English
Related papers

Related papers: PAT: Accelerating LLM Decoding via Prefix-Aware At…

200 papers

Self-attention is an essential component of large language models (LLM) but a significant source of inference latency for long sequences. In multi-tenant LLM serving scenarios, the compute and memory operation cost of self-attention can be…

Machine Learning · Computer Science 2024-08-02 Lu Ye , Ze Tao , Yong Huang , Yang Li

Prefix-sharing among multiple prompts presents opportunities to combine the operations of the shared prefix, while attention computation in the decode stage, which becomes a critical bottleneck with increasing context lengths, is a…

Large language models (LLMs) are increasingly employed for complex tasks that process multiple generation calls in a tree structure with shared prefixes of tokens, including few-shot prompting, multi-step reasoning, speculative decoding,…

Computation and Language · Computer Science 2025-03-10 Jinwei Yao , Kaiqi Chen , Kexun Zhang , Jiaxuan You , Binhang Yuan , Zeke Wang , Tao Lin

Transformer-based large language models (LLMs) are now deployed to hundreds of millions of users. LLM inference is commonly performed on batches of sequences that share a prefix, such as few-shot examples or a chatbot system prompt.…

Machine Learning · Computer Science 2024-05-14 Jordan Juravsky , Bradley Brown , Ryan Ehrlich , Daniel Y. Fu , Christopher Ré , Azalia Mirhoseini

LLM decoding is bottlenecked for large batches and long contexts by loading the key-value (KV) cache from high-bandwidth memory, which inflates per-token latency, while the sequential nature of decoding limits parallelism. We analyze the…

Machine Learning · Computer Science 2025-05-28 Ted Zadouri , Hubert Strauss , Tri Dao

Large language models (LLMs) increasingly play an important role in a wide range of information processing and management tasks in industry. Many of these tasks are performed in large batches or even offline, and the performance indicator…

Computation and Language · Computer Science 2026-04-23 Zhen Zheng , Xin Ji , Taosong Fang , Fanghao Zhou , Chuanjie Liu , Gang Peng

Attention efficiency is critical to large language model (LLM) inference. While prior advances optimize attention execution for individual requests (e.g., FlashAttention), production LLM serving relies on batching requests with highly…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-02-09 Rui Ning , Wei Zhang , Fan Lai

High-throughput inference serving is essential for applications built on large language models (LLMs). Existing serving frameworks reduce request-level and batch-level bubbles through batching and scheduling, but often overlook bubbles…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-05-25 Fengyao Bai , Hongbin Zhang , Zhitao Chen , Jiangsu Du , Zhiguang Chen , Yutong Lu

High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV cache) memory for each request is huge and grows and shrinks…

Machine Learning · Computer Science 2023-09-13 Woosuk Kwon , Zhuohan Li , Siyuan Zhuang , Ying Sheng , Lianmin Zheng , Cody Hao Yu , Joseph E. Gonzalez , Hao Zhang , Ion Stoica

The efficient deployment of large language models (LLMs) in online settings requires optimizing inference performance under stringent latency constraints, particularly the time-to-first-token (TTFT) and time-per-output-token (TPOT). This…

Data Structures and Algorithms · Computer Science 2026-01-05 Gregory Dexter , Shao Tang , Ata Fatahi Baarzi , Qingquan Song , Tejas Dharamsi , Aman Gupta

This study introduces bifurcated attention, a method designed to enhance language model inference in shared-context batch decoding scenarios. Our approach addresses the challenge of redundant memory IO costs, a critical factor contributing…

PagedAttention is a popular approach for dynamic memory allocation in LLM serving systems. It enables on-demand allocation of GPU memory to mitigate KV cache fragmentation -- a phenomenon that crippled the batch size (and consequently…

Machine Learning · Computer Science 2025-01-30 Ramya Prabhu , Ajay Nayak , Jayashree Mohan , Ramachandran Ramjee , Ashish Panwar

Attention mechanisms underpin the success of large language models (LLMs), yet their substantial computational and memory overhead poses challenges for optimizing efficiency and performance. A critical bottleneck arises as KV cache and…

Computation and Language · Computer Science 2025-07-24 Luoyang Sun , Cheng Deng , Jiwen Jiang , Xinjian Wu , Haifeng Zhang , Lei Chen , Lionel Ni , Jun Wang

Each request in LLM inference goes through two phases: compute-bound prefill and memory-bandwidth-bound decode. To improve GPU utilization, recent systems use hybrid batching that combines the prefill and decode phases of different requests…

Machine Learning · Computer Science 2025-02-18 Aditya K Kamath , Ramya Prabhu , Jayashree Mohan , Simon Peter , Ramachandran Ramjee , Ashish Panwar

Large Language Models (LLMs) encounter severe memory inefficiencies during long-context inference due to conventional handling of key-value (KV) caches. In this work, we introduce a novel integration of PagedAttention with PyTorch's…

Machine Learning · Computer Science 2025-06-10 Thomas Joshi , Herman Saini , Neil Dhillon , Antoni Viros i Martin , Kaoutar El Maghraoui

Multimodal large language models (MLLMs) are plagued by exorbitant inference costs attributable to the profusion of visual tokens within the vision encoder. The redundant visual tokens engenders a substantial computational load and…

Computer Vision and Pattern Recognition · Computer Science 2026-02-03 Jiedong Zhuang , Lu Lu , Ming Dai , Rui Hu , Jian Chen , Qiang Liu , Haoji Hu

Large language models encounter critical GPU memory capacity constraints during long-context inference, where KV cache memory consumption severely limits decode batch sizes. While existing research has explored offloading KV cache to DRAM,…

Machine Learning · Computer Science 2026-03-31 Qiuyang Zhang , Kai Zhou , Ding Tang , Kai Lu , Cheng Li , Zhenyu Yang , Peng Xu , Jiguang Wan

LAPS identifies and disaggregates requests with different prompt lengths in LLM serving to reduce TTFT latency. While recent systems have decoupled the prefill and decode stages to improve throughput, they still rely on unified scheduling…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-01-29 Jianshu She , Zonghang Li , Hongchao Du , Shangyu Wu , Wenhao Zheng , Eric Xing , Zhengzhong Liu , Huaxiu Yao , Jason Xue , Qirong Ho

Large Vision-Language Models (VLMs) have achieved remarkable success in multi-modal reasoning, but their inference time efficiency remains a significant challenge due to the memory overhead during decoding, especially when the query and…

Computer Vision and Pattern Recognition · Computer Science 2026-03-26 Fatih Ilhan , Gaowen Liu , Ramana Rao Kompella , Selim Furkan Tekin , Tiansheng Huang , Zachary Yahn , Yichang Xu , Ling Liu

Serving large language models (LLMs) efficiently remains challenging due to the high memory and latency overhead of key-value (KV) cache access during autoregressive decoding. We present \textbf{TinyServe}, a lightweight and extensible…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-09-17 Dong Liu , Yanxuan Yu
‹ Prev 1 2 3 10 Next ›