English
Related papers

Related papers: AMLA: MUL by ADD in FlashAttention Rescaling

200 papers

Multi-Head Latent Attention (MLA), introduced in DeepSeek-V2, improves the efficiency of large language models by projecting query, key, and value tensors into a compact latent space. This architectural change reduces the KV-cache size and…

Hardware Architecture · Computer Science 2026-04-10 Robin Geens , Marian Verhelst

Attention mechanisms, particularly within Transformer architectures and large language models (LLMs), have revolutionized sequence modeling in machine learning and artificial intelligence applications. To compute attention for increasingly…

Hardware Architecture · Computer Science 2025-06-02 Kosmas Alexandridis , Vasileios Titopoulos , Giorgos Dimitrakopoulos

Multi-Head Attention (MHA) is a critical computational kernel in transformer-based AI models. Emerging scalable tile-based accelerator architectures integrate increasing numbers of tightly-packed processing elements (PEs) with tensor units.…

Hardware Architecture · Computer Science 2025-05-27 Chi Zhang , Luca Colagrande , Renzo Andri , Thomas Benz , Gamze Islamoglu , Alessandro Nadalini , Francesco Conti , Yawei Li , Luca Benini

Multi-Head Latent Attention (MLA) is a recent attention mechanism adopted in state-of-the-art LLMs such as DeepSeek-v3 and Kimi K2. Thanks to its novel formulation, MLA allows two functionally equivalent but computationally distinct kernel…

Machine Learning · Computer Science 2026-02-13 Ahmet Caner Yüzügüler , Ahmet Çelik , Jiawei Zhuang , Lukas Cavigelli

FlashAttention series has been widely applied in the inference of large language models (LLMs). However, FlashAttention series only supports the high-level GPU architectures, e.g., Ampere and Hopper. At present, FlashAttention series is not…

Transformer-based large language models (LLMs) rely heavily on intensive matrix multiplications for attention and feed-forward layers, with the Q, K, and V linear projections in the Multi-Head Self-Attention (MHA) module constituting a…

Hardware Architecture · Computer Science 2025-05-22 Richie Li , Sicheng Chen

Linear RNNs with gating recently demonstrated competitive performance compared to Transformers in language modeling. Although their linear compute scaling in sequence length offers theoretical runtime advantages over Transformers, realizing…

Machine Learning · Computer Science 2025-12-30 Maximilian Beck , Korbinian Pöppel , Phillip Lippe , Sepp Hochreiter

Multi-head latent attention (MLA) is designed to optimize KV cache memory through low-rank key-value joint compression. Rather than caching keys and values separately, MLA stores their compressed latent representations, reducing memory…

Computation and Language · Computer Science 2025-09-09 Guihong Li , Mehdi Rezagholizadeh , Mingyu Yang , Vikram Appia , Emad Barsoum

Key-Value (KV) cache memory and bandwidth increasingly dominate large language model inference cost in long-context and long-generation regimes. Architectures such as multi-head latent attention (MLA) and hybrid sliding-window attention…

Computation and Language · Computer Science 2026-04-08 Zhen Cheng , Hao-Bo Yang , Wan-Yi Huang , Jin-Long Li

Computational workloads composing traditional transformer models are starkly bifurcated. Multi-Head Attention (MHA) and Grouped-Query Attention are memory-bound due to low arithmetic intensity, while FeedForward Networks are compute-bound.…

While FP8 attention has shown substantial promise in innovations like FlashAttention-3, its integration into the decoding phase of the DeepSeek Multi-head Latent Attention (MLA) architecture presents notable challenges. These challenges…

Machine Learning · Computer Science 2026-04-29 Yifan Zhang , Zunhai Su , Shuhao Hu , Rui Yang , Wei Wu , Yulei Qian , Yuchen Xie , Xunliang Cai

Multi-Head Latent Attention (MLA), introduced in DeepSeek-V2, compresses key-value states into a low-rank latent vector, caching only this vector to reduce memory. In tensor parallelism (TP), however, attention heads are computed across…

Machine Learning · Computer Science 2025-08-26 Xiaojuan Tang , Fanxu Meng , Pingzhi Tang , Yuxuan Wang , Di Yin , Xing Sun , Muhan Zhang

While the Transformer architecture dominates many fields, its quadratic self-attention complexity hinders its use in large-scale applications. Linear attention offers an efficient alternative, but its direct application often degrades…

Computer Vision and Pattern Recognition · Computer Science 2026-01-15 Kewei Zhang , Ye Huang , Yufan Deng , Jincheng Yu , Junsong Chen , Huan Ling , Enze Xie , Daquan Zhou

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

The Transformer-based Whisper model has achieved state-of-the-art performance in Automatic Speech Recognition (ASR). However, its Multi-Head Attention (MHA) mechanism results in significant GPU memory consumption due to the linearly growing…

Sound · Computer Science 2026-03-03 Sen Zhang , Jianguo Wei , Wenhuan Lu , Xianghu Yue , Wei Li , Qiang Li , Pengcheng Zhao , Ming Cai , Luo Si

We provide an optimized implementation of the forward pass of FlashAttention-2, a popular memory-aware scaled dot-product attention algorithm, as a custom fused CUDA kernel targeting NVIDIA Hopper architecture and written using the…

Machine Learning · Computer Science 2023-12-20 Ganesh Bikshandi , Jay Shah

As vision-language models (VLMs) tackle increasingly complex and multimodal tasks, the rapid growth of Key-Value (KV) cache imposes significant memory and computational bottlenecks during inference. While Multi-Head Latent Attention (MLA)…

Computer Vision and Pattern Recognition · Computer Science 2026-01-19 Xiaoran Fan , Zhichao Sun , Tao Ji , Lixing Shen , Tao Gui

Transformer-based deep learning models are increasingly deployed on energy, and DRAM bandwidth constrained devices such as laptops and gaming consoles, which presents significant challenges in meeting the latency requirements of the models.…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-08-26 Aadesh Deshmukh , Venkata Yaswanth Raparti , Samuel Hsu

Large Language Model (LLM) deployment is increasingly shifting to cost-efficient accelerators like Google's Tensor Processing Units (TPUs), prioritizing both performance and total cost of ownership (TCO). However, existing LLM inference…

Performance · Computer Science 2026-04-20 Jevin Jiang , Ying Chen , Blake A. Hechtman , Fenghui Zhang , Yarong Mu

Large language models (LLMs) face significant challenges in processing long contexts due to the linear growth of the key-value (KV) cache and quadratic complexity of self-attention. Existing approaches address these bottlenecks separately:…

Computation and Language · Computer Science 2026-04-17 Zeng You , Yaofo Chen , Qiuwu Chen , Ying Sun , Shuhai Zhang , Yingjian Li , Yaowei Wang , Mingkui Tan
‹ Prev 1 2 3 10 Next ›