相关论文: TASP: Topology-aware Sequence Parallelism

Linear sequence modeling approaches, such as linear attention, provide advantages like linear-time training and constant-memory inference over sequence lengths. However, existing sequence parallelism (SP) methods are either not optimized…

Efficient parallelization of Large Language Models (LLMs) with long sequences is essential but challenging due to their significant computational and memory demands, particularly stemming from communication bottlenecks in attention…

Sequence parallelism (SP) serves as a prevalent strategy to handle long sequences that exceed the memory limit of a single device. However, for linear sequence modeling methods like linear attention, existing SP approaches do not take…

Sequence parallelism (SP), which divides the sequence dimension of input tensors across multiple computational devices, is becoming key to unlocking the long-context capabilities of generative AI models. This paper investigates the…

We present tensor and sequence parallelism (TSP), a parallel execution strategy that folds tensor parallelism and sequence parallelism onto a single device axis. In conventional multi-dimensional parallelism layouts, tensor parallelism (TP)…

The quadratic computational complexity of standard attention mechanisms presents a severe scalability bottleneck for LLMs in long-context scenarios. While hybrid attention mechanisms combining Full Attention (FA) and Sparse Attention (SA)…

We propose a new attention mechanism with linear complexity, ATP, that fixates \textbf{A}ttention on \textbf{T}op \textbf{P}rincipal keys, rather than on each individual token. Particularly, ATP is driven by an important observation that…

Efficient parallelism is necessary for achieving low-latency, high-throughput inference with large language models (LLMs). Tensor parallelism (TP) is the state-of-the-art method for reducing LLM response latency, however GPU communications…

The Segment Anything Model (SAM) achieves strong open-vocabulary segmentation, but its ViT-based image encoders dominate inference latency and memory. Existing activation compression methods, such as token merging, reduce the token length…

Self-attention in transformer models is an incremental associative memory that maps key vectors to value vectors. One way to speed up self-attention is to employ GPU-compatible vector search algorithms based on standard partitioning methods…

Video Large Language Models (VLLMs) incur substantial prefilling cost due to the large number of visual tokens. While attention-based token pruning offers a promising acceleration strategy, applying it at shallow decoder layers often causes…

Long-context understanding is crucial for many NLP applications, yet transformers struggle with efficiency due to the quadratic complexity of self-attention. Sparse attention methods alleviate this cost but often impose static, predefined…

Our formulation reveals that the reduction across the sequence axis can be efficiently computed in parallel through a tree reduction. Our algorithm, called Tree Attention, for parallelizing exact attention computation across multiple GPUs…

With the rapid expansion in the scale of large language models (LLMs), enabling efficient distributed inference across multiple computing units has become increasingly critical. However, communication overheads from popular distributed…

Attention-based large language models (LLMs) have transformed modern AI applications, but the quadratic cost of self-attention imposes significant compute and memory overhead. Dynamic sparsity (DS) attention mitigates this, yet its hardware…

While Large Vision-Language Models (LVLMs) demonstrate exceptional multi-modal capabilities, the quadratic computational cost of processing high-resolution visual tokens remains a critical bottleneck. Though recent token reduction…

Training large language models (LLMs) is fundamentally constrained by limited device memory and costly inter-device communication. Although pipeline parallelism alleviates memory pressure by partitioning models across devices, it incurs…

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…

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…

As long-context inference becomes central to large language models (LLMs), attention over growing key-value caches emerges as a dominant decoding bottleneck, motivating sparse attention for scalable inference. Fixed-budget top-k sparse…