# S2-Attention: Hardware-Aware Context Sharding Among Attention Heads

arXiv: https://arxiv.org/abs/2407.17678 · HTML: https://mlsanity.com/paper/2407.17678/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 20 references that are in the map.
- First posted on arXiv 25 Jul 2024.

## Builds on (most cited first, 17 of 17)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [Big Bird: Transformers for Longer Sequences](https://mlsanity.com/paper/2007.14062/) (2020)
- [Reformer: The Efficient Transformer](https://mlsanity.com/paper/2001.04451/) (2020)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://mlsanity.com/paper/2406.17557/) (2024)
- [Efficient Transformers: A Survey](https://mlsanity.com/paper/2009.06732/) (2020)
- [Axial Attention in Multidimensional Transformers](https://mlsanity.com/paper/1912.12180/) (2019)
- [Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer](https://mlsanity.com/paper/2203.03466/) (2022)
- [Retrieval Head Mechanistically Explains Long-Context Factuality](https://mlsanity.com/paper/2404.15574/) (2024)
- [Data Engineering for Scaling Language Models to 128K Context](https://mlsanity.com/paper/2402.10171/) (2024)
- [Blockwise Self-Attention for Long Document Understanding](https://mlsanity.com/paper/1911.02972/) (2019)
- [Pyramid-BERT: Reducing Complexity via Successive Core-set based Token Selection](https://mlsanity.com/paper/2203.14380/) (2022)
- [Efficient Language Adaptive Pre-training: Extending State-of-the-Art Large Language Models for Polish](https://mlsanity.com/paper/2402.09759/) (2024)

## Abstract (arXiv)

Sparse attention, which selectively attends to a subset of tokens in the context was supposed to be efficient. However, its theoretical reduction in FLOPs has rarely translated into wall-clock speed-up over its dense attention counterparts due to the lack of hardware-aware optimizations like FlashAttention. Meanwhile, it remains unclear whether sparse attention can maintain the model's quality at a scale of today's large language models (LLMs) and how. This paper presents Sparsely-Sharded(S2) Attention, a Triton library that provides kernel optimization for sparse attention customizable at both per-head and per-context-range levels. S2-Attention enables the exploration of novel and high-performance sparse attention techniques, which we demonstrate through extensive ablations across a wide range of sparse attention designs at various model scales. From these insights, we present several basic guidelines to design sparse attention that can achieve not only practical efficiency improvements, but also strong downstream performance. To achieve high parallelization and optimized memory IO, sparse attention should shard the context heterogeneously across attention heads, where each head attends to a different subset of tokens while collectively covering the full context. Meanwhile, we find hybrid architectures combining sparse and dense attention particularly beneficial in practice. S2-Attention achieves wall-clock speedup of 8.79X, 15.87X, 25.3X compared to the strong FlashAttention-2 baseline with strong downstream performance on-par with full attention and perfect retrieval performance at a 128k context length. At inference, for 7B models, our model, with the help of our S2-Attention kernel, achieves 4.5x speed-up compared to dense counterparts. S2-Attention is released with easy-to-customize APIs for direct usage in Megatron and vLLM.