AttentionLib: A Scalable Optimization Framework for Automated Attention Acceleration on FPGA

The self-attention mechanism is a fundamental component within transformer-based models. Nowadays, as the length of sequences processed by large language models (LLMs) continues to increase, the attention mechanism has gradually become a bottleneck in model inference. The LLM inference process can b...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Proceedings - Design, Automation, and Test in Europe Conference and Exhibition s. 1 - 7
Hlavní autoři: Liu, Zhenyu, Zhou, Xilang, Sun, Faxian, Chen, Jianli, Yu, Jun, Wang, Kun
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: EDAA 31.03.2025
Témata:
Attention mechanisms
Computational modeling
Engines
Field programmable gate arrays
FPGA
Fusion
Hardware
HLS
Manuals
MLIR
Optimization
Scalability
Space exploration
Transformer
Transformers
ISSN:1558-1101
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The self-attention mechanism is a fundamental component within transformer-based models. Nowadays, as the length of sequences processed by large language models (LLMs) continues to increase, the attention mechanism has gradually become a bottleneck in model inference. The LLM inference process can be separated into two phases: prefill and decode. The latter contains memory-intensive attention computation, making FPGA-based accelerators an attractive solution for acceleration. However, designing accelerators tailored for the attention module poses a challenge, requiring substantial manual work. To automate this process and achieve superior acceleration performance, we propose AttentionLib, an MLIR-based framework. AttentionLib automatically performs fusion dataflow optimization for attention computations and generates high-level synthesis code in compliance with hardware constraints. Given the large design space, we provide a design space exploration (DSE) engine to automatically identify optimal fusion dataflows within the specified constraints. Experimental results show that AttentionLib is effective in generating well-suited accelerators for diverse attention computations and achieving superior performance under hardware constraints. Notably, the accelerators generated by AttentionLib exhibit at least a 25.1 × improvement compared to the baselines solely automatically optimized by Vitis HLS. Furthermore, these designs outperform GPUs in decode workloads, showcasing over a 2× speedup for short sequences.
ISSN:1558-1101
DOI:10.23919/DATE64628.2025.10992963