arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2501.11286cs.AR
收藏

注意力机制的混合光子-数字加速器

Hybrid Photonic-digital Accelerator for Attention Mechanism

Huize Li, Dan Chen, Tulika Mitra

AI总结:

提出HyAtten,一种基于光子的注意力加速器,通过信号比较器分类信号并集成低分辨率转换器及数字电路,最小化转换开销,相比最先进光子Transformer加速器,实现9.8倍性能/面积和2.2倍能效/面积提升。

摘要:

基于注意力的Transformer的广泛采用和巨大的计算资源需求,激发了对高效硬件加速器的需求。与基于数字的加速器不同,由于光子学具有高能效和超快处理速度,人们对其探索的兴趣日益增长。然而,显著的信号转换开销限制了基于光子的加速器的性能。在这项工作中,我们提出了HyAtten,一种基于光子的注意力加速器,旨在最小化信号转换开销。HyAtten引入了一个信号比较器,根据信号是否可由低分辨率转换器处理,将信号分为两类。HyAtten集成了低分辨率转换器来处理所有低分辨率信号,从而提高了光子计算的并行性。对于需要高分辨率转换的信号,HyAtten使用数字电路代替信号转换器,以减少面积和延迟开销。与最先进的基于光子的Transformer加速器相比,HyAtten实现了9.8倍的性能/面积和2.2倍的能效/面积提升。

英文摘要:

The wide adoption and substantial computational resource requirements of attention-based Transformers have spurred the demand for efficient hardware accelerators. Unlike digital-based accelerators, there is growing interest in exploring photonics due to its high energy efficiency and ultra-fast processing speeds. However, the significant signal conversion overhead limits the performance of photonic-based accelerators. In this work, we propose HyAtten, a photonic-based attention accelerator with minimize signal conversion overhead. HyAtten incorporates a signal comparator to classify signals into two categories based on whether they can be processed by low-resolution converters. HyAtten integrates low-resolution converters to process all low-resolution signals, thereby boosting the parallelism of photonic computing. For signals requiring high-resolution conversion, HyAtten uses digital circuits instead of signal converters to reduce area and latency overhead. Compared to state-of-the-art photonic-based Transformer accelerator, HyAtten achieves 9.8X performance/area and 2.2X energy-efficiency/area improvement.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑