The increasing complexity and scale of Deep Neural Networks (DNNs) necessitate specialized tensor accelerators, such as Tensor Processing Units (TPUs), to meet various computational and energy efficiency requirements. Nevertheless, designing optimal TPU remains challenging due to the high domain expertise level, considerable manual design time, and lack of high-quality, domain-specific datasets. This paper introduces TPU-Gen, the first Large Language Model (LLM) based framework designed to automate the exact and approximate TPU generation process, focusing on systolic array architectures. TPU-Gen is supported with a meticulously curated, comprehensive, and open-source dataset that covers a wide range of spatial array designs and approximate multiply-and-accumulate units, enabling design reuse, adaptation, and customization for different DNN workloads. The proposed framework leverages Retrieval-Augmented Generation (RAG) as an effective solution for a data-scare hardware domain in building LLMs, addressing the most intriguing issue, hallucinations. TPU-Gen transforms high-level architectural specifications into optimized low-level implementations through an effective hardware generation pipeline. Our extensive experimental evaluations demonstrate superior performance, power, and area efficiency, with an average reduction in area and power of 92\% and 96\% from the manual optimization reference values. These results set new standards for driving advancements in next-generation design automation tools powered by LLMs.

翻译：深度神经网络（DNN）日益增长的复杂性和规模，要求使用专门的张量加速器（如张量处理单元，TPU）来满足不同的计算和能效需求。然而，由于所需领域专业知识水平高、手动设计耗时巨大以及缺乏高质量、领域特定的数据集，设计最优的TPU仍然具有挑战性。本文介绍了TPU-Gen，这是首个基于大语言模型（LLM）的框架，旨在自动化精确及近似TPU的生成过程，重点关注脉动阵列架构。TPU-Gen由一个精心策划、全面且开源的数据集支持，该数据集涵盖了广泛的空间阵列设计和近似乘累加单元，从而能够为不同的DNN工作负载实现设计复用、适配和定制。该框架利用检索增强生成（RAG）作为在数据稀缺的硬件领域构建LLM的有效解决方案，以解决最引人关注的幻觉问题。TPU-Gen通过一个高效的硬件生成流程，将高级架构规范转化为优化的低级实现。我们广泛的实验评估展示了其在性能、功耗和面积效率方面的卓越表现，与手动优化的参考值相比，面积和功耗平均降低了92%和96%。这些结果为推动基于LLM的下一代设计自动化工具的进步树立了新的标准。