FedChip: Federated LLM for Artificial Intelligence Accelerator Chip Design
This work addresses data privacy and domain-specific training issues in AI hardware design automation, enabling collaborative model enhancement among chip design parties, though it appears incremental as it builds on existing federated learning and LLM methods.
The paper tackles the problem of automating AI hardware chip design using Large Language Models (LLMs) while addressing data privacy concerns, resulting in FedChip, a federated fine-tuning approach that improves design quality by over 77% compared to high-end LLMs.
AI hardware design is advancing rapidly, driven by the promise of design automation to make chip development faster, more efficient, and more accessible to a wide range of users. Amongst automation tools, Large Language Models (LLMs) offer a promising solution by automating and streamlining parts of the design process. However, their potential is hindered by data privacy concerns and the lack of domain-specific training. To address this, we introduce FedChip, a Federated fine-tuning approach that enables multiple Chip design parties to collaboratively enhance a shared LLM dedicated for automated hardware design generation while protecting proprietary data. FedChip enables parties to train the model on proprietary local data and improve the shared LLM's performance. To exemplify FedChip's deployment, we create and release APTPU-Gen, a dataset of 30k design variations spanning various performance metric values such as power, performance, and area (PPA). To encourage the LLM to generate designs that achieve a balance across multiple quality metrics, we propose a new design evaluation metric, Chip@k, which statistically evaluates the quality of generated designs against predefined acceptance criteria. Experimental results show that FedChip improves design quality by more than 77% over high-end LLMs while maintaining data privacy