Jointly Optimizing Wirelength and Thermal Fields for Chip Placement

Boning Wang, Daoyang Hua, Xianqi Chen, Qian Zheng, Gang Pan, Wen Yao, De Ma

Keywords: Reinforcement Learning, Chip Design, Thermal Placement

Abstract: Macro placement is a crucial and complex issue in chip design. In recent studies, reinforcement learning (RL) has demonstrated outstanding performance in optimizing chip wirelength, but this leads to thermally inefficient design. Additionally, due to the specialized expertise necessary for creating chip benchmarks and the constraints imposed by confidentiality agreements, there exists a scarcity of publicly available chip thermal placement benchmarks. This work introduces a reinforcement learning-based thermal placement model that can optimize both wirelength and peak temperatures. We also strictly followed the chip design process and established a macro thermal placement benchmark. This significantly reduces the entry barriers for researchers, facilitating benchmarking and result replication. Compared to other models, our model notably diminishes the chip's peak temperature of the chip while slightly extending wirelength, thereby improving the chip's heat dissipation efficiency.

Primary Area: applications to robotics, autonomy, planning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 9663