Summary of Chiplet Placement Order Exploration Based on Learning to Rank with Graph Representation, by Zhihui Deng et al.
Chiplet Placement Order Exploration Based on Learning to Rank with Graph Representation
by Zhihui Deng, Yuanyuan Duan, Leilai Shao, Xiaolei Zhu
First submitted to arxiv on: 7 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed paper tackles the challenge of determining the optimal placement order for chiplets in integrated circuits (ICs) using reinforcement learning (RL). The RL framework, named RLPlanner, utilizes graph representation to select the most suitable placement order for each chiplet-based system. This approach aims to optimize system performance by considering factors like temperature and inter-chiplet wirelength. Compared to traditional methods based on descending chiplet area or interconnect wires, the proposed learning-to-rank (LTR) network demonstrates improved results, specifically reducing total inter-chiplet wirelength by 10.05% and peak system temperature by 1.01%. The paper’s contributions include developing a novel LTR approach for IC placement optimization, leveraging graph representation to capture spatial relationships between chiplets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper talks about how to make tiny electronic chips work together really well. Right now, people are using different types of chips and putting them on special boards called interposers. The order in which these chips are placed matters a lot because it affects how hot they get and how much they use special wires to talk to each other. The researchers came up with a new way to decide the best order for placing the chips, using something called “reinforcement learning.” This method helps find the best order by looking at different options and seeing which one works the best. In this case, the new method resulted in a 10% reduction in the use of special wires and a 1% decrease in how hot the chips got. |
Keywords
* Artificial intelligence * Optimization * Reinforcement learning * Temperature
