Summary of A Reinforcement Learning-based Task Mapping Method to Improve the Reliability Of Clustered Manycores, by Fatemeh Hossein-khani and Omid Akbari
A Reinforcement Learning-Based Task Mapping Method to Improve the Reliability of Clustered Manycores
by Fatemeh Hossein-Khani, Omid Akbari
First submitted to arxiv on: 26 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 A novel reinforcement learning-based method for task mapping in manycore systems is proposed, aiming to improve reliability while meeting performance demands. The approach considers aging mechanisms like NBTI, HCI, TC, and EM, which become more significant as system scale increases. The method consists of three steps: bin packing using DBSCAN clustering, followed by Q-learning-based task-to-bin mapping and task-to-core mapping. Unlike state-of-the-art techniques, this proposed method can be performed during runtime without requiring offline parameter calculation. Evaluation on SPLASH2 and PARSEC benchmark suite applications shows a 27% increase in mean time to failure (MTTF) compared to previous task mapping methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Manycore systems are becoming increasingly important, but they also pose challenges in managing reliability while meeting performance demands. This paper proposes a new way to map tasks on these systems using reinforcement learning. The method considers different aging mechanisms that can affect the system’s performance over time. It consists of three steps: first, it groups tasks into clusters based on their temperature; then, it uses an algorithm called Q-learning to assign each task to a core in a way that minimizes thermal variations. Unlike previous methods, this approach can be used during runtime without needing any special setup. The results show that this method can improve the reliability of manycore systems by up to 27%. |
Keywords
» Artificial intelligence » Clustering » Reinforcement learning » Temperature
