Chiplet-Gym: Optimizing Chiplet-Based AI Accelerator Design With Reinforcement Learning
IEEE Transactions on Computers · 2025-01 · 74 (1) · 43-56
DOI: 10.1109/tc.2024.3457740 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Computers, 2025-01. Verfasst von Kaniz Mishty und 1 weiteren Autor:innen. Die Arbeit wurde bisher 20-mal zitiert (Crossref) und nennt 56 Literaturangaben. Gefördert von: National Science Foundation.
Abstract (Originalsprache)
Modern Artificial Intelligence (AI) workloads demand computing systems with large silicon area to sustain throughput and competitive performance. However, prohibitive manufacturing costs and yield limitations at advanced tech nodes and die-size reaching the reticle limit restrain us from achieving this. With the recent innovations in advanced packaging technologies, chiplet-based architectures have gained significant attention in the AI hardware domain. However, the vast design space of chiplet-based AI accelerator design and the absence of system and package-level co-design methodology make it difficult for the designer to find the optimum design point regarding Power, Performance, Area, and manufacturing Cost (PPAC). This paper presents Chiplet-Gym, a Reinforcement Learning (RL)-based optimization framework to explore the vast design space of chiplet-based AI accelerators, encompassing the resource allocation, placement, and packaging architecture. We analytically model the PPAC of the chiplet-based AI accelerator and integrate it into an OpenAI gym environment to evaluate the design points. We also explore non-RL-based optimization approaches and combine these two approaches to ensure the robustness of the optimizer. The optimizer-suggested design point achieves $1.52\boldsymbol{\times}$ 1.52 × throughput, $0.27\boldsymbol{\times}$ 0.27 × energy, and $0.89\boldsymbol{\times}$ 0.89 × cost of its monolithic counterpart at iso-area.
Förderung
- National Science Foundation
Zitieren
APA
Mishty, K., & Sadi, M. (2025). Chiplet-Gym: Optimizing Chiplet-Based AI Accelerator Design With Reinforcement Learning. IEEE Transactions on Computers, 74(1), 43-56. https://doi.org/10.1109/tc.2024.3457740
Vancouver
Mishty K, Sadi M. Chiplet-Gym: Optimizing Chiplet-Based AI Accelerator Design With Reinforcement Learning. IEEE Transactions on Computers. 2025;74(1):43-56. doi:10.1109/tc.2024.3457740
BibTeX
@article{Mishty2025,
author = {Mishty, Kaniz and Sadi, Mehdi},
title = {Chiplet-Gym: Optimizing Chiplet-Based AI Accelerator Design With Reinforcement Learning},
journal = {IEEE Transactions on Computers},
year = {2025},
volume = {74},
number = {1},
pages = {43--56},
doi = {10.1109/tc.2024.3457740}
}
EUVIMED