Next-generation artificial intelligence (AI) workloads are posing challenges of scalability and robustness in terms of execution time due to their intrinsic evolving data-intensive characteristics. In this paper, we aim to analyse the potential bottlenecks caused due to data movement characteristics of AI workloads on scale-out accelerator architectures composed of multiple chiplets. Our methodology captures the unicast and multicast communication traffic of a set of AI workloads and assesses aspects such as the time spent in such communications and the amount of multicast messages as a function of the number of employed chiplets. Our studies reveal that some AI workloads are potentially vulnerable to the dominant effects of communication, especially multicast traffic, which can become a performance bottleneck and limit their scalability. Workload profiling insights suggest to architect a flexible interconnect solution at chiplet level in order to improve the performance, efficiency and scalability of next-generation AI accelerators.
@article{arxiv.2410.22262,
title = {Communication Characterization of AI Workloads for Large-scale Multi-chiplet Accelerators},
author = {Mariam Musavi and Emmanuel Irabor and Abhijit Das and Eduard Alarcon and Sergi Abadal},
journal= {arXiv preprint arXiv:2410.22262},
year = {2025}
}