Benchmarking and Dissecting the Nvidia Hopper GPU Architecture
Published in International Parallel and Distributed Processing Symposium (IPDPS), 2024
This paper provides a comprehensive benchmarking and analysis of the Nvidia Hopper GPU architecture. Through micro-benchmarking and multi-level characterization, we reveal key architectural features, bottlenecks, and performance behaviors of Hopper, offering insights for GPU programmers and system researchers.
Recommended citation: Weile Luo, **Ruibo Fan**, Zeyu Li, et al., "Benchmarking and Dissecting the Nvidia Hopper GPU Architecture," in *Proceedings of the IEEE International Parallel and Distributed Processing Symposium (IPDPS)*, 2024.
Download Paper | Code | Download Bibtex