@misc{indiciae21d58015b4f6,
  title = {Dissecting GPU Utilization for LLM Inference on Nvidia Hopper},
  author = {Mohammad Siavashi and Gerald Q. Maguire Jr. and Dejan Kostic and Marco Chiesa},
  year = {2026},
  url = {https://arxiv.org/abs/2609.12923},
  note = {Source identifier: 2609.12923}
}