@article{hpuhighbandwidthprocessingunitfor,
  title = {HPU: High-Bandwidth Processing Unit for Scalable, Cost-effective LLM Inference via GPU Co-processing},
  author = {Myunghyun Rhee and Joonseop Sim and Taeyoung Ahn and Seungyong Lee and Daegun Yoon and Euiseok Kim and Kyoung Park and Youngpyo Joo and Hosik Kim},
  year = {2025},
  eprint = {2504.16112},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2504.16112v1},
}