[ISCA 2025] Seungjae Moon and Junseo Cha’s paper on Hybe: GPU-NPU Hybrid System for Efficient LLM Inference with Million-Token Context Window is accepted

Congratulations!

We have a paper accepted to ACM/IEEE International Symposium on Computer Architecture (ISCA), 2025

“Hybe: GPU-NPU Hybrid System for Efficient LLM Inference with Million-Token Context Window”

  • *Seungjae Moon, *Junseo Cha, Hyunjun Park, Joo-Young Kim (*equal contribution)