1 paper
Dowon Son, Yonggon Park, Hyunuk Cho +5
This work investigates the potential benefits and technical challenges of using high-bandwidth flash (HBF) for large language model (LLM) inference. HBF has gained increasing atten…