一作MICRO 2026Oral· Systolic & Dataflow Innovations for LLM/GNN Inference
One Chiplet to Build Them All: Cost-Efficient Hardware Specialization for Stage-Separated LLM Inference
Haonan Zhu · Mengdi Wang · Haoyi Jiang · Danqing Zhang · Lian Liu · Shixin Zhao · Yutian Zhou · Yudong Pan · Tianhua Han · Yinhe Han · Ying Wang
效率与系统OpenReview