SMOOTH: Hardware-Assisted Fine-Grained On-Chip Memory Management for Efficient On-Device LLM Inference

Seulki Kim, Bokyeong Kim, Kyeonghyeon Ryu, Yeji Jung, Hwanjun Lee, Sungju Kim, Yunhyeong Jeon, and Daehoon Kim^* IEEE/ACM International Symposium on Computer Architecture (ISCA), 2026