InfAMAX: Bridging the Compute-Memory Gap in Intel AMX for Efficient LLM Inference

Kyeonghyeon Ryu, Yunhyeong Jeon, Gyeongseo Park, Minho Kim, and Daehoon Kim*. IEEE Computer Architecture Letters (CAL), 2026
DOI
10.1109/LCA.2026.3693303

Abstract

Keywords

Intel AMX, Cache Optimization, Large Language Model Inference, NUMA-aware Systems, Performance Analysis.