CASL’s paper accepted to IEEE Transactions on Computers
A research paper entitled “Hotness- and Lifetime-Aware Data Placement and Migration for High-Performance Deep Learning on Heterogeneous Memory Systems” has been accepted for publication in IEEE Transactions on Computers. The paper is co-authored by Myeonggyun Han, Jihoon Hyun, Seongbeom Park, and Prof. Woongki Baek at Computer Architecture and Systems Lab. (CASL), CSE, UNIST.
This work proposes HALO, hotness- and lifetime-aware data placement and migration for high-performance deep learning on heterogeneous memory systems. HALO dynamically places and migrates the tensors on heterogeneous memory nodes based on their hotness and lifetime characteristics. HALO significantly improves the performance and energy efficiency of various deep learning applications.
IEEE Transactions on Computers is one of the top-tier journals in the field of computer systems.