Two Papers Accepted at DATE 2027
We are pleased to announce that two of our papers have been accepted for presentation at the Design, Automation and Test in Europe Conference (DATE 2027).
Both papers address challenges in racetrack-memory computing, ranging from efficient and reliable neural-network execution to data-placement optimization.
Misalignment-aware Binarized Neural Networks on Racetrack Memories
Leonard David Bereholschi, Simon Kurz, Mikail Yayla, Asif Ali Khan, Kuan-Hsun Chen, and Jian-Jia Chen.This paper builds on our previous work presented at SAMOS and introduces the idea of misalignment awareness in the training phase.
SHIFTLESS: Spectral Seeding and Median-Guided Refinement for Racetrack-Memory Data Placement
Elijah Seth Cishugi, Asif Ali Khan, Kuan-Hsun Chen, and Marco Ottavi.This paper investigates data-placement optimization for racetrack memories and results from our ongoing collaboration across the University of Twente, TU Dresden, and Paderborn University.
Congratulations to all co-authors!