Close

Session

Research Manuscript: AI on Fire: Compute-in-Memory and Multiplication-Free Acceleration for the Next Era
DescriptionAI hardware is at an inflection point, demanding radical efficiency and performance leaps. This session dives into groundbreaking compute-in-memory (CIM) architectures, multiplication-free acceleration, and novel arithmetic paradigms that slash power and latency while boosting throughput. From LUT-based DNN accelerators to hybrid analog-digital computing and STT-MRAM-powered CIM, these innovations are setting AI hardware ablaze with new possibilities.
Event TypeResearch Manuscript
TimeTuesday, June 2410:30am - 12:00pm PDT
Location3003, Level 3
Topics
Design
Tracks
DES2A: In-memory and Near-memory Computing Circuits