UPDF AI

Aquabolt-XL: Samsung HBM2-PIM with in-memory processing for ML accelerators and beyond

J. Kim,Shinhaeng Kang,17 Authors,N. Kim

2021 · DOI: 10.1109/HCS52781.2021.9567191
IEEE Hot Chips Symposium · 42 Citations

TLDR

Using PIM to overcome memory bottleneck • Although various bandwidth increase methods have been proposed, it is physically impossible to achieve a breakthrough increase.