As Samsung explains, most computing systems today are based on the von Neumann architecture. This sequential processing approach involves separate CPU and memory units with data constantly being shuffled between them, creating a bottleneck.
https://www.techspot.com/news/88672-samsung-develops-high-bandwidth-memory-onboard-ai-processing.html?utm_source=dlvr.it&utm_medium=blogger
https://www.techspot.com/news/88672-samsung-develops-high-bandwidth-memory-onboard-ai-processing.html?utm_source=dlvr.it&utm_medium=blogger