New memory architecture targets AI inference bottlenecks

Lightbits Labs Ltd. today is introducing a new architecture aimed at addressing one of the most stubborn bottlenecks in large-scale artificial intelligence inference: the growing mismatch between the ...

SOCAMM2 Is The Memory Standard AI Is Looking For

AI infrastructure can't evolve as fast as model innovation. Memory architecture is one of the few levers capable of accelerating deployment cycles. Enter SOCAMM2 ...
Electronic Design

New Architecture Enhances Memory Test and Repair

Generic test and repair approaches to embedded memory have hit their limit. Smaller feature sizes, such as 130 nm and 90 nm, have made it possible to embed multiple megabits of memory into a single ...
VentureBeat

GAM takes aim at “context rot”: A dual-agent memory architecture that outperforms long-context LLMs

For all their superhuman power, today’s AI models suffer from a surprisingly human flaw: They forget. Give an AI assistant a sprawling conversation, a multi-step reasoning task or a project spanning ...
Semiconductor Engineering

RPU: A Chiplet-Based Architecture To Address The Challenges of the Modern Memory Wall (Harvard University)

A Reasoning Processing Unit”. Abstract “Large language model (LLM) inference performance is increasingly bottlenecked by the memory wall. While GPUs continue to scale raw compute throughput, they ...
insideHPC

Untether AI Unveils At-Memory Compute Architecture at Hot Chips

PALO ALTO — Untether AI, an at-memory computation company for artificial intelligence (AI) workloads, today announced at the HOT CHIPS 2022 conference its next-generation architecture for accelerating ...