Nvidia shrinks LLM memory 20x without changing model weights

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory costs and time-to-first-token by up to 8x for multi-turn AI applications.

New KV cache compaction technique cuts LLM memory 50x without accuracy loss

MIT researchers developed Attention Matching, a KV cache compaction technique that compresses LLM memory by 50x in seconds — ...
EDN

Last-level cache has become a critical SoC design element

LLC, positioned between external memory and internal subsystems, stores frequently accessed data close to compute resources.
Semiconductor Engineering

A Primer On Last-Level Cache Memory For SoC Designs

System-on-chip (SoC) architects have a new memory technology, last level cache (LLC), to help overcome the design obstacles of bandwidth, latency and power consumption in megachips for advanced driver ...