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 — without the hours of GPU training that prior methods required.

Nvidia says it can shrink 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 ...
Electronic Design

Hardware Compression Works at the Memory Cache Level

How lossless data compression can reduce memory and power requirements. How ZeroPoint’s compression technology differs from the competition. One can never have enough memory, and one way to get more ...