MIT Scientist Shrinks Active Memory Footprint by 50%
-
最新报道：MIT researchers led by Professor Srinivas Devadas have developed a breakthrough memory optimization technique that reduces active memory footprint by 50-75% while maintaining performance. The novel approach separates frequently accessed "hot" memory from less-used "cold" data, dynamically compressing the latter. This innovation could significantly lower energy consumption in data centers and improve efficiency for memory-intensive applications like AI and blockchain. The team's prototype demonstrated these substantial memory savings without compromising speed, potentially offering major cost reductions for cloud computing services and high-performance computing systems.
下一篇
Little Pepe Surpasses $3 Million Presale Milestone, BlockDAG Raises $332 Million
- Pepe Coin has surpassed a $3 million milestone in its presale, while BlockDAG has raised $...
24小时热点
热点专题
407366