AMD Stretches Server DRAM with Flash Extended Memory
AMD has introduced a new technology that extends server DRAM capacity by using flash memory as an extended memory tier. This approach allows systems to access larger memory pools at lower cost, leveraging flash storage to supplement traditional DRAM for memory-intensive workloads.
Background
AMD is developing a technology that lets servers treat a pool of flash storage (similar to SSDs) as if it were main memory (DRAM), bridging the gap between fast-but-expensive DRAM and slow-but-cheap storage. This matters because AI and large-scale database workloads increasingly demand far more memory than a single server can economically fit in DRAM. The approach — sometimes called CXL-attached memory expansion — would allow servers to access terabytes of memory at speeds much closer to DRAM than to a traditional SSD, while costing far less per gigabyte than adding more RAM sticks. AMD is competing with Intel and other chipmakers who are pursuing similar memory-tiering and disaggregated-memory strategies. The article likely discusses how this "flash extended memory" works, what hardware changes it requires, and whether it's a genuine breakthrough or a niche solution for specific workloads.