Compute Express Link
An open standard, first released in 2019, that runs over PCIe wiring and lets processors share memory with accelerators and memory expanders.
Also: CXL
Compute Express Link reuses the physical layer of PCI Express but adds protocols for cache-coherent memory sharing. A processor and an accelerator can then work on the same data without copying it, and servers can add memory on cards or in separate boxes, pooling it between machines. Intel led the first specification in 2019, and the CXL Consortium now includes most of the industry. It is supported in recent Xeon and EPYC server processors, and memory makers sell CXL memory modules.