AMD, Lightelligence, MemVerge, NVIDIA, Samsung & Supermicro Optical Connectivity

Optical CXL in the Datacenter

Subscribe to the Memory Fabric Forum Community News

Receive information about events, online courses, new products and solutions, and early access programs.

Photowaveâ„¢ Optical Interconnect Demo at FMS

Interconnects remote devices using CXL over lower-latency fiber optic cable, extending reach to enable memory pooling at pod scale and beyond. This allows for scalable CXL fabrics in the composable data center.

Participate in the Memory Fabric Forum Community

The content in this website, our events, and in our news is crowd-sourced from people like you. Contact us about speaking at Memory Fabric Forum events, contributing blogs or an online course.


Average GPU-PCIe-Transmit-Traffic:

CXL: 882MB/s, System Memory 857MB/s, Disk: 582MB/s, MemVerge: 493MB/s

Average GPU-PCIe-Receive-Traffic:

CXL: 2365MB/s, System Memory: 2609MB/s, Disk: 1887MB/s, MemVerge: 2173MB/s

CXL Mem. Achieves ~2.4x
Higher Throughput than Disk