Server-Scale Programmable Photonic Fabric to Interconnect Accelerators Within Servers (Cornell University, Lightmatter)

Server-Scale Programmable Photonic Fabric to Interconnect Accelerators Within Servers (Cornell University, Lightmatter)

A new technical paper titled “Morphlux: Programmable chip-to-chip photonic fabrics in multi-accelerator servers for ML” was published by researchers at Cornell University and Lightmatter. Abstract “We optically interconnect accelerator chips (e.g., GPUs, TPUs) within compute servers using newly viable programmable chip-to-chip photonic fabrics. In contrast, today, commercial multi-accelerator compute servers that are workhorses of ML, use electr…