PowerBI Data Analyst - Create visualizations and dashboards from scratch
Get 20% off all career paths from fullstack to AI
Overview
Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Learn about optical connectivity technologies essential for scaling AI xPUs beyond single rack configurations in this conference talk by Rang-Chen (Ryan) Yu, VP Marketing at TeraHop. Explore the economic drivers behind multi-rack AI compute cluster scaling and examine various low-power optical technologies including TRO (Tunable Receive Optics), LPO (Linear Pluggable Optics), NPO (Near Packaged Optics), and CPO (Co-Packaged Optics). Discover innovative approaches such as slow and wide technologies utilizing massive numbers of uLED channels modulated at low speeds to achieve low picojoule-per-bit energy efficiency. Analyze the advantages and disadvantages of each optical technology solution and identify the most promising candidates for next-generation multi-rack xPU scale-up applications in large AI computing environments.
Syllabus
Optical Technology Considerations for AI xPU Scale Up
Taught by
Open Compute Project