Facebook Unveils Its New AI Hardware Under their latest AI Server Big Sur

By CIOReview | Thursday, December 17, 2015
773
1209
246

FREMONT, CA: Deep Learning is a new area of Machine Learning research, which has been introduced with the objective of moving Machine Learning closer to Artificial Intelligence. It is the science of getting computers to act without being explicitly programmed. Facebook recently announced that its latest AI server designs has been made open source, continuing on the company’s course of letting others share new hardware designs. Facebook’s new AI hardware is powered by GPUs, processors built originally to run graphics but increasingly used in high-performance computing. The company’s latest AI server, called Big Sur, packs eight of Nvidia’s Tesla GPUs, each consuming up to 300 watts.

Big Sur servers are compatible with Open Rack, the custom server rack Facebook designed and contributed to the Open Compute Project. The AI server can run in the company’s regular data center, which is unusual for high-performance computers that often need specialized high-capacity cooling systems. Big Sur is built for training neural networks, key to nearly any type of modern AI research, for example, projects related to language, text or video.

“Leveraging NVIDIA’s Tesla Accelerated Computing Platform, Big Sur is twice as fast as our previous generation, which means we can train twice as fast and explore networks twice as large. And distributing training across eight GPUs allows us to scale the size and speed of our networks by another factor of two,” commented Kevin Lee and Serkan Piantino, Researchers, Facebook AI Research.