NVIDIA and Microsoft Partner to Build Massive Cloud AI Computer

Today NVIDIA announced a partnership with collaboration with Microsoft to build one of the most powerful AI supercomputers in the world. The Super Computer will be powered by Microsoft Azure’s infrastructure combined with NVIDIA GPUs.

"Azure's cloud-based AI supercomputer includes powerful and scalable ND- and NC-series virtual machines optimized for AI distributed training and inference. It is the first public cloud to incorporate NVIDIA's advanced AI stack, adding tens of thousands of NVIDIA A100 and H100 GPUs, NVIDIA Quantum-2 400Gb/s InfiniBand networking and the NVIDIA AI Enterprise software suite to its platform."



Trying to do the math on the AI power of tens of thousands of h100 is a bit staggering. A petaFLOP equals one thousand million million (1015) floating-point operations per second. One h100 has a performance of 32 petaFLOPS. To put it in perspective According to DUKE University a human brain can hit up to 1000 petaflops. So something like every 30 h1100's cards is smarter than any of us... well, at least processing speed. That's insanely quick processing power, and at some point, I gave up trying to analogize and was drooling over what my FPS on COD Modern Warefare 2 would be. (No way I'm alone with that thought -- raise your hand if you're with me.)

The current fastest Super Computer is Frontier which runs on an AMD / Radeon / Linux combo and can hit 1.102 quintillion operations per second.