Nvidia Built Its Own Supercomputer

Believe it or not, Nvidia scored a place in the TOP500 most powerful supercomputers in the world. Dominated by China and US, the top is based on the Linpack benchmark. Now, Nvidia can boast about their in-house supercomputer, too #hardwaremagic

The DGXSaturnV is ranked 28th on the Top500 list of Supercomputers. It features 124 DGX-1 units based on the new Tesla Pascal P100 GPUs and 63.488GB of RAM. As you can imagine, with these numbers, Nvidia claims it’s more efficient than a Xeon Phi system. Specifically, Nvidia’s supercomputer delivers an a whooping 9.46 GigaFlops/Watt! This means a 42% improvement in efficiency over the most efficient machine on the Top500 list released this summer.

One of its first goals is going to be helping Nvidia design GPUs faster and more efficiently: “We’re also training neural networks to understand chipset design and very-large-scale-integration, so our engineers can work more quickly and efficiently. Yes, we’re using GPUs to help us design GPUs. Most importantly, SATURNV’s power will give us the ability to train — and design — new deep learning networks quickly.”