NVIDIA partners up with Microsoft and IBM, also launches the DGX SATURNV
NVIDIA's endeavor into the realm of supercomputing, artificial intelligence, and Deep Learning recently passed several milestones, as the graphics company officially broke ground with Microsoft and IBM (respectively) in the field of artificial intelligence, as well as announced the launch of their new and more power efficient DGX SATURNV supercomputer.
The next step for AI with Microsoft
NVIDIA's partnership with Microsoft comes with the latter's purchase of a DGX-1 unit from NVIDIA, in their effort to optimize their first purpose-built AI framework for Microsoft Azure. Additionally, the procurement of NVIDIA's new supercomputer will also allow Microsoft to provide their Cognitive Toolkit across Azure and on-premise.
“We stand at the beginning of the next era, the AI computing era, powered by a new computing model,” Jen-Hsun Huang, CEO and founder of NVIDIA, said about the collaboration. “Our close collaboration with Microsoft means companies have the fastest AI platform, the most scalable solution with NVIDIA DGX-1 and Tesla GPUs, and the best tools to transform any product or service."
Microsoft's Cognitive Toolkit is designed to train and evaluate deep learning algorithms. Originally done with the use of CPU power, the presence of the DGX-1 now means that Microsoft can utilize the prowess that is GPU-driven calculations, as well as the full power of NVIDIA's Pascal GPUs with NVLink.
Partnering up with IBM
On the other side of the fence, IBM and NVIDIA went into collaboration in order to create new Deep Learning tools that would enable computers to think and learn as a human would, but at a much faster pace.
Known as IBM PowerAI, the tool runs on IBM's recently announced AI-built server that, once again, features NVIDIA's NVLink interconnect technology. Specs-wise, the server reportedly provides more than twice the performance over comparable servers with four Telsa P100 GPUs running AlexNet with Caffe. Not only that, but that same configuration can also outperform eight Tesla M40 GPU-based configurations.
"PowerAI democratizes deep learning and other advanced analytic technologies by giving enterprise data scientists and research scientists alike an easy to deploy platform to rapidly advance their journey on AI,” said Ken King, General Manager, OpenPOWER IBM. "Coupled with our high performance computing servers built for AI, IBM provides what we believe is the best platform for enterprises building AI-based software, whether it’s chatbots for customer engagement, or real-time analysis of social media data."
Clients who are using IBM's Deep Learning solutions include groups such as the Human Brain Project, Nimbix, the city of Yachay in Ecuador, and SC3 Electronics.
As per our earlier report, NVIDIA also announced their new DGX SATURNV supercomputer. To keep things brief, the DGX SATURNV is powered by their new Tesla P100 GPUs (NVIDIA failed to specify how many exactly), delivering as much as 9.46 GFLOPs per watt, which translates as a 42 percent improvement over the 6.67 GFLOPs that's delivered by the most efficient machine on the Top500 list of supercomputers that was released back in June 2015. How many calculations does that equate to? Well, all that power allows the DGX SATURNV to perform one quintillion floating-point operations per second.
Energy efficiency is a key component of the DGX SATURNV, with it being 2.3 times more efficient than Intel's own Xeon Phi Knights Landing.
As you may well know, the DGX SATURNV is not the first supercomputer to be powered by NVIDIA's Tesla P100 GPU. That honor goes to the DGX-1, and that supercomputer has been adopted by higher learning and research institutions around the world, chief among which being Monash University in Melbourne, Australia, as well as researchers at Stanford and New York University for Open AI.
For more news from NVIDIA, follow us here.