GPU-accelerated Microsoft Cognitive Toolkit, from Microsoft Azure to the cloud,
Available on-premise on NVIDIA DGX-1

NVIDIA announced a collaboration with Microsoft to accelerate AI advancements in the enterprise. This will enable more companies to join the AI revolution, the company said.

This collaboration will enable enterprises to have an AI platform that spans from their data centers to the Microsoft cloud, using optimized enterprise AI frameworks built to run on NVIDIA Tesla GPUs in Microsoft Azure or on-premises.

“We are at the dawn of an era of AI computing, powered by the power of a new computing model,” said Jen-Hsun Huang, co-founder and CEO of NVIDIA. “The close collaboration between NVIDIA and Microsoft means that enterprises have the fastest AI platform, the most scalable solution based on NVIDIA DGX-1 and Tesla GPUs, and the best tools to transform any product or service.”


NVIDIA co-founder and CEO Jen-Hsun Huang announces collaboration with Microsoft at SC16

“We’re working to make it possible for every organization to use AI to build smarter products and solve the world’s most pressing problems,” said Harry Shum, executive vice president of Microsoft’s AI and Research Group. “By working closely with NVIDIA and leveraging the power of GPU-accelerated systems, we’ve made Cognitive Toolkit and Microsoft Azure the fastest, most versatile AI platform available, so AI can now be applied to any business.”

By Nvidia and MicrosoftThis jointly optimized platform runs the new Microsoft Cognitive Toolkit (formerly CNTK) on NVIDIA GPUs, including the NVIDIA DGX-1™ supercomputer with NVLink™ technology and Pascal GPUs, and Azure N-Series virtual machines, currently in preview. This combination delivers unprecedented performance and ease of use when leveraging data for deep learning.

This allows businesses to leverage AI to make better decisions, launch new products and services faster, and deliver better customer experiences. This is why AI is being applied across all industries, with the number of companies NVIDIA collaborates with on deep learning growing 194x in just two years, to over 19,000. Industries such as healthcare, life sciences, energy, financial services, automotive, and manufacturing are already leveraging deep insights based on massive amounts of data.

The Microsoft Cognitive Toolkit trains and evaluates deep learning algorithms faster than any other toolkit. It also efficiently scales across a variety of environments, from CPUs to GPUs and multiple machines, while maintaining accuracy. NVIDIA and Microsoft have been working closely to accelerate the Cognitive Toolkit on GPU-based systems and the Microsoft Azure cloud.

NVIDIA and Microsoft plan to continue collaborating to optimize the Cognitive Toolkit for NVIDIA GPUs within Azure and as part of a hybrid cloud AI platform when connected to NVIDIA DGX-1 on-premises.