Nvidia NGC To Supercharge AI On VMware Cloud, vSphere And Tanzu

‘We are joining forces now to say, ‘hey, use this same platform and run your workloads,’ which essentially means that we can bring AI to every enterprise on top of the platform that they are already using,’ VMware executive Krish Prasad says of the upcoming integration of Nvidia’s NGC hub for GPU-accelerated software.


Nvidia is expanding its partnership with VMware and integrating its Nvidia NGC software hub with VMware vSphere, Cloud Foundation and Tanzu, giving virtualization users faster and better ways to deploy GPU-accelerated AI applications in existing hybrid cloud environments.

The two companies announced the expanded partnership Tuesday at VMworld 2020, saying that VMware’s new Project Monterey, a new kind of hybrid cloud architecture, can take advantage of Nvidia’s SmartNIC data processing units to deliver purpose-built performance for AI, machine learning, high-throughput and data-hungry applications.

[Related: VMware vSphere Now Supports AMD EPYC’s ‘Powerful’ SEV Features]

Sponsored post

The companies did not say when VMware’s Project Monterey or the Nvidia NGC integration with VMware platforms would become available.

The announcement was made as Nvidia takes on Intel in pursuit of becoming the “world’s premier computing company,” which the chipmaker plans to achieve by acquiring British chip designer Arm for $40 billion — a move that would allow it to expand beyond GPUs and interconnects into CPUs. It also builds off a partnership announced between VMware and Nvidia last year that allows users to run GPU-accelerated virtual machines for AI, high-performance computing and analytics workloads.

“Nvidia and VMware will help customers transform every enterprise data center into an accelerated AI supercomputer,” Nvidia CEO and co-founder Jensen Huang said in a statement. “Nvidia DPUs will give companies the ability to build secure, programmable, software-defined data centers that can accelerate all enterprise applications at exceptional value.”

The two companies said Nvidia NGC will accelerate the deployment and simplify the management of AI workloads for a variety of industries, from health care to manufacturing, in containers and virtual machines while allowing them to run alongside enterprise applications on VMware’s platforms.

“We are joining forces now to say, ‘hey, use this same platform and run your workloads,’ which essentially means that we can bring AI to every enterprise on top of the platform that they are already using,” Krish Prasad, senior vice president and general manager of VMware’s cloud platform business, said in a pre-briefing with journalists.

With the integration of Nvidia NGC, which stands for Nvidia GPU Cloud, users of vSphere, Tanzu and VMware Cloud Foundation will gain access to Nvidia’s containers, AI models and industry-specific software development kits that are optimized for its GPUs in cloud environments.

“As much as people may think of Nvidia as a hardware company, we are more so a software company today, and the reason is because we have spent a lot of energy on developing the algorithms for artificial intelligence that can be run on the GPU and take advantage of the GPU,” Manuvir Das, head of enterprise computing at Nvidia, said. “We have worked with the ecosystem to modify and tune existing AI algorithms so that they can run well on GPUs and really exploit the massive parallel computing power that GPUs make available. And in the process, we have built a large ecosystem.”

Matthew DuBell, a consulting solutions architect for the Business and Analytics Advisors group at World Wide Technology, a top Nvidia and VMware partner in St. Louis, Mo., said the upcoming NGC integration into VMware’s platforms will make implementation of AI applications easier for organizations, particularly those with stricter protocols or those who lack open-source talent.

“I think in the AI space, the one thing we haven‘t seen is a lot of people come out with an at-scale machine learning or AI platform,” he said. “A lot of times we say AI Ops, we’re talking more about IT service, understanding their performance, their infrastructure. Whereas if I’m a developer, how do I go from wild-west development to QA to at-scale production seamlessly? The more that’s integrated, the better that is for those developers.”

Combined with VMware’s Bitfusion solution, which enables GPU virtualization for AI workloads, the Nvidia NGC integration will also play well with the need for organizations to drive down hardware costs and use virtualization to simplify the deployment of high-performance applications for employees working remotely, according to DuBell. It’s a play that makes sense, especially right now, because of the coronavirus pandemic forcing many to work from home.

“Given the current world state, most of my customers are asking, ‘how do I do really powerful engineering without buying several thousand dollars’ worth of desktops and sending them out to people‘s homes?’ So being able to have that integrated service with VMware’s expertise and Bitfusion, to me, makes complete sense,” he said

As for VMware’s support for Nvidia’s SmartNIC data processing units, specifically the Nvidia BlueField-2, for its Project Monterey data center architecture, DuBell said such technology will become increasingly important as organizations need to better insights into network performance and resolve bottlenecks. That’s something he has encountered, for instance, in deploying multiple Nvidia DGX systems.

“I have a lot of clients buy one or two DGX-1s or [DGX-2s], and they did interesting, unique things with that when they really want to go to scale and hook up five, 10, 20 of them. That‘s where the network performance and insight became critical,” he said.