Nvidia is busy this week at the Computex 2021 Taipei virtual tech show, announcing an expansion of its Nvidia certified nascent server program, a range of new models of Nvidia Bluefield DPU servers and the upcoming availability of its core ordering platform which will include a subscription option. for its DGX SuperPods so that customers can try them out.
As part of its expanded Certified Servers program, which was initially unveiled in April at Nvidia’s own GTC21 conference, dozens of new servers are on the way. certified to run the full suite of Nvidia AI business software, giving customers more options for demanding workloads in traditional data centers or hybrid cloud infrastructures.
New partner servers using the company’s latest version were also announced. Bluefield-2 data processing units, including from ASUS, Dell Technologies, GIGABYTE, QCT and Supermicro.
Nvidia’s announcements also included the news that the Nvidia basic ordering platform, which is currently only available to Early Access customers after being unveiled at GTC21 in April, will be offered in conjunction with NetApp as a premium monthly subscription with Nvidia DGX SuperPod AI supercomputers and NetApp data management services. .
The new products are part of the company’s ongoing democratization of AI, Manuvir Das, head of corporate IT at Nvidia, said in a May 27 briefing with reporters on the news. .
“The work we’re doing with the ecosystem is really to prepare it now to fully participate in this next wave of AI democratization, where AI is used by every business on the planet rather than early adopters.” , said Das. “This is really the theme of what we talked about at Computex.”
This democratization includes using the software tools, libraries, frameworks and other pieces of Nvidia that the company has built and integrating them into what it calls the Nvidia AI enterprise software, Das said.
Servers certified to run Nvidia AI Enterprise software
This strategy is behind the company’s news that it is certifying its enterprise AI software suite on the latest wave of servers from partners such as ASUS, Advantech, Altos, ASRock Rack, ASUS, Dell Technologies, GIGABYTE, Hewlett Packard Enterprise, Lenovo, QCT and Supermicro. Currently, the number of Certified Servers is over 50. The Certified Server Program aims to help customers in the healthcare, manufacturing, retail and financial services industries find the consumer servers they have. need, depending on the company.
Nvidia systems include certifications for running VMware vSphere, Nvidia Omniverse Enterprise for design collaboration and advanced simulation, and Red Hat OpenShift for AI development, as well as data engineering and learning. Cloudera automatic.
Systems can be acquired in a wide range of prices and performance levels and can be equipped with a wide range of Nvidia hardware, including Tensor Core A100, A40, A30 or A10 GPUs as well as BlueField-2 DPUs or ConnectX-6 adapters.
A previous group of Nvidia certified servers were unveiled in April at GTC21.
New DPU Bluefield-2 servers
With that new series of servers equipped with DPU-2, Nvidia is expanding the lineup to give customers more options for finding the right servers for their needs, according to the company. Servers are intended for workloads, including software-defined networking, software-defined storage, or traditional enterprise applications, which can benefit from the DPU’s ability to accelerate, offload, and isolate workloads from within. infrastructure for networking, security and storage, according to Nvidia. DPU servers can also benefit systems running VMware vSphere, Windows, or hyperconverged infrastructure solutions for AI and machine learning applications, graphics-intensive workloads, or traditional business applications.
Nvidia’s BlueField DPUs are designed to move infrastructure tasks from the processor to the DPU, making more server processor cores available to run applications and increasing server and data center efficiency, the report says. society.
BlueField-2 DPU accelerated servers are expected this year.
Nvidia Base Command and SuperPod subscriptions
For customers, the idea behind Nvidia Base Command Platform and its related DGX SuperPod subscription option is that it can help companies move their AI projects from prototypes to production faster.
The Base Command software platform, designed for large-scale, multi-user, multi-team AI development workflows hosted on-premise or in the cloud, enables researchers and data scientists to work simultaneously on computing resources accelerated, according to Nvidia.
The basic cloud-hosted ordering platform will be offered in conjunction with NetApp, including an option to try a subscription-based DGX SuperPod, Das said. NetApp all-flash storage is also included. More information on these options will be released later this week, according to Nvidia.
The basic ordering platform works with DGX systems and other Nvidia accelerated computing platforms, such as those offered by its cloud service provider partners. Many features of Base Command were unveiled by the company at GTC21. Base Command Manager is used to manage resources on an on-premises DGX SuperPod. Base Command Platform provides a wide range of controls to manage workflows from anywhere and allows the subscription service hosted with NetApp to be offered.
Das said this was the first time DGX SuperPod subscriptions had been offered and the move came after being requested by customers. “All the hardware is hosted by Nvidia in the Equinix data centers,” he said. “And customers can go into that environment and rent access to a SuperPod or a smaller part of the SuperPod, and they can rent it for a few months at a time.”
For customers, this new option can provide a simple and easy-to-use experience for AI, Das said.
“What we’re doing here is we’re really lowering the barrier to entry to experience this state-of-the-art system and equipment, and we’re democratizing that way,” he said. It is expected that once customers try SuperPods, they will buy theirs and use them more widely, he added.
So far, the basic Nvidia ordering platform with NetApp is only available to early access customers. The monthly subscription price starts at $ 90,000.
Analysts on the latest Nvidia news
So what do industry analysts think of Nvidia’s Computex ads?
“Nvidia is clearly moving up the value chain, from chips and systems to software and possibly data centers,” Karl Freund, Founder and Senior Analyst at Cambrian AI Research, Told Corporate AI. “The announcements will appeal to companies starting their AI journey, with a wide enough range of software to develop, manage and collaborate on AI applications. “
And while booting up on a cloud instance of a $ 90,000-per-month DGX SuperPod may seem rich, it provides an easy ramp for customers, with no hardware to buy and install, and no additional software needed. -he declares.
“Eliminating the hassle will help businesses get started with AI,” Freund said. “When ready for production, these Base Command customers can purchase DGX systems, systems from their server vendor, or deploy them to public clouds, all with the same software. “
Another analyst, James kobielus, senior research director for communications and data management at data research, training and analytics consultancy TDWI, said he was impressed by Nvidia’s willingness to help customers produce the full line of its AI software.
Most notable is the Core Control Platform, which provides cloud-based access for AI development teams to Nvidia’s most powerful DGX SuperPod AI supercomputer, as well as the management suite of data from NetApp, ”Kobielus said. “Once this offering becomes available on the Google Cloud Marketplace later in the year, I expect many companies to select Nvidia Base Command Platform for their development of machine learning applications to be deployed in cloud environments. hybrid and to run various Nvidia Certified Systems from Nvidia Partners in support of high performance enterprise applications.
Bob sorensen, a Hyperion Research analyst, said Corporate AI that Nvidia’s DPU servers provide HPC server vendors with opportunities to develop new capabilities for intelligent, targeted compute capabilities where customers need it.
“The added benefit is that these devices can help offload data management responsibilities from processors, freeing them up for more processor-relevant tasks,” Sorensen said. “Indeed, one could argue that DPUs like these could be the harbinger of a new form of composable computing-based HPC design, which seeks to break down and distribute discrete server functions over discrete server functions. specific smart devices dispersed in a traditional HPC architecture. “
Rob Enderle, Senior Analyst at Enderle Group, said Nvidia appears to be setting itself up to make a significant breakthrough in enterprise servers. “Their DPU technology is mind blowing,” Enderle said. “This frees up significant CPU resources, which can then be applied to other projects. This is particularly ideal for cloud solutions where you need great flexibility.
The importance of this technology is notable, he said.
“This is just the start of what should be the biggest effort to move technology from X86 servers in over a decade,” said Enderle. “This initiative is just the start and coupled with their Arm HPC development kit with Gigabyte, it anticipates an endgame where x86 will become obsolete.”