Nvidia announces certified server systems for AI workloads
Nvidia Corp. today announced the first batch of what it calls “Nvidia-Certified Systems” for organizations that want to run artificial intelligence workloads at scale.
The company said it has worked with server makers, including Dell Technologies Inc., Hewlett Packard Enterprise Co., Gigabyte Technology, Inspur Group and Super Micro Computer Inc., to certify that their systems meet its best design practices and can deliver an optimal performance for the most advanced machine learning and data analytics tasks.
The new systems are powered by Nvidia’s most advanced A100 graphics processing unit and, combined with high-speed Mellanox network adapters, give companies different options to run their AI workloads in corporate data centers or at the network edge.
Adel El Hallak, director of product management for Nvidia’s GPU Cloud, said in a press briefing that “AI has gone mainstream” and that customers are looking for assured functionality, performance, scalability and security.
“Doing AI at scale is difficult and up until now has been a do-it-yourself program,” El Hallak said. He added that the newly certified systems will help to “turn something that was previously complex into something that’s turnkey.”
The company explained that each of the certified systems has been tested across a broad range of AI workloads, ranging from jobs that require multiple compute nodes to those that need just a fraction of the power of a single GPU. Each of them have been optimized to run AI applications from Nvidia’s NGC catalog, which is the company’s hub for GPU-optimized AI applications.
Certification involves passing tests on AI workloads including deep learning training and inference, machine learning algorithms, intelligent video analytics and network and storage offload, using the most popular AI frameworks in the NGC catalog, Nvidia explained.
“We’re tapping into real workloads people use and we’re testing at scale,” El Hallak said.
Analyst Holger Mueller of Constellation Research Inc. told SiliconANGLE that Nvidia is on the path to succeed in AI with both its software and its hardware, and that “certified systems” are a classic and proven strategy for the latter.
“Company executives like platform certified systems as they ensure viability and portability,” Mueller said. “More importantly, this makes it possible for enterprises to run AI locally, in a future-proof way, as Nvidia has managed to have its platforms supported in all major public clouds as well. The result is that Nvidia becomes a compute platform for AI, enabling workload portability between on-premises systems and the public cloud for next-generation applications.”
Nvidia said 14 systems have been certified to provide accelerated computing at launch, including the Dell EMC PowerEdge R7525 and R740 rack servers, Gigabyte’s R281-G30, R282-Z96, G242-Z11, G482-Z54, G492- Z51 systems, HPE’s Apollo 6500 Gen10 System and HPE ProLiant DL380 Gen10 Server, Inspur’s NF5488A5 server and the Supermicro A+ Server AS -4124GS-TNR and AS -2124GQ-NART.
Each of those systems carries the “Nvidia-Certified Systems” badge that certifies they meet the company’s best design practices and can handle the most demanding AI workloads. Enterprise support is available across the full software stack, including support for open-source code.
Nvidia said about 70 systems from 11 system makers are currently engaged with its program, and that it expects to announced more Nvidia-Certified Systems soon.
Image: Nvidia
A message from John Furrier, co-founder of SiliconANGLE:
Your vote of support is important to us and it helps us keep the content FREE.
One click below supports our mission to provide free, deep, and relevant content.
Join our community on YouTube
Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.
THANK YOU