Dell integrates more AI power into updated server portfolio

As demand for artificial intelligence (AI) continues to grow, there is a corresponding growth in demand for hardware that has been optimized for machine learning (ML) and AI workloads.

It’s an opportunity that Dell Technologies is looking to capture, with its new lineup of servers announced today at the Dell PowerEdge .Next event. The new PowerEdge server lineup includes hardware for general purpose workloads, edge and cloud, as well as purpose-built systems for AI/ML. The updated Dell server lineup benefits from CPU improvement from the Intel Xeon 4th generation scalable processors that were announced last week. 

Dell is also integrating technology from Nvidia, including GPUs as well as support for the Nvidia AI Enterprise suite of tools, into its PowerEdge lineup.

“Given the explosive growth of data, and the need for businesses to gain insights from that data, AI implementation is one of the fastest growing segments in the industry,” Rajesh Pohani, vice president of portfolio and product management for PowerEdge, HPC and core compute, told VentureBeat. “To fulfill those needs, Dell’s next-generation PowerEdge server portfolio has been built for AI, and specifically includes new purpose-built AI-enabled servers for most complex training workloads.”

Nvidia and Dell extend partnership for AI

Dell and Nvidia also announced that the full Nvidia AI stack is now available on 15 new Dell PowerEdge servers. 

In a press briefing, Justin Boitano, VP of enterprise and edge computing at Nvidia, said the new range of Nvidia-accelerated Dell PowerEdge systems draws from Nvidia’s full stack of AI, including H100 GPUs, BlueField data processing units (DPUs) and the Nvidia AI Enterprise software suite.

“Dell is a very special partner to us — they’re the largest server OEM in the world — and I think the exciting thing about this announcement is that it’s the broadest portfolio of servers that we’ve ever launched together,” Boitano said. “It’s going to provide a range of systems from training systems to volume servers to run scale-out workloads like Spark in a data center, as well as AI inference workloads even at the edge.”

Boitano said that new Dell PowerEdge servers that integrate the H100 GPU will help organizations process data faster, which will enable more AI models to make it into production on Dell hardware. He also highlighted the potential of the BlueField DPUs. 

“Nvidia BlueField DPUs can offload and accelerate the networking and operating system stacks in the data center, which means businesses using DPUs can cut data center energy use by up to 25%, potentially saving them millions of dollars in energy costs,” Boitano said. 

Dell embeds AI deep across its server portfolio

In a Dell press briefing, Pohani said that his company’s new lineup of PowerEdge servers are being targeted at multiple use cases, including general compute, edge and cloud.

The Dell PowerEdge XE servers portfolio is purpose built for AI/ML workloads, but they aren’t the only servers where Dell is embedding AI capabilities. Dell’s AI capabilities are being bolstered with the Nvidia partnership, as well as functionality it gains with Intel’s latest server CPU technology.

The Intel Xeon 4th generation scalable CPU that is part of the Dell PowerEdge server update integrates a series of capabilities, including Intel Deep Learning Boost and Intel Advanced Matrix Extensions (AMX), that can help to support AI/ML workloads, even without a discrete GPU present.

“Our entire portfolio is strongly AI-enabled with over 80% of our products supporting discrete GPUs, and all of our servers supporting CPU-based AI, whether it’s in the data center or the edge,” Pohani said.