Dell has announced the expansion of its AI edge server lineup by adding the latest Intel Gaudi 3 AI to the Dell PowerEdge product line. It is expected to serve as another option for companies deploying services and products based on generative AI models.

Dell Technologies announced on the 15th that it is adding support for the 'Intel Gaudi 3' AI accelerator to its high-performance AI server, 'Dell PowerEdge XE9680'.

Dell expects that the expanded lineup of XE9680, designed to effectively support data-intensive and complex workloads, will provide more choices to enterprise customers considering utilizing AI.

The Dell PowerEdge XE9680 is a model equipped with eight GPUs on an x86 server architecture, providing performance specialized for AI operations. With the integration of Intel Gaudi 3 accelerator into Dell PowerEdge, the system is expected to be tailored to the computing requirements associated with generative AI workloads. This is part of Dell's strategy to provide AI acceleration infrastructure that has both versatility and performance.

It has been reported that the 'PowerEdge XE9680' server model equipped with the Intel Gaudi 3 AI accelerator is planned to be released domestically within this year.

Designed with a focus on reliability and scalability, the XE9680 adds Gaudi3 accelerators to improve data throughput with up to 32 DDR5 memory DIMM slots in server configuration options, and provides expanded connectivity and bandwidth with 16 EDSFF3 flash storage drives and eight PCIe Gen 5.0 slots.

Equipped with two 4th generation Intel Xeon Scalable processors supporting up to 56 cores per processor, it is optimized for data processing and analysis of high-difficulty AI/ML workloads.

The XE9680 combines advanced processing capabilities with an efficient air-cooling design. Powered by the Gaudi3 accelerator, the PowerEdge XE9680 offers new networking capabilities directly coupled to the accelerator via six OSFP 800GbE ports. This enables direct connectivity to external accelerator fabrics without the need to deploy external NICs in the system, simplifying the infrastructure and reducing the total cost of ownership and complexity of the infrastructure.

The Intel Gaudi 3 AI accelerator has the essential specifications for generative AI workloads, including 64 customizable and programmable Tensor Processor Cores (TPCs), 128 GB of HBMe2 memory capacity, 3.7 TB of memory bandwidth, and 96 MB of onboard SRAM.

Additionally, Gaudi3 has the advantage of an open ecosystem and partnership-based optimization and model library framework support. Supports easy migration with development tools that simplify the transition of existing code bases.

The Intel Gaudi-3 specialized media decoder is designed for AI vision applications. It supports a wide range of preprocessing tasks, simplifying video-to-text conversion and improving the performance of enterprise AI applications.

“Dell is pushing the boundaries of AI development and continuously setting new standards through a broad collaboration strategy,” said Kim Kyung-jin, president and CEO of Dell Technologies Korea. “We plan to focus on helping customers accelerate their AI journey, prepare for the performance demands of the future, and continue this journey safely, no matter where they are.”

Meanwhile, Dell was recently named a leader in the AI field in the Forrester Wave report. It received high scores in both strategy and offering aspects, being evaluated as driving generative AI performance in an end-to-end environment by providing comprehensive solutions that enable IT and data scientists to apply AI and increase productivity.