HPE Unveils AI-Optimized Solutions Powered by AMD for Advanced Model Training

HPE
Image Courtesy: HPE

Hewlett Packard Enterprise (NYSE: HPE) has introduced the HPE ProLiant Compute XD685, a powerful new system designed to tackle complex AI model training. Featuring the latest 5th Gen AMD EPYC™ processors and AMD Instinct™ MI325X accelerators, this system is built to accelerate the development of large language models, natural language processing, and multi-modal training.

The ProLiant Compute XD685 delivers high-performance, secure, and energy-efficient AI clusters, making it easier for organizations to deploy cutting-edge AI solutions. The growing demand for AI’s transformative potential in sectors like healthcare, workforce productivity, and climate research is driving the need for advanced computing solutions. To meet this demand, AI providers, governments, and large-scale model builders need flexible, high-performance technology that can be quickly deployed.

“Training massive language models requires robust, scalable computing power and specialized services, and it is uniquely positioned to deliver these capabilities,” said Trish Damkroger, Senior Vice President and General Manager of HPC & AI Infrastructure Solutions at the company. “Our collaboration with AMD in the new ProLiant Compute XD685 brings cutting-edge innovation to the market, helping AI developers achieve faster breakthroughs while maintaining energy efficiency and security.”

The HPE ProLiant Compute XD685 offers a modular 5U chassis design that can be customized to support various GPUs, CPUs, and cooling methods. By incorporating AMD CDNA™ 3 architecture and AMD Instinct™ MI325X accelerators, the system provides outstanding AI performance with reduced power consumption. The optional direct liquid cooling feature enhances efficiency, making it an ideal solution for organizations seeking to meet sustainability goals while managing growing power demands.

The HPE ProLiant Compute XD685 offers high performance with fewer GPUs, utilizing AMD Instinct™ MI325X accelerators based on CDNA™ 3 architecture, which deliver exceptional AI performance for both training and inference tasks. Its high-bandwidth HBM3E memory ensures improved efficiency. Optional direct liquid cooling, supported by HPE’s expertise in cooling technology, enhances sustainability while lowering operational costs.

To ensure swift deployment, HPE provides installation services that minimize downtime, accelerating time to market for AI clusters. The XD685 also features enhanced security through HPE iLO technology, offering protection from supply chain to system-level threats.

Designed for optimized rack density, the 5U chassis accommodates up to eight AMD Instinct™ MI325X or MI300X accelerators and the latest AMD EPYC™ processors. Additionally, HPE Performance Cluster Manager simplifies cluster management, allowing for rapid setups and proactive system maintenance.

In addition to launching the XD685, HPE announced that its ProLiant Gen11 servers, equipped with AMD EPYC processors, have set 48 new world records for performance and power efficiency in enterprise computing workloads. The HPE ProLiant Compute XD685 is available for order now, with general availability expected in early 2025. Various configurations will be rolled out over the course of the year.

Hewlett Packard Enterprise (HPE) is a global leader in edge-to-cloud technology, helping organizations drive innovation and improve performance by unlocking the value of their data. With offerings that span cloud services, computing, AI, and more, HPE delivers intelligent, open technology solutions that transform the way people live and work.