Do you need a highly performant, secure and energy-efficient AI cluster? The HPE ProLiant Compute XD685 leverages a purpose-built architecture and is designed to accelerate large, complex AI model training and tuning workloads. Powered by eight of the latest accelerators from NVIDIA® or AMD, this server is optimized for use in large language model training, natural language processing, and multimodal training. Optional direct liquid cooling (DLC) helps organizations meet escalating power requirements, advance sustainability goals, and lower operational costs. Managed by HPE iLO, the server offers seamless management and robust security. HPE Services provides expertise for global installation and quick deployment of large AI clusters, offering stability and operational excellence.
Nouveautés
- Support for eight NVIDIA H200 SXM Tensor Core GPUs air-cooled or liquid-cooled, or eight NVIDIA Blackwell GPUs liquid-cooled, or eight AMD Instinct™ MI300X air-cooled
- Modular chassis designed by Hewlett Packard Enterprise. 5U for liquid-cooled and 6U for air-cooled configurations.
- 24x DDR5 6400 RDIMMs and 12 channels per CPU DDR5 w/ ECC
- Built-in security features with HPE iLO 6
- Seamless cluster management with HPE Performance Cluster Manager (HPCM)
- 2x 5th Generation AMD EPYC™ processors
Caractéristiques
Speed and Flexibility to Increase your Competitive Advantage
The HPE ProLiant Compute XD685 provides a choice of eight NVIDIA H200 Tensor Core or NVIDIA Blackwell GPUs or eight AMD Instinct™ MI300X accelerators, delivering leading AI performance for training and tuning tasks.
In liquid-cooled configurations, a 5U chassis designed by HPE helps you optimize pod deployment, leveraging a compact 8-node-per-rack arrangement.
HPE Performance Cluster Manager expedites initial cluster setup and software deployment, and provides continued resilience and reliability.
HPE Factory Express Services builds, integrates, validates, fully tests, and customizes your solution in the factory, facilitating quicker on-site deployment.
You get flexibility and agility with a customized IT environment that lets you run the right workload at the right place, at the right time. Whether you want a solution on-premises, or deploy a hybrid model, we offer flexible consumption and deployment models to suit your requirements and goals.
Sustainable and Safe AI Clusters to Increase Efficiencies and Protect Against Threats
The HPE ProLiant Compute XD685 offers a DLC (direct liquid cooling) option that helps you meet escalating power requirements and advance sustainability goals, leveraging HPE’s decades of expertise and market leadership in deploying the world’s largest liquid-cooled IT environments.
HPE iLO offers seamless server management with built-in security innovation to protect against escalating security threats.
Silicon root of trust from HPE, our fundamental approach to security, provides a Zero- Trust architecture at the silicon level and helps enable completely server essential firmware that is free from malware or compromised code.
HPE Financial Services allows you to turn aging infrastructure into capital in a secure, globally consistent, and environmentally responsible way, without disrupting your AI production.
Proven Expertise Worldwide to Support your Global Footprint and Growth Ambitions
The HPE ProLiant Compute XD685 is backed by a comprehensive portfolio of HPE Services to build, test, and deploy your solution with ongoing support, entirely available as stand-alone or consumption-based services, to accelerate the deployment of your environment.
The robust and secure global supply chain from HPE and our proven ability to deploy massive custom AI clusters anywhere in the world with superior serviceability enable you to establish your presence globally with confidence.
Advisory and Professional Services, offering best-in-class methodologies and tools to enhance performance and increase return on investment (ROI), help you navigate the challenges of a large AI environment.
HPE Tech Care Service combines deep technical expertise with a leading-edge digital experience to provide personalized, prompt, and reliable product support, so you can get the utmost of your global IT investments.
- AMD is a trademark of Advanced Micro Devices, Inc. ENERGY STAR is a registered mark owned by the U.S. government. NVIDIA is a trademark and/or registered trademark of NVIDIA Corporation in the U.S. and other countries. All third-party marks are property of their respective owners.
Liens connexes
HPE Services
- HPE Financial Services
- Consulting Services
- HPE Managed Services
- HPE Complete Care Service
- HPE Tech Care Service
- HPE Lifecycle Services
- HPE Education Services
HPE ProLiant Compute XD685 at hpe.com