Custom configured, quickly delivered!
Your shopping cart is empty
Buy Online or Call 800-237-0402
Search
Designed for AI applications, the HPE ProLiant Compute XD685 Server features 2x chassis, a 5U platform with direct liquid cooling or a 6U chassis with air cooling. For easy expansion, the 2-socket platform supports a limited selection of 5th generation AMD EPYC processors with up to 96 physical cores and 192 virtual threads. At full capacity it can support up to 6TB of memory across 24 DDR5 memory module slots. It supports up to 6x NVIDIA SXM form factor GPUs.
Offering a flexible scale-out platform for Large Language Models (LLM), complex AI models, and Deep Learning (DL) workloads, the HPE ProLiant Compute XD685 Server can be outfitted with either up to 6x NVIDIA H200 SXM Tensor Core or NVIDIA B200 featuring Blackwell architecture, depending on air or direct liquid cooling (DLC). Additional GPU options include AMD Instinct MI300X for air-cooled platforms and the Intel Gaudi 3 GPUs. Processors are limited to a thermal design power (TDP) range of up to 400W and can offer up to 384MB of L3 cache. Featuring 12-channel memory architecture and 12x memory module slots per socket, at full capacity it can support up to 6TB of Registered DIMMs (RDIMM). With 5th Gen AMD EPYC CPUs memory can run at up to 6400MT/s. There are a few different storage options, all supporting EDSFF storage devices.
Storage Options*
With 8x half-height, half-length (HHHL) PCIe 5.0 x16 slots and 4x full-height, half-length (FHHL) slots, this platform can be outfitted with a number of PCIe-based Network Interface Controllers (NICs) and storage controllers. To boot the system, there is also a dedicated slot for an NVMe hot plug boot optimized storage device that includes 2x NVMe M.2 storage devices that can be combined in a hardware RAID 1 for resiliency. The boot optimized storage device can also be mounted in the front of the chassis. 6x 3000W hot-plug power supply units (PSU) at the back of the system deliver 1+N redundancy. For the DLC option, with the addition of the optional power shelf, it can support up to 9x PSUs.
For management of the system, HPE offers integrated Lights Out 6.0 (iLO6) which is embedded in the system. Also supported is the HPE Performance Cluster Manager (HPCM). HPCM is a software utility designed to manage high-performance computing clusters and supercomputers. It enables provisioning, managing, and monitoring of essential systems and can scale up to 100,000 nodes.
If you know what you want but can't find the exact configuration you're looking for, have one of our knowledgeable sales staff contact you. Give us a list of the components you would like to incorporate into the system, and the quantities, if more than one. We will get back to you immediately with an official quote.