Delta Sprut XL
GPGPU platform based on the latest Intel® Xeon® 6 processors with the ability to install up to 20 NVIDIA H200 accelerators or up to 25 NVIDIA RTX PRO 6000 SE accelerators
Description
Delta Sprut XL is a Russian platform created for deploying inference and training artificial intelligence models (AI), as well as for solving highly complex tasks in the fields of high-performance computing and scientific modeling (HPC).
Specifications
- Capacity for up to 20 NVIDIA H200 or up to 25 NVIDIA RTX PRO 6000 SE
- Scalability thanks to the platform's modularity from 4 to 25 accelerators
- Significant performance increase by combining every 4 accelerators through NVLink
- Up to 172 productive Granite Rapids P-cores or 288 energy-efficient Sierra Forest E-cores
- Up to 8 TB of DDR5 RAM up to 6400 MT/s, support for MCR DIMM memory with frequencies up to 8000 MT/s
- Extensive monitoring and control functionality in the software Delta BMC
Up to 20 accelerators
(NVIDIA H200)
OR
Up to 25 accelerators
(NVIDIA RTX PRO 6000 SE)
Appearance
Applications
Artificial intelligence and machine learning
With record GPU density per server, Delta Sprut XL is a unique device for building Artificial Intelligence systems.
HPC
Ensuring maximum performance of HPC clusters with Delta Sprut XL, opening up new possibilities for parallel computing
Scientific modeling
Delta Sprut XL allows you to work with the most resource-intensive applications thanks to its high level of computing performance and large amount of high-speed memory
Advantages
Scalability
Possibility of phased expansion of the complex to 25 GPUs per server
NVLink support
The platform supports the combination of every 4 NVIDIA H200 accelerators through NVLink
Unsurpassed level of performance
The platform is built on the latest Intel® Xeon® 6 processors and modern NVIDIA H200 or NVIDIA RTX PRO 6000 SE accelerators
Photo
