Density Optimized System Solution Providing Leadership Performance for HPC and AI

Key Features

  • Performance Optimized for Compute Density

  • Intel® Xeon® Platinum 9200 Processors Advanced Performance

  • Density Optimized 2U Rack Server with Air-Cooled and Liquid-Cooled Options

  • Built with Intel Quality, Reliability, and Performance

  • Accelerating Time to Market with Innovative Data Center Solutions

author-image

投稿者:

Performance Optimized for Compute Density

The Intel® Server System S9200WK product family is a purpose built, performance-optimized data center block ideal for use in high performance computing (HPC) and Artificial Intelligence applications. Designed for the Intel® Xeon® Platinum 9200 series processors, with up to 24 DDR4 DIMM slots per compute module, the S9200WK family maximizes processor and memory bandwidth to provide leadership performance for the most demanding compute use requirements.

Intel® Xeon® Platinum 9200 Processors Advanced Performance

  • Leadership CPU performance per socket with Intel’s highest core count, Intel® Xeon® Platinum 9200 processors
  • Double the memory bandwidth for memory-intensive workloads with 12 memory channels per CPU, 24 memory channels per compute module
  • New Intel® Deep Learning Boost (Intel® DL Boost) Instructions for data analytics greatly accelerates inference performance
  • Multi-Chip packaging optimized for density and performance

Density Optimized 2U Rack Server with Air-Cooled and Liquid-Cooled Options

  • Up to 4 compute modules per 2U chassis which can support multiple compute module types in a single chassis
  • 2 CPU compute module design with advanced cooling technology for high flow rate air or liquid cooled for CPUs, VRs, DIMMs, and memory VRs for high heat capture ratio
  • Up to 350W processor TDP for high performance workloads in a 2U air cooled chassis, up to 400W processor TDP with liquid-cooled versions
  • Up to 2 x16 PCIe* slots in 1U compute modules, up to 4 x16 PCIe* slots in 2U compute modules for network expansion options
  • Support for 2x M.2 SATA/NVMe* storage devices per 1U compute module, up to 2x M.2 SATA/NVMe* and 2x U.2 NVMe storage devices per 2U compute module
  • Hot-swappable compute modules, storage1, fans, and power supplies

Built with Intel Quality, Reliability, and Performance
Intel® Server Products are backed by Intel’s design excellence and manufacturing expertise to deliver processing power with high levels of flexibility, manageability, and reliability. Product and design quality is paired with 3-year standard warranties and robust technical and incident resolution support to ensure customer satisfaction.

Accelerating Time to Market with Innovative Data Center Solutions
The S9200WK product family is the highest performance member of Intel® Data Center Systems (Intel® DCS). These fully validated, unbranded server systems include Intel’s latest data center technology― already optimized to work better together―allowing partners to accelerate time to market with reliable data center solutions. The process of configuring and validating the hardware components of solutions that are tuned to meet specific customer requirements is a complex and resource intensive process. Intel® Data Center Systems (Intel® DCS) based on the S9200WK product family reduce this complexity, making it easier to build innovative server solutions that can support the demands of today’s data center workloads.

Highly Integrated, High Density Compute Solution
The Intel® Server System S9200WK Data Center Systems (Intel® DCS) can be configured to support a wide range of memory, storage, and I/O options. Solutions are configured using Intel® Xeon® Platinum 9200 processors, Intel® Server System S9200WK Compute Modules, and Intel® Server Chassis FC2000. The Intel® Server Chassis FC2000 family allows flexible configuration of different functionality compute modules in a single chassis further extending the benefits of Intel® Server System S9200WK Data Center Systems (Intel® DCS). This allows a single data center systems to address both compute and service node functionality simultaneously, reducing cost and increasing functionality.