4U

GPU
NF5468A5 (4U 8GPU AI Server with AMD EPYC™)
  • High-performance balanced design of 4U 8GPU to provide leading performance in the ultimate space
  • N+N redundant PSUs to allow 8 x 350 W GPUs to run at full speed
  • 2 AMD EPYC™ processors, 128 cores, and 3 xGMI links
  • Various global AI accelerator card partners like NVIDIA, AMD, Intel, Cambricon, and Enflame
GPU
NF5468M6 (4U 4-16 GPU AI Server)
  • Supports up to 20x PCIe GPUs and accelerators in 4U
  • Supports the latest NVIDIA A40, A100
  • Supports AMD Instinct MI100 GPUs powered by the latest Matrix Core technology
  • 3 flexible GPU configurations
High-Density
i48M6 (4U 8-Node Server)
  • high density performance with 8x dual-socket computing and storage nodes in 4U
  • powered by 3rd Generation Intel Xeon Scalable processors
  • up to 48x I/O expansions for data-intensive applications
Storage
NF5466M6 (4U 2-Socket Storage)
  • 2x 3rd Generation Intel Xeon Scalable processors
  • enhanced I/O scalability with 13x PCIe expansions
  • extreme storage capacity of up to 828TB with 46x 3.5-inch drives
Compute
NF8480M6 (4U 4-Socket)
  • 4x 3rd Generation Intel Xeon Scalable processors
  • up to 18TB of memory with Intel Optane Persistent Memory 200
  • dynamic scalability with up to 18 PCIe Gen3 slots
Storage
NF5476M5 (4U 2-Socket Server)
  • 1PB storage in 4U 800mm depth system
  • supports 60* 3.5” SATA HDDs, 4* U.2 NVMe/SATA SSDs, 2* M.2 SSDs
  • 12 DIMMs and 6 memory channels per processor
GPU
NF5488M5-D (4U 8GPU Server with A100)
  • 8x fully interconnected NVLink 3.0 NVIDIA A100 GPUs, 2x second-generation Intel® Xeon® scalable processors, HBM2e memory in 4U
  • The 4U 8GPU design helps AI users to efficiently complete the construction of AI infrastructure and development environment
  • NVIDIA CUDA Delta system and Intel platform development ecosystems minimize business migration costs, enabling fast deployment.
GPU
NF5488A5 (4U 8GPU Server with A100)
  • 2 second generation AMD EPYC™ processors
  • 8 NVIDIA A100 GPUs with NVLink 3.0 interconnect
  • 5 petaFLOPS AI performance
  • Highest single server performance in MLPerf v0.7 benchmark
High-Density
NF8480M5 (4U 4-Socket)
  • Extreme performance: CPUs of up to 28 cores and 56 threads, and max 205W TDP
  • 16GB 24* 2666 MT/ s NVDIMM memory
  • 12* hot swap NVMe SSD
  • 2* dual-width GPUs each delivering 20 trillion FP16 performance
  • Flexible expansion: up to 48* 2.5”/3.5” hard drives, 12* U.2 NVMe SSD, 2 built-in M.2 SSDs
High-Density
i48 (4U 4/8-Node)
  • 2 configurations: 4-node and 8-node
  • Hybrid deployment of computational and storage node types
  • Stable and reliable with fully redundant design
  • Scale on demand with 4 PCIe expansion cards and 8 PCIe 3.0 x16 expansion slots
Compute
TS860M5 (4U 8-Socket)
TS860M5 4U 8-Socket SAP HANA Mission Critical Server
  • Extreme computing power: 4U contains 8x Intel Xeon Skylake Gold/Platinum series processors
  • Supports 2.5-inch NVMe SSD
  • 12T high performance memory, suitable for HPC large memory applications
  • Flexible and reliable: fully modular redundant design for power-off protection
  • Intelligent system monitoring
GPU
NF5468M5 (4U 16GPU AI Inference Server)
NF5468M5 4U 16GPU AI Inference Server
  • High Performance for Al Model Training
  • High Reliability for More Al Application Scenarios
  • High Efficiency for Al Edge Inference
Storage
NF5486M5 (4U 106-Bay)
  • Ultra-high density 4U storage server supporting 106 3.5″ HDDs and 1.25PB storage capacity
  • Energy saving on-demand power controls reduce system power consumption
  • Stable and reliable with enhanced structural design
  • Ideal for video and data storage