High performance
Supports 1st or 2nd generation Intel® Xeon® Scalable series processors.
high capacity
12 memory channels 24x DIMM support. 3TB memory support.
Strong expansion
11 PCIe expansion slots. It has high-power GPU support and 2 x M.2 ports.
2.5 petaFLOPS AI
2,496 TFLOPS Tensor Performance
311.88 TFLOPS of FP16 Performance
77.96 TFLOPS of FP32 Performance
38.98 TFLOPS of FP64 Performance
Deep Learning SDK, Frameworks, AI Libraries, Applications, Operating System and Optimised Drivers
System Specifications
Up to 64-Core AMD or INTEL processor
4 x NVIDIA A100 80GB or 40GB HBM2
27,648 total NVIDIA CUDA® Cores
Server-grade Motherboard
Up to 2TB 3200Mhz DDR4 Memory
Up to 2TB PCI-E M.2 NVME disk for OS
7,68 TB PCI-E U.2 NVME disk for DATA
Up to 2 x 16TB SATA 6Gb HDDs in RAID Optional PCI-E NVME disk for RAID CACHE
Up to 2000W Power Supply
Linux / Ubuntu - Red Hat Enterprise- Centos
NVIDIA-qualified driver
NVIDIA® CUDA® Toolkit
NVIDIA® cuDNN™
Caffe, Theano, Torch, BIDMach, Tensorflow Anaconda Data Science Platform
Management Controller Port 1 Gbit
Dual Port 10G BASE-T Ethernet
Nominal Operating Temperature 5–30 ºC