DGX Systems

Developed to meet the demands of AI and analytics, NVIDIA® DGX™ Systems are built on the revolutionary NVIDIA Volta™ GPU platform. Combined with innovative GPU-optimized software and simplified management tools, these fully-integrated solutions deliver groundbreaking performance and results. NVIDIA DGX Systems are designed to give data scientists the most powerful tools for AI exploration—from your desk to the data center to the cloud.

A variety of businesses are choosing deep learning to process geometrically increasing data. With this solution, they can utilize the data to develop product, service, and process to keep their competency

NVIDIA DGX Station comes preloaded with the same software stack found in all DGX solutions. This innovative, integrated software stack includes access to popular deep learning frameworks, each optimized by NVIDIA deep learning experts and updated monthly. It also comes with NVIDIA DIGITS deep learning training application, third-party accelerated solutions, the NVIDIA deep learning SDKs such as cuDNN and cuBLAS, CUDA toolkit, fast multi-GPU collectives (called NCCL), and NVIDIA drivers.

※ NVIDIA DGX-1, NVIDIA DGX Station Common

NVIDIA GPUs 8X Tesla V100 8X Tesla P100 4X Tesla V100
TFLOPS(GPU FP16) 960 170 480
CPU Dual 20-Core Intel Xeon E5-2698 v4 2.2 GHz Intel Xeon E5-2698 v4 2.2GHz (20-Core)
NVIDIA CUDA® Core 40,960 28,672 20,480
NVIDIA Tensor Core 5,120 N/A 2,560
Max. Power Cost 3,200 W 1,500 W
System Memory 512 GB 2,133 MHz DDR4 LRDIMM 256GB LRDIMM DDR4
Storage 4X 1.92 TB SSD RAID 0 데이터 : 3X 1.92TB SSD RAID 0
OS : 1X 1.92TB SSD
Network Dual 10GbE, Max. 4 IB EDR 이중 10Gb LAN
GPU Memory System Up to 128GB System Up to 64GB
Software Ubuntu Linux Host OS Ubuntu Desktop Linux OS
Turbo Ability Comparison 96x Faster Training Possible NVIDIA DGX-1
47x Faster Training Included