NVIDIA® DGX Station™ – Your deep learning desktop

vc-single-img-ws

NVIDIA workstation GPU for data science teams

NVIDIA DGX deep learning PC supports accelerated analytics, deep learning training, and inferencing, all within a single versatile platform, that packs the computing capacity 400 CPU’s into a desk-friendly package that consumes less than 1/20th the power. It’s the only personal AI supercomputer that integrates four Volta powered V100 GPUs, with innovations including NVIDIA NVLink and Tensor Core architecture.

DGX-Station-Performance-IMG

Breakthrough performance and precision - powered by Volta

DGX-Station-Accelerate-IMG

Start experimenting in hours, not weeks, powered by DGX Stack

DGX-Station-Form-Factor-IMG

Revolutionary form factor - designed for your office, whisper-quiet

DGX-Station-AI-Access-IMG

Access to AI expertise and enterprise grade support

Special discount for education and AI startups*

Key Features

Groundbreaking AI in your office

1. GPUs

  • 4X NVIDIA V100 32 GB GPUs
  • 500 TFLOPS (Mixed Precision)
  • 20,480 Total NVIDIA CUDA® Cores
  • 2,560 Tensor Cores

2. SYSTEM MEMORY

  • 256 GB RDIMM DDR4

3. GPU INTERCONNECT

  • NVIDIA NVLink, Fully Connected 4-Way

4. STORAGE

  • Data: 3 x 1.92 TB SSD RAID 0
  • OS: 1 x 1.92 TB SSD

5. CPU

  • Intel Xeon E5-2698 v4
  • 2.2 GHz 20-Core

6. NETWORKING

  • DUAL 10 GBASE-T LAN (RJ45)

7. DISPLAYS

  • 3X DisplayPort,
  • 4K Resolution

8. COOLING

  • Water-Cooled

9. POWER

  • 1500 W

Integrated software stack for enhanced performance

Linear Scalable Performance

DGX Station V100 Linear Scalable Performance 1

NVIDIA DGX Station with 4x V100 32 GB. ResNet-50 V1.5 Training, Volta with Tensor Cores (mixed precision), 20.3_py3 optimized containers from NGC.

Score: Images per second

Translation and Speech Performance

DGX Station V100 Translation and Speech Performance

NVIDIA DGX Station with 4x V100 32 GB. Volta with Tensor Cores (mixed precision). PyTorch 20.03_py3 optimized containers from NGC.

Score: Total Tokens per Second (GNMT v2), Output Samples per Second (waveGlow)

DGX Station: DL Optimizations

DGX Station V100 DL Optimizations 3

NVIDIA DGX Station with 4x V100 32 GB. ResNet-50 Training, Volta with Tensor Cores (mixed precision), 17.09, 18.09py3, 18.11py3 DGX optimized containers from NGC.

Score: Images per Second

Fully Integrated Software for Instant Productivity

NVIDIA DGX Station software stack

Advantages:

  • Instant productivity with NVIDIA optimized deep learning frameworks, pre-trained models, and scripts: Pytorch, TensorFlow, MXNet, and many others

  • Performance optimized across the entire stack

  • Faster Time-to-Insight with pre-built, tested, and ready to run framework containers

  • Flexibility to use different versions of libraries like libc, cuDNN, CYDA, python in each framework container

Unparalleled deep learning training performance

Delivers 47X faster training

DGX Station performance projected based on DGX (with V100) Workload: ResNet50, 90 epochs to solution | CPU Server: Dual Xeon E5-2699 v4, 2.6 GHz. Projections subject to chance.

Get started within 1 hour with NVIDIA DGX Station

Maximize productivity

Deploy quickly and simply

Plug-and-play setup that takes teams from power-on to data science and AI research

NVIDIA GPU cloud and support

Access to NVIDIA’s vast deep learning knowledge, expertise, and the latest software updates

NVIDIA® DGX Station™

Accelerate your data science and AI research today

NVIDIA DGX Station™ Whitepaper

NVIDIA DGX Station gives computational scientists, data scientists and AI researchers the fastest start in deep learning, HPC, and data science from the convenience of their desks. Understand how the DGX Station will allow you to experiment at your desk and extend same deep learning software across DGX systems and the cloud.

From your desk to the data center to the cloud

Purpose-built from the ground up, NVIDIA DGX Systems are designed to give data scientists the most powerful tools for AI exploration.

NVIDIA DGX Software Stack

NVIDIA DGX Software Stack

DOWNLOAD PDF
Greater Productivity & ROI

Greater Productivity & ROI

DOWNLOAD STUDY
Access to AI expertise

Access to AI expertise

DOWNLOAD PDF

Deploying an AI platform can be complex and time-consuming, including the effort required to build a platform and engineer a software stack that’s optimized for maximum performance.  Together with NVIDIA, AMAX AI provides the total solutions that enables you to get results faster.

We'll design a system that matches your specifications.

Choose from workload-optimized, pre-configured system or customize a workstation for your unique needs.

*Education & AI startup discount is subject to approval, please call for more details.

Enter your keyword