Features

The complete solution for all common deep learning frameworks

Die Deep Learning Box ist für alle gängigen Deep Learning Frameworks optimiert und bereits fertig eingerichtet für TensorFlow, Caffe2, Theano oder Torch
Deploy. Connect. Train.

 

Tensorflow Deep LearningCaffe Deep LearningTheano Deep LearningTorch Deep Learning NVIDIA digits NVIDIA Preferred Solution Provider CADnetwork NVIDIA CUDNN Library

Let's Get Ready to Train

Start training neural networks without having to worry about hardware, installation and setup issues. The Deep Learning Appliance is already set up for Tensorflow, Caffe2 or a framework of your choice.

 

 

Deep Learning Machine Learning Tensorflow

Multi-GPU support

The Deep Learning Appliance supports up to eight high-performance NVIDIA GPUs, delivering impressive AI results in the shortest possible time.

 

 

Deep Learning Box Rack 8GPU

 

Features
up to eight
NVIDIA Tesla V100 GPUs
Features
Dual processor
Intel Xeon SP
Features
up to 1.5TB DDR4
memory
Features
PCIe NVMe SSD
24x HDD / SSD

With dual Intel® Xeon® SP processors

Dual Intel Xeon

 

Two Intel Xeon Scalable processors with up to 56 cores drive the Deep Learning Appliance and deliver maximum performance for demanding calculations. The integrated DDR4 memory controller enables a memory expansion of up to 1.5TB. In total, the two Xeon SP processors provide 80 PCI Express lanes, so that there is no bottleneck when using eight GPUs.

Technical Data

Technical Data

  • Processor
  • GPU Support
  • Memory max.
  • Storage
  • Network
  • Connector ports
  • Expansion slots
  • Remote management
  • Power supply
  • Software
  • Dimensions
  • Warranty

Deep Learning Appliance 4U

  • 2x Intel® Xeon™ SP Skylake processors 1st and 2nd generation
    up to 56 Cores
  • 8x NVIDIA Tesla V100 32GB
    SXM2 module with NVLink technology
  • up to 6TB 3DS DDR4-RAM
    2933 MHz ECC registered
    24 slots
  • 2x NVMe M.2 SSD 2260,2280,22110
    16x 2,5" SSD/HDD SATA 6GB/s
    Optional 8x 2.5" NVMe
  • 2x 10 Gigabit LAN Intel
    1x management LAN
  • 2x USB 3.0; 3x LAN; 1x D-Sub (VGA)
  • 4x PCI-E 3.0 x16 LP with GPUDirect RDMA slots
    2x PCI-E 3.0 x16 LP slots
  • IPMI 2.0 with KVM and Virtual Media via LAN
  • 2200W power supply redundant (2+2) 80+ Titanium
    100/240V, 50/60Hz
  • Ubuntu 18.04 LTS
    NVIDIA GPU driver, CUDA, CuDNN
    Ready-to-Brain Docker Container with
    Tensorflow, Caffe2 & Torch
  • 4U chassis
    447 x 178 x 805 mm WHD
  • 36 months, optionally expandable

Buy

Here you can find our standard configurations:

 

 

Downloads