System Thumbnail

Recommended Systems for NVIDIA DIGITS/Caffe

Share:

 

Compact
GPU Workstation

Purchase

Compact, quiet and cool -- GPU accelerated workstation for Deep Learning workloads in a package small enough to be taken as a carry-on on an airplane.

"DIGITS"
GPU Workstation

Purchase

Full "DIGITS" configuration using our highest quality motherboard with up to 4 GPU's at full X16. Up to 512GB of RAM and a wide range of storage options.

 


Which system is right for you?

Compact GPU Workstation

  • Well utilized for Deep Learning workloads with DIGITS and Caffe
  • Compact, but not "too" small
  • Efficient and quiet cooling under heavy load
  • Optional Airline compliant carrying case

The small size of this workstation is not necessarily it's most important feature

  • Tested with million image DNN classification jobs
  • 1 or 2 NVIDIA Pascal GPU's for compute
  • Intel Xeon or core i7 CPU (up to 22-core)
  • Up to 256GB mem
  • Recommended hardware configs: ( other options available )
    • 1 or 2 Titan X Pascal or GTX 1070
    • Intel E5 1630v4 4-core 
    • 64 or 128GB mem
    • 1 or 2TB system SSD 

We have tested this system using the NVIDIA DIGITS software stack with Caffe and found it to give very good performance under heavy load.

"This system always makes me smile when I have it under load"

-- Dr D.B. Kinghorn


DIGITS GPU Workstation

  • Updated version of our  "DIGITS" workstation
  • Best workstation configuration for GPU focused workloads like DNN with Caffe  
  • Can train GoogLeNet on a 1 million ImageNet subset for 30 epocs in 8hr
  • Highest quality motherboard
    • 4 Full X16, PLX switched, metal reinforced PCIe slots
    • converged 10G network
  • Optimal chassis with excellent cooling and quiet operation

Our main platform for GPU accelerated Machine Learning applications 

  • Recommended hardware configs ( other options available )
    • 2 or 4 Titan X Pascal GPU's
    • Intel Xeon E5 1660v4 8-core or core-i7 6950X 10-core
    • 128 or 256GB memory
    • 1TB system SSD, 2TB data SSD, 4GB storage HD

We have trained deep neural networks with complex models and large data sets, utilizing 4 TitanX GPU's with this system.

Note: Some workloads may not scale well on multiple GPU's You might consider using 2 GPU's to start with unless you are confident that your particular usage and job characteristics will scale to 4 cards. We can pre-wire for 4 cards for easy expansion.  

 

Looking for more detail on our testing and why we chose this hardware? Have a look at the testing blog post!


NVIDIA DIGITS, Caffe, and Machine Learning Articles:

If you are configuring a system for workloads like training DNN's with NVIDIA DIGITS and Caffe, we have a number of articles that you may be interested in:

NVIDIA DIGITS with Caffe - Performance on Pascal GPUs
Main DIGITS/Caffe hardware testing article

NVIDIA DIGITS Install
Simple DIGITS install on Ubuntu 14.04

Install Ubuntu 16.04 or 14.04 and CUDA 8 and 7.5 for NVIDIA Pascal GPU
More general multi-version CUDA setup on Ubuntu 14.04 and 16.04 


Why Choose Puget Systems?


Highly Reliable Product Line

We do not add a part to our product line unless we feel we can stand behind it. You can feel confident that any selection you make on our website is a quality product.

Fast Build Times

By keeping inventory of our most popular parts, and maintaining a short supply line to parts we need, we are able to offer an industry leading ship time of 7-10 business days on nearly all our system orders.

We're Here, Give Us a Call!

We make sure our representatives are as accessible as possible, by phone and email. At Puget Systems, you can actually talk to a real person!

Lifetime Support/Labor Warranty

Even when your parts warranty expires, we continue to answer your questions and even fix your computer with no labor costs.

Click here for even more reasons!