DEEP Gadget provides a silent, reliable, and powerful deep learning platform powered by up to seven water-cooled GPUs and ready-to-use software stack.
Noise is no longer an essential part of GPU computing. DEEP Gadget runs at 55 dB* even at full load. This is much quieter than other air-cooled GPU servers (more than 70 dB) as well as normal conversation (60 dB). DEEP Gadget can even be placed in your office.
* Standalone model
Seven GPUs in a single DEEP Gadget system provides up to 100 TFLOPS of single-precision floating-point performance, and 700 TFLOPS of deep learning performance enabled by Volta Tensor Cores. A closed-circuit water cooling system keeps high-density GPUs at a low temperature. It completely prevents GPU throttling, which significantly degrades the performance of GPUs.
DEEP Gadget is delivered with pre-installed software stacks including an operating system and deep learning frameworks, such as TensorFlow, PyTorch, and Keras.
DEEP Gadget uses water cooling to ensure enterprise-level lifetime and reliability in cost-effective gaming GPU.
All prices are subject to change without any prior notice. Tax and shipping cost are not included.
|Form Factor||Standalone (W*H*D 405*611*626 mm) or 4U Rackmount|
|# of CPU Sockets||1 or 2|
|CPU||Intel Xeon Scalable 2nd Generation (Cascade Lake)|
|Maximum # of GPUs||7|
One of the following:
|Primary Storage||250 GB, 500 GB, 1 TB, or 2 TB M.2 NVMe SSD|
One of the following for a standalone model:
|Motherboard||ASUS WS C621E SAGE or X299 SAGE|
|Power Supply||2x 1300 Watts (80 PLUS Platinum)|
|Operating System||Ubuntu 18.04 LTS|
|Libraries||CUDA Toolkit, cuBLAS, cuDNN, NCCL, NCCL 2|
|Deep Learning Frameworks||TensorFlow, PyTorch / Caffe2, Keras, Caffe, Theano, MXNet, Cognitive Toolkit|
You can customize a DEEP Gadget system and make an order online.Configure & Order