NVIDIA DGX-1 supercomputer is aimed at deep learning and AI

In addition to the Tesla P100 GPU, NVIDIA also announced DGX-1 Deep Learning System "supercomputer in a box".  The system has been designed to tackle the complex worlds of deep learning and AI, areas of research which requires massive amount of computing power. 

The company says that the DGX-1 is equivalent of 250 servers relying only on Intel processors, and offers 170 Teraflops of half-precision performance.

The DGX-1 uses eight Tesla P100 GPUs each having 16GB of HBM2 RAM to work with, and packs 2x Intel Xeon E5-2698 v3 (16 core, Haswell-EP) processor , 512GB DDR4 RAM, 4 x Samsung's 1.92TB SSD RAID, Dual 10GbE, 4 IB EDR networking, and requires a maximum of 3,200W.

NVIDIA is positioning this machine for serious research purposes. Massachusetts General Hospital is the first customers using the DGX-1 for its clinical data center. The AI will be used in the hospital to learn about and diagnose heart disease using radiology and pathology data and will use its archive of 10 billion medical images to create a deep learning neural network.

The NVIDIA DGX-1 is priced at $129,000 and is now up for pre-order. It will be available in the U.S in June, followed by other regions in Q3 2016.

Via