If you are impressed with the recently released Nvidia Quadro GV100, get ready to embrace the all-new DGX-2 which packs a huge 16 x Tesla V100s with 32GB of HBM2, each for a total of 512GB of HBM2.
The Power House
Undoubtedly, the 350 lbs DGX-2 is a powerhouse for professionals and content creators. The unit consumes 10kW of power to produce 2 PFLOPs of compute performance, Nvidia has connected all of its Tesla V100s with 32GB of HBM2 through its new NVSwitch technology.
With the latest NVSwitch technology, achieving the 14.4TB/sec of bandwidth is now possible. The best possible clients of Nvidia for the DGX-2 can be super-powered machines for AI or deep learning scenarios.
Patrick Moorhead, President & Principal Analyst, Moor Insights & Strategy of Nvidia spoke in an interview:
NVIDIA combined Volta architecture with the new, state of the art platform improvements and created DGX-2. The biggest feature in DGX-2 is the new NVSwitch, which enables memory sharing of 16 GPUs. This improves performance as well as reducing latency as the GPUs don’t need to go to main memory as much.