NVIDIA® DGX Systems
Inspired by the demands of deep learning and analytics, NVIDIA® DGX™ Systems are the essential instruments for AI research built on the new NVIDIA® Ampere™ Architecture in the DGX A100 and Volta™ GPU platform in the DGX Station.
The third generation DGX provides universal infrastructure for AI, with unmatched capabilities for training, inference and big data analytics.
- 5 petaFLOPS of performance in 6RU
- 8 GPU’s with 320GB combined memory
- 6x NVLink Switches with 4.8TB/s bandwidth
- 9 x Mellanox ConnectX-6 200Gb/s NICs
The Personal Supercomputer for Leading-Edge AI Development.
- 3x the performance for Deep Learning training
- 100x in speed-up on large data set analysis, compared with a 20 node Spark server cluster
- 5x increase in I/O performance over PCIe-connected GPU’s with NVIDIA® NVLink technology