NVIDIA releases DGX Station A100, a compact AI supercomputer for researchers and enterprises.

NVIDIA announced the DGX Station A100 on November 16, 2020, at SC20, positioning it as "the world's only petascale workgroup server." The second-generation system accelerates demanding machine learning and data science workloads for teams working in corporate offices, research facilities, labs or home offices. Delivering 2.5 petaflops of AI performance, the DGX Station A100 features four of the latest NVIDIA A100 Tensor Core GPUs fully interconnected with NVIDIA NVLink, providing up to 320GB of GPU memory to speed breakthroughs in enterprise data science and AI.

The system is the only workgroup server that supports NVIDIA's multi-instance GPU (MIG) technology, which allows a single DGX Station A100 to provide up to 28 separate GPU instances to run parallel jobs and support multiple users without impacting system performance. Charlie Boyle, vice president and general manager of DGX systems at NVIDIA, stated: "DGX Station A100 brings AI out of the data center with a server-class system that can plug in anywhere. Teams of data science and AI researchers can accelerate their work using the same software stack as NVIDIA DGX A100 systems, enabling them to easily scale from development to deployment." While requiring no data-center-grade power or cooling, it features the same remote management capabilities as NVIDIA DGX A100 data center systems.

Organizations worldwide have adopted DGX Station for AI innovation across industries including education, financial services, government, healthcare and retail. BMW Group Production is using it to develop and deploy AI models that improve operations. DFKI, the German Research Center for Artificial Intelligence, is building models that tackle critical challenges including computer vision systems that help emergency services respond rapidly to natural disasters. Lockheed Martin is developing AI models that use sensor data and service logs to predict maintenance needs and improve manufacturing uptime and worker safety. Pacific Northwest National Laboratory is conducting federally funded research in support of national security with a focus on energy resiliency. NTT Docomo, Japan's leading mobile operator with over 79 million subscribers, uses DGX Station to develop AI-driven services such as its image recognition solution.

For complex conversational AI models like BERT Large inference, DGX Station A100 is more than 4x faster than the previous generation, and delivers nearly a 3x performance boost for BERT Large AI training. The system is available with four 80GB or 40GB NVIDIA A100 Tensor Core GPUs, providing options for data science and AI research teams to select according to their unique workloads and budgets. For advanced data center workloads, NVIDIA DGX A100 systems will be available with new 80GB GPUs, doubling GPU memory capacity to 640GB per system. The Cambridge-1 supercomputer being installed in the U.K. for healthcare research and the University of Florida HiPerGator AI supercomputer are among the first installations of NVIDIA DGX SuperPOD systems with DGX A100 640GB. Both DGX Station A100 and DGX A100 640GB systems became available that quarter through NVIDIA Partner Network resellers worldwide, with an upgrade option available for NVIDIA DGX A100 320GB customers.