Because of the COVID-19 pandemic, graphics giant NVIDIA said its first GPU based on the Ampere architecture, A100, is in full production and shipping to customers worldwide. According to an announcement made at the GPU Technology Conference (GTC) the A100 is also built for data analytics, scientific computing, and cloud graphics.
At the heart of the talk by NVIDIA Founder and CEO Jensen Huang was a vision of how data centers, the engine rooms of the new global knowledge economy are evolving, and how NVIDIA and Mellanox, acquired in an agreement that closed last month, are driving those changes together.
“The data center is the new computing machine,” Huang said, adding that NVIDIA is driving silicon-based efficiency improvements, how CPUs and GPUs interact, the entire software stack, and eventually all data centers.
NVIDIA A100 provides NVIDIA’s eight generations of GPUs with the greatest generational performance leap, Huang said, adding it’s in full production and shipping to customers around the world. These are implemented by eighteen of the world’s leading service providers and network builders, including Alibaba Cloud, Amazon Web Services, Baidu Cloud, Cisco, Dell Technologies, Google Cloud, Hewlett Packard Enterprise, Microsoft Azure, and Oracle. The A100, and the NVIDIA Ampere architecture on which it is based, are boosting efficiency by up to 20 times over its predecessors, Huang said. The coronavirus pandemic revised original plans for the keynote to be presented live at NVIDIA’s GPU Technology Conference in San Jose late March.
Huang explained in his address reported from his California home’s kitchen, that Nvidia is collaborating with researchers and scientists to use GPUs and AI computing to handle, reduce, contain, and monitor the pandemic. “Researchers and scientists applying Nvidia accelerated computing to save lives is the perfect example of our company’s purpose — we build computers to solve problems normal computers cannot,” Huang said.
Huang also announced that NVIDIA GPUs will power major software applications to accelerate three critical uses: large data management, recommendation systems, and real-time, conversational AI building.
These new tools come as the productivity of machine learning has pushed businesses to gather more and more data. “That positive feedback is causing us to experience exponential growth in the amount of data that is collected,” Huang said.
Huang declared support for NVIDIA GPU acceleration on Spark 3.0 to help organizations of all kinds keep up, describing the Big Data Analytics engine as “one of the world’s most significant applications in today.”