NVIDIA CEO Jensen Huang delivered a belated keynote address for GPU Technology Conference (GTC) 2020 yesterday (May 14). The delay was due to the COVID-19 pandemic which caused the event, which was to be held in March, to go digital.
But, the wait was perhaps for the better because the NVIDIA chief announced a a host of products based on its new Ampere architecture, which sports a staggering 54 billion plus transistors to become the world’s largest seven-nanometer processor.
Topping the announcements was the first Ampere-based GPU, the NVIDIA A100 (above), which delivers unifies AI training and inferencing and is 20 times faster than its predecessor. It’s a universal workload accelerator that is built for data analytics, scientific computing and cloud graphics.
“Ampere is going to give customers incredible benefits. In the future, the computing unit is going to be the entire data centre,” said Huang.
“For the first time, scale-up and scale-out workloads can be accelerated on one platform. NVIDIA A100 will simultaneously boost throughput and drive down the cost of data centres,” he added.
One of the first systems to sport the new chip is the NVIDIA DGX A100 (above), which features eight NVIDIA A100 GPUs interconnected with NVIDIA NVLink.
As the third generation of NVIDIA’s AI supercomputer-in-a-box, it delivers five petaflops of AI performance and consolidates the power and capabilities of an entire data centre into a single flexible platform for the first time.
Available immediately from US$199,000, its early adopters include Australia’s Harrison.ai, Thailand’s Chulalongkorn University and Vietnam’s VInAI Research.
Then, there’s the HGX A100, a server building block in the form of integrated baseboards in multiple GPU configurations. The four-GPU HGX A100 offers full interconnection between GPUs with NVLink, while the eight-GPU configuration offers full GPU-to-GPU bandwidth through NVIDIA NVSwitch.
With the new MIG technology, it can be configured as 56 small GPUs, each faster than NVIDIA T4, all the way up to a giant eight-GPU server with 10 petaflops of AI performance.
Two new EGX Ai platform products (above) are also based on Ampere to deliver high-performance, secure AI processing at the edge.. They are the EGX A100 for larger commercial off-the-shelf servers and the small EGX Jetson Xavier NX for micro-edge servers.
With the enterprise Ampere-based GPUs off the block, consumers are waiting excitedly on enjoying the same computing power with the next generation of GeForce cards.