The latest and greatest GPU architecture has been enabling our data center needs for quite a while now, and the power these accelerators deliver have solved some of the world’s biggest challenges, such as high-performance computing, data analytics, deep learning training and inference. Since the release of the NVIDIA Turing back in August 2018, we’ve been awaiting the next launch of the latest and greatest GPU. Just a month and a half after their GTC Digital Online Event, NVIDIA has announced its Ampere architecture and A100 Tensor Core GPU products that make up the next generation graphics architecture behind the world’s fastest server GPU.

Based on the new NVIDIA A100 Tensor Core GPU architecture, the GPU features third-generation Tensor Core technology with significantly higher performance over the previous generation to accelerate diverse workloads. It is more than just a chip though, as it builds on an ecosystem of deep learning frameworks and development tools: CUDA-X HPC libraries, over 700 HPC applications, and over 1.7 million developers. When combined with accelerated containerized software stacks from NGC, developers achieve unprecedented performance at scale well beyond Moore’s law that are accelerating time to discovery. The A100 powers QCT servers and provides Mellanox HDR Infiniband (IB), a new Multi-Instance GPU (MIG) technology, and NVIDIA NVSwitch™, which offers 600GB/sec of GPU-to-GPU interconnect, for scaling up and enabling AI inference and training in a broad range of fields, including healthcare, finance, and retail, which are key elements in global high performance computing market for enterprise and hyperscale. Prior to this announcement, QCT had begun using the NVIDIA HGX™ A100 8-GPU to expand and accelerate hyperscale data centers. The outstanding results are not only in the efficient scaling of thousands of GPUs with these integrated technologies, but also in the precision and training of some of the most complex AI networks at unprecedented speeds. Additionally, QCT plans to add the new NVIDIA EGX A100 configuration to its edge server portfolio to deliver enhanced security and unprecedented performance at the edge. The EGX A100 converged accelerator combines a Mellanox SmartNIC with GPUs powered by the new NVIDIA Ampere architecture, so enterprises can run AI at the edge more securely.

Follow QCT on Facebook, Linkedin, and Twitter to receive the latest news and announcements.