QTC, otherwise known as Quanta Cloud Technology which is a global data center solution provider, announced a new addition when it comes to its portfolio of NVIDIA-Certified Systems. This is known as QuantaGrid D43N-3U. It is a server that fits well alongside QCT’s existing high-performance QuantaGrid as well as QuantaPlex family of servers, which also enables customers the ability to leverage the latest, next-generation NVIDIA Tensor Core GPUs for advanced AI workloads, and this even includes those that run on virtualized infrastructure alongside the NVIDIA AI Enterprise software suite.
The NVIDIA Tensor Core GPUs on QCT systems are built on the ecosystem of AI framework, which is based on NVIDIA’s NGC catalog, CUDA-X libraries, and over 2.3 million developers with 1.800 GPU optimized applications that help enterprises solve some of the most critical challenges within their business.
Mike Yang, President of QCT, had the following to say on the subject: “Utilizing NVIDIA accelerated computing technologies, QCT is able to deliver innovative server designs for businesses to achieve faster results, quicker ROI, and reduced total cost of ownership. With the QuantaGrid D43N-3U, we are giving our customers a building block for an accelerated data center that simplifies deployments and consolidates a variety of configurations into a single solution for high-performance workloads.”
Since it is optimized for high-performance workloads, it also reduces the time to deployment for IT administrators and brings together compute acceleration alongside high capacity memory, PCIe Gen4, as well as high-speed, secure networking to servers while also handling three optimized configurations in a single 3U server design.
When it comes to accelerated HPC/AI training, cloud gaming, or even professional virtualization, this D43N-3U can even support double-width NVIDIA Ampere architecture-based GPUs, and this includes:
- NVIDIA A100 Tensor Core GPU
- NVIDIA RTX A6000
- NVIDIA A30 Tensor Core GPU
All of these are connected through the NVIDIA NVLink bridge board, which connects each pair of GPUs.
If we take a look at a scenario of mid-spec cloud gaming, it requires a lot of concurrent users, and the D43N-3U can fit up to eight NVIDIA A10 GPUs in order to reduce the total cost of ownership. For users that are looking to get better power efficiency out of their devices for AI multi-stream inference or intelligent video analytics, this same model can be set up with 8 NVIDIA T4 accelerators.
.@QuantaQCT today announced #QuantaGrid D43N-3U, a new addition to its broad portfolio of NVIDIA-Certified #GPU servers, delivering maximum acceleration for various workloads with three different GPU configurations. Learn more: https://t.co/l5VKrNHJv2 #GTC21 #HPC #AI @NVIDIAGTC pic.twitter.com/0G9hEoPglF
— QCT (@QuantaQCT) April 13, 2021
When it comes to Quanta Cloud Technology or QCT, they are a global data center solutions provider that combines the effectiveness of hyper-scale hardware with the infrastructure software from many leaders in the industry in order to solve the next-generation data center design as well as operational challenges. The QCT serves cloud service providers and enterprises running public, hybrid, and even private clouds.
Throughout history, these types of workloads would typically require a multitude of servers on a larger system in order to be accomplished.