NVIDIA brings state-of-the-art data center cooling system and AI advancement to COMPUTEX
American tech company NVIDIA brought the latest advancement for AI to COMPUTEX this year. As per its intro, the company went from making chips for game machines to having its tech recognized by a plethora of fields to gain the recognition it has today, from helping to create blockbuster movies to powering medical imaging devices.
For COMPUTEX HYBRID 2022, the company focused its presentation on its latest tech and software built for AI application in data centers, robotics as well as on gaming and content creation. The most eye-catching theme from NVIDIA this year was none other than how its latest tech will help cultivate the growth of AI.
Let’s Bring the “Cool”
Taking the stage first, NVIDIA’s Hyperscale and HPC Ian Buck explained how the information stored and the speed of processing such info are practically evolving data centers into AI factories.
The phenomenon then drove the company to engineer tech that are comparative in all levels from hardware to software, from chips to infrastructure to systems,” Buck said.
To power such devices to process massive data at super speed, Buck introduced the Hopper GPU chip, Grace CPU chip and BlueField DPUs that sets the framework which can collaborate seamlessly with the company’s Quantum and Spectrum networking switches.
Talking about the Grace CPU chip and Grace Hopper GPU chip, which networks both chip on one board, NVIDIA senior VP for hardware engineering Brian Kelleher lauded that the chips will help advance the fields of AI, data analytics, scientific computing, and hyperscale computing.
The company also announced that the Taiwan Titan partners who will adopt the so-called Grace-powered systems will be the likes of Foxconn Industrial Internet, ASUS, QCT, GIGABYTE and Wiwynn.
With NVIDIA’s “hot” new products, heat naturally follows all the computing power and speed that comes with it.
To both cool the temperature of the techs as well as keeping in line with the global “green” initiative, NVIDIA introduced the A100 PCIe GPU, its first data center PCIe GPU with direct chip liquid cooling.
The company said the GPU allows for businesses like cloud computing companies to lower energy use in their facilities via liquid cooling thanks to the addition of a water block component that is added to reduce data center hardware’s dependency on other cooling devices.
Advancing Artificial Intelligence
NVIDIA’s VP of Embedded and Edge Computing, Deepu Talla also touched on AI’s influence on automation and robotics, saying “We are entering the age of robotics — autonomous machines that are keenly aware of their environment and that can make smart decisions about their actions.”
Talla spoke about NVIDIA’s a robotics platform, Isaac, whose four pillars make the creation of AIs faster and easier, provide state-of-the-art simulation of a machine’s real-world application in a virtual setting that is almost true to life, construct the actual machines following tweaks learned from simulations, and managing fleets of the machines.
Talla also addressed the NVIDIA Jetson AGX Orin developer kit, which features the company’s Ampere architecture GPU, Arm Cortex-A78AE CPUs, next-generation deep learning and vision accelerators, high-speed interfaces, faster memory bandwidth, and multimodal sensor support that can manage many AI applications all at once.
Taiwan-based partners also announce and showcase Jetson new products during COMPUTEX including AAEON, ADLINK, Advantech, Aetina, AverMedia, Axiomtek, EverFocus, Neousys, Onyx and Vecow.