NVIDIA is envisioning a future where GPUs are leveraged to perform tasks too challenging for CPUs.
In the post-Moore’s Law future, NVIDIA is aiming to position its core competency in GPUs as accelerators that work in concert with traditional processors for general-purpose computing. NVIDIA GPUs are incorporated in 22 of the 25 most energy-efficient supercomputers, according to the Green500 list published in June 2019. At MWC Los Angeles, NVIDIA announced a handful of collaborations late Monday that bring GPUs from supercomputers that live in research labs to edge computing solutions that live on-premises, where data is being generated.
NVIDIA announced the EGX Edge computing platform, which it bills as a “high-performance, cloud-native platform that lets organizations harness rapidly streaming data” for AI, IoT, and “5G-based services.” The platform combines the CUDA-X general-purpose GPU (GPGPU) compute libraries with hardware ranging from the Jetson Nano single-board computer to full racks of NVIDIA T4 servers.
SEE: Special report: Managing AI and ML in the enterprise (free PDF) (TechRepublic)
The platform can be used with NVIDIA’s Metropolis application, which was designed for smart cities and intelligent video analytics use cases, as well as the new NVIDIA Aerial SDK for 5G virtualized radio access networks (5G vRAN).
Ericsson is utilizing EGX and Aerial for a 5G vRAN product, aimed at telecoms looking to virtualize operations, aiding in efforts to adequately distribute resources for emerging use cases, including augmented and virtual reality, as well as gaming.
Likewise, Red Hat is working with NVIDIA on making artificial intelligence (AI), data science, and machine learning more accessible at the edge, using EGX and OpenShift.
Microsoft is also partnering with NVIDIA, with the Metropolis video analytics framework operating on Azure IoT Edge and Azure Machine Learning offerings, as well as a T4 GPU-powered Azure Data Box Edge appliance.
NVIDIA named a handful of early adopters of the EGX, foremost among them Walmart’s “Intelligent Retail Lab,” a grocery store that tracks data using Metropolis to alert associates to “restock shelves, open up new checkout lanes, retrieve shopping carts and ensure product freshness in meat and produce departments.”