NVIDIA EGX: DPU Edge Computing Solution
NVIDIA today announced the launch of the new NVIDIA EGX AI platform for edge computing. The new solution can be applied in manufacturing, agriculture, logistics, telecommunications, and media to quickly and efficiently deploy AI systems.
The EGX platform combines the power of an NVIDIA Ampere GPU and Bluefield-2 DPU on a single PCIe card to create a compact and powerful solution. DPU Bluefield-2 includes Arm cores that can be used to accelerate GPU data and other tasks.
In addition to the hardware novelty, NVIDIA has introduced an updated software stack for EGX, available on NVIDIA NGC and in a separate software catalog, which additionally offers pre-configured AI tools for various industry solutions. These app platforms are supported by leading AI software vendors such as Deep North, Everseen, and Whiteboard Coordinator.
NVIDIA has partnered with leading OS and hypervisor vendors including Canonical, Cloudera, Red Hat, SUSE, and VMware to support and optimize the new NVIDIA software that can run on both bare metal and virtualized environments.
NVIDIA is committed to deploying AI analytics at the edge, right inside offices, factories, warehouses, cell towers, schools, stores, and banks. This should make such technologies more accessible to a wide range of consumers.
NVIDIA also announced the NVIDIA Fleet Command Early Access Program, which is a hybrid cloud platform with real-time data processing capabilities that will be delivered through software as a service (SaaS). Fleet Command will simplify the management of servers across multiple sites and reduce the burden on IT departments with proactive monitoring and an easy-to-use graphical console management system.
NVIDIA EGX systems are available from many certified server manufacturers: ASUS, Atos, Dell Technologies, Fujitsu, GIGABYTE, H2C, Inspur, Lenovo, Quanta, and Supermicro.
