Nvidia's vGPU July 2020 software provides new or expanded support for Suse Linux, VMware vSphere, and immersive collaboration.
With this update, vGPU is supported on Suse Linux Enterprise Server virtual machines, initially via Nvidia vComputeServer software for GPU virtualisation for AI and data science workloads.
“Demand for accelerated computing has grown beyond specialised HPC environments into virtualised data centres,” said Suse global CTO Brent Schroeder.
“To ensure the needs of business leaders are met, Suse and Nvidia have worked to simplify the use of Nvidia virtual GPUs in Suse Linux Enterprise Server. These efforts modernise the IT infrastructure and accelerate AI and ML workloads to enhance high-performance and time-sensitive workloads for Suse customers everywhere.”
GPU-enabled VMs are now supported by vSphere 7's Distributed Resource Scheduler. The 'assignable hardware' feature allows a VM to be automatically placed on a host that has exactly the right GPU and profile available before powering the VM on. This also works when a VM fails over under vSphere High Availability.
The new capability means new VMs can be deployed in a few minutes instead of requiring a manual process that could take hours.
vGPU July 2020 adds 120Hz VSync support at resolutions up to 4K, making it possible to deliver an even smoother VR or AR experience across 5G and Wi-Fi networks to the extent that it is indistinguishable from native tethered configurations.
“Streaming AR/VR over Wi-Fi or 5G enables organisations to truly take advantage of its benefits, enabling immersive training, product design and architecture and construction,” said VMware director of AR/VR Matt Coppinger.
“We’re partnering with Nvidia to more securely deliver AR and VR applications running on VMware vSphere and Nvidia Quadro Virtual Workstation, streamed using Nvidia CloudXR to VMware’s Project VXR client application running on standalone headsets.”
Other improvements include a better user experience and manageability for demanding workloads such as Omniverse AEC Experience for realtime photorealistic rendering of architectural and engineering models; provision for the host and guest vGPU software to be different versions, simplifying upgrades; and the ability to manage vGPU powered VMs with the latest release of VMware vRealize Operations.
“In a centralised computing environment with virtualised GPU technology, users no longer have to be tied to their physical workstations," said Jon Peddie Research/Cadalyst analyst Alex Herrera.
"As proven recently through remote work companies can turn on a dime, enabling anywhere/anytime access to big data without compromising on performance.”
Nvidia vGPU July 2020 will be released shortly.