Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about NVIDIA's approach to bringing vGPU technology upstream through this technical conference talk that explores the architecture, implementation details, and future roadmap for virtualized GPU support. Discover how NVIDIA vGPU technology enables high-performance GPU capabilities in virtualized environments, supporting everything from graphics-intensive virtual desktops to AI and data science applications while combining hardware performance benefits with virtualization flexibility and manageability. Explore the proposed software architecture based on SR-IOV, where each vGPU is represented by a PCI Virtual Function managed through the standard Linux VFIO framework, and understand how the NVIDIA vGPU VFIO driver functions as a variant driver exposing standard userspace interfaces. Examine critical features including vGPU type selection, runtime creation and teardown of vGPU instances, and live migration capabilities, while learning how the driver interacts with a core driver responsible for hardware management. Gain insights into the architectural goals that enable the core driver to support DRM for host graphics, other NVIDIA GPU use cases, and the VFIO driver for vGPU functionality, and discover the upstream roadmap and areas where community input is most valuable for future development.
