This course covers the learning outcomes and goals of implementing a full GPU virtualization solution with mediated pass-through. Students will learn about the architecture of Intel Processor Graphics, pass-through, trap and emulation, render engine sharing, display management, graphics memory resource partitioning, secure isolation in gVirt, configurations, performance analysis for Linux and Windows VMs, scalability, overhead analysis, and summary of the technology. The course teaches skills related to GPU virtualization, performance optimization, secure isolation, and scalability. The teaching method includes theoretical explanations, architectural diagrams, experimental results, and performance analysis. The intended audience for this course includes virtualization engineers, system architects, GPU developers, and individuals interested in GPU virtualization technology.
A Full GPU Virtualization Solution with Mediated Pass-Through
Overview
Syllabus
Intro
GPU Usages
Virtualization Usages
Motivations for GPU Virtualization
GPU Virtualization Requirements
GPU Virtualization Approaches
Challenges
Architecture of Intel Processor Graphics
Pass Through
Trap and Emulation
Render Engine Sharing
Display Management
Graphics Memory Resource Partitioning
Address Space Ballooning Inconsistent view due to graphics memory partition
Secure Isolation in gVirt
Vulnerability from Direct Execution
Smart Shadowing
Lazy Shadowing
Configurations
Linux VM Performance
Windows VM Performance
Scalability
Overhead Analysis
Summary
Taught by
USENIX
