Get 20% off all career paths from fullstack to AI
Learn EDR Internals: Research & Development From The Masters
Overview
AI, Data Science & Cloud Certificates from Google, IBM & Meta — 40% Off
One plan covers every Professional Certificate on Coursera. 40% off Coursera Plus Annual.
Unlock All Certificates
Explore the concept of Passthrough/Headless GPU in this 20-minute conference talk by Tina Zhang and Vivek Kasireddy from Intel. Delve into the challenges and proposed solutions for MultiGPU platforms, covering topics such as Blob Resources, Sync Objects, and GPU Copies. Gain insights into the advantages and potential drawbacks of this technology, and understand its implications for the future of GPU utilization in computing environments.
Syllabus
Introduction
Agenda
MultiGPU Platform Introduction
Challenges
Proposed Solution
Background
Blob Resources
Sync Objects
GPU Copies
Advantages Cons
Summary
Thank you
Taught by
Linux Foundation