GPU Accelerated Containers on Apple Silicon with libkrun and podman machine

Explore GPU acceleration for containerized AI workloads on Apple Silicon in this DevConf.US 2024 conference talk. Discover how podman machine and libkrun work together to enable running GPU-accelerated applications from containers on macOS. Learn about the challenges of running machine learning models locally on desktop systems, particularly with Apple's M-series GPUs. Understand the role of virtualization in using containers on macOS and how libkrun serves as a hypervisor backend for podman machine. Gain insights into recent enhancements that allow workloads to leverage Apple GPU acceleration. Witness a demonstration of an AI workload offloaded to the host GPU on macOS, and learn how to make the most of your hardware when running AI workloads in containers on Apple Silicon devices.