Linux Kernel Autotuning for Optimal Performance

Explore an innovative solution for automating Linux kernel parameter tuning in this conference talk from the Linux Plumbers Conference. Delve into the challenges of optimizing performance across diverse workloads in large-scale data centers, and discover how machine learning algorithms like Bayesian optimization can outperform manual tuning. Examine the design and architecture of a comprehensive autotuning system, with a focus on memory management optimization. Gain insights into specific case studies demonstrating the effectiveness of this approach, and consider the potential for an in-kernel machine learning framework to further enhance Linux kernel optimization in kernel-space.