Effectively Measure and Reduce Kernel Latencies for Real-time Constraints
Explore effective methods for measuring and reducing kernel latencies to meet real-time constraints in this informative conference talk. Delve into the concept of latency, examining how it relates to Linux scheduler latency, deferred execution methods, and task priorities. Learn experimental techniques to compare response latencies of different methods across Intel and ARM architectures. Discover how system load and user space task priorities impact latency. Gain insights into analyzing and improving Linux performance scalability in multi-core environments for real-time applications. Benefit from the expertise of speakers Jim Huang, who specializes in real-time and virtualization for Linux-based robots, and Chung-Fan Yang, a student focused on real-time Linux kernel experimentation and optimization techniques.