Brendan Gregg
The rising cost of AI datacenters is a call for performance engineering: not just to use the best tools and techniques of yesterday, but to build new AI performance capabilities that are easier and faster. In 2000, performance experts were wizards, calculating system behavior from limited metrics and tools (e.g., from vmstat, iostat, and mpstat): It was the "performance math" era, and had many blind spots. By 2015, custom dashboards, dynamic instrumentation, eBPF, flame graphs, and an expectation of open source, enabled direct measurement almost everywhere, creating the "performance tools" era: Many tools, but also little time to learn them. By 2030 we will be well into the "performance prompts" era: AI prompts, skills, and agents to analyze all sources, including flame graphs, eBPF, Ftrace, PMCs, and MSRs. This talk summarizes these eras and shares performance prompts for code analysis, benchmarking, system metrics, and CPU analysis.

Brendan Gregg is an internationally renowned expert in computing performance. He works at OpenAI on performance optimization and cost reduction for ChatGPT. Previously at Netflix and Sun Microsystems, he authored Systems Performance and BPF Performance Tools (Addison-Wesley), and received the USENIX LISA Outstanding Achievement award. He has delivered industry-leading performance for various products, and created widely used performance tools, methodologies, and visualizations, including flame graphs. His work is the basis for multiple startups.

