Feedback Driven Optimization: Improving Program Performance Through Compilation

Explore Feedback Driven Optimisation (FDO) in this 58-minute Linux Foundation conference talk. Delve into the history, challenges, and recent advancements of this code optimization technique. Learn about AutoFDO, a new approach originating from Google, and its implementation in GCC and LLVM. Discover real-life examples of successful FDO deployments and gain insights into performance analysis and optimization. Examine topics such as instrumentation-based FDO, swap optimization, profile optimization results, and sampling profile quality. Understand the limitations and practical applications of FDO in modern software development.

Introduction
Terminology
FDO Manual
Limit Instrumentation Based FDO
Example Program
Swap Optimization
Swap Optimization Results
Profile Optimization Results
Problems with FDO
Feedback Directed Optimization
Sampling Profile Quality
Non Perfect Condition
Where is it being used