More Speed and Simplicity - Practical Data-Oriented Design in C++

Explore Data-Oriented Design (DOD) as a practical alternative to traditional object-oriented programming in this conference talk from NDC TechTown 2025. Learn how prioritizing data layout over class hierarchies can unlock significant performance gains through improved cache efficiency while simultaneously simplifying code that processes data. Follow along with a step-by-step refactoring demonstration that transforms conventional OOP design into a data-first approach, complete with performance benchmarks and analysis of how data access patterns directly influence execution speed. Master key techniques including Structure-of-Arrays (SoA) versus Array-of-Structures (AoS) implementations, understanding their impact on both performance and code clarity. Discover how DOD can complement rather than replace OOP principles, and examine practical C++ techniques for creating safe, expressive APIs that balance complexity management with performance optimization. Gain insights into when and how to apply data-oriented thinking to make C++ code both faster and more conceptually clear, with pragmatic consideration of DOD's strengths in performance and data logic simplicity alongside its potential limitations.