Beyond OOP: A Journey From Cache Misses to MASS Hits
summary
Shows that poor object layout causes frequent 64-byte cache line misses and that reorganizing data into contiguous arrays yields orders-of-magnitude speedups. Demonstrates using Unreal Engine’s Mass Entity fragments/archetypes/processors instead of nested pointers, builds a toy pulsing-cube system and compares approaches, and reports a real measurement: iterating 10,000 entities stays under 1 ms with Mass while OOP costs scale far worse. Practical migration advice and use cases included.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:48CPU cache lines
- 8:41OOP problems in games
- 18:19MASS data-driven dev
- 26:15MASS architecture
- 32:03OOP vs MASS in games
- 33:28Q&A: MASS meaning
- 36:41Q&A: OOP to MASS rewrite
- 40:33Q&A: Migration paths
related talks
57mOptimising Data Building In Game Development
01h 21mData Oriented Design and Entity Component System Explained
28mRuntime optimization patterns - from games to your application
01h 00mUnreal Engine Game Optimization on a Budget
43mIntroduction to Data-Oriented Design in Unity
39mFantastic Bottlenecks and Where to Find Them
31mOptimization Techniques in UDK/UE3
50mProductivity in C++ Game Development: The Mind of an AAA Game Developer