Runtime optimization patterns - from games to your application
summary
Shows concrete patterns: using dirty flag to skip 49,400 unchanged entities in Cesium (50,000 cats, only ~600 update/sec) drops CPU from busy to under 2%; using an object pool (preallocate ~5,000 objects) removes allocation/fragmentation spikes for thousands of short-lived particle objects; reorganizing structures for data locality turns random-access passes that are ~90% slower into fast linear scans. Includes links to Cesium, Phaser examples and a memory-in-JS article.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 2:25Who am I
- 3:16Motivation
- 4:03Design patterns
- 7:41CPU
- 8:06CaesiumJS
- 8:24Washington DC
- 9:29Monitor
- 9:42Summary
- 10:10Rendering cycle
- 11:13General generalize
- 11:47Dirty flag
- 12:11Results
- 14:19Memory
- 15:43Memory allocation
- 16:45Fragmentation
- 17:41Object Pool
- 18:48Get New
- 19:20Pool
- 21:50Connection
- 22:10Bus
- 22:45Old Bus
- 23:14Cache
- 23:41Cache miss
- 26:23Data performance
- 26:58Motivation considerations
related talks
40mHow to Improve JavaScript Performance
23mOptimizing HTML5 Games: 10 Years of Learnings
57mOptimising Data Building In Game Development
41mBeyond OOP: A Journey From Cache Misses to MASS Hits
42mPractical Optimizations
28mOptimizations You Should Have Implemented Yesterday
39mFantastic Bottlenecks and Where to Find Them
20mOptimizing assets for graphics heavy Web Applicatoins: tips and tricks from the game dev background