How we make Angular fast
summary
Ivy aims to make Angular noticeably faster by changing data layout and avoiding megamorphic property lookups and overuse of object shapes The talk shows concrete tools and numbers: microbenchmarks with V8 flags reveal inline/DeOpt pitfalls, then analysis with DeOptigate and Optigate and --prof uncovers a 2.5× slowdown at five shapes and up to a 50× slowdown at 10,000 shapes The built LView/TNode array strategy reduces memory and speeds runtime and the dev mode keeps debuggability.
AI-written summary · about
chapters (10)
- 0:00Intro
- 1:21Performance tests
- 6:47The fix
- 10:17Object layout
- 14:39Benchmarking
- 20:30Explanation
- 27:27Profiler
- 31:00Logical Trees
- 33:12Storage
- 35:49NGDev Mode
related talks
46mAdvancements in Angular
37mThe Secrets Behind Angular’s Lightning Speed
22mAngular’s new renderer Ivy: a deep dive into the heart of Angular
45mWhat's New with Angular
31mThe Rise of a new Engine: Angular Ivy
41mThe theory of Angular Ivy
54mBuilding Fast Angular Applications by Default
25mIvy Is Coming!