Insights into Entity Component Systems
summary
Claims a working Vienna Entity Component System VECS that uses slot maps and archetypes to prioritize cache-friendly, per-frame throughput. Shows concrete numbers: parallelization yields up to 50% speedup on large workloads, a live debugger handles snapshots and watches for runs with ~300,000 entities (~3.22 MB), and an LLM can generate a simple ECS quickly but not with VECS-level optimizations. Source and slides are linked on the conference page.
AI-written summary · about
chapters (18)
- 0:00Entity Component Systems intro
- 2:32Traditional OO problems
- 5:04How ECS solves issues
- 8:16Core ECS operations
- 12:12Strong-typed components
- 13:16Efficient entity iteration
- 15:02Choosing an ECS
- 17:32Performance considerations and data organization
- 20:41Cache efficiency benchmark
- 25:17Slot maps and archetypes
- 28:50Parallelizing Vienna ECS
- 32:03ECS thread pools
- 35:48Parallel performance gains
- 37:41Multi-threading future
- 39:41Debugging ECS live
- 41:10ECS snapshots analysis
- 42:56Live ECS monitoring
- 47:17ECS conclusion
related talks
43mClosing Keynote
01h 06mSIMD in C++20: eve of a new era
20mA Tour of the C++ Engine which Powers That Design MMO Called Figma
01h 05mC++ for Enterprise Applications
01h 00mJust-in-Time Compilation
46mEnarx - Attested, Secured Execution with AMD’s SEV
52mA Taste of Elixir - The language of the future you can have today
01h 03mThe most valuable values