Optimising Games for a Complex World
Talkat GCAP 2022: Ambition · on Wednesday, 2022-10-05· at 12:00· Room 109, 1 Convention Centre Place South Wharf, Victoria 3006·
summary
Describes a pipeline that turns per-frame telemetry into dashboards and stats: client telemetry (aggregated every 15–30 seconds), stored in Influx and visualized with Grafana + Influx telemetry, Python + pandas reprocessing millions of profiler rows into pandas box‑plots and KPIs (example: 90% time ≥55 FPS, 95th percentile 32.8 ms). Shows a real problem — emulation was 10–15% faster than real hardware — and a working mitigation: per-main and worker performance scalar (plus emergency mode) that throttles spawn rates, AI budgets and shadow/animation costs to hit targets.
AI-written summary · about
chapters (8)
- 0:00Game optimization challenges
- 2:12Measuring frame rate
- 7:15Performance telemetry
- 16:25Prioritizing optimizations
- 23:08Data science for performance
- 39:05Scaling techniques
- 45:02Key takeaways
- 47:25Q&A
related talks
01h 03mMaximizing Your Game's Performance in Unreal Engine
43mOptimizing animations in Unreal Engine 4
36mDebug, Analyze and Optimize Games with Intel Tools
58mUE4 Performance and Profiling
01h 00mUnreal Engine Game Optimization on a Budget
41mA Deep Look Inside Games: GameSDK and Unity Adaptive Performance
21mOptimization: a broad intro
57mOptimising Data Building In Game Development